david/ipxe
david
/
ipxe
Archived
1
0
Fork 0
This repository has been archived on 2020-12-06. You can view files and clone it, but cannot push or open issues or pull requests.
ipxe/src/net/aoe.c

1106 lines
28 KiB
C
Raw Normal View History

/*
* Copyright (C) 2006 Michael Brown <mbrown@fensystems.co.uk>.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation; either version 2 of the
* License, or any later version.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
* 02110-1301, USA.
*
* You can also choose to distribute this program under the terms of
* the Unmodified Binary Distribution Licence (as given in the file
* COPYING.UBDL), provided that you have satisfied its requirements.
*/
FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
#include <stddef.h>
#include <string.h>
2007-01-19 02:13:12 +01:00
#include <stdio.h>
#include <stdlib.h>
#include <errno.h>
#include <assert.h>
#include <byteswap.h>
#include <ipxe/list.h>
#include <ipxe/if_ether.h>
#include <ipxe/iobuf.h>
#include <ipxe/uaccess.h>
#include <ipxe/netdevice.h>
#include <ipxe/features.h>
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
#include <ipxe/interface.h>
#include <ipxe/xfer.h>
#include <ipxe/uri.h>
#include <ipxe/open.h>
#include <ipxe/ata.h>
#include <ipxe/device.h>
#include <ipxe/aoe.h>
/** @file
*
* AoE protocol
*
*/
FEATURE ( FEATURE_PROTOCOL, "AoE", DHCP_EB_FEATURE_AOE, 1 );
struct net_protocol aoe_protocol __net_protocol;
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
struct acpi_model abft_model __acpi_model;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/******************************************************************************
*
* AoE devices and commands
*
******************************************************************************
*/
/** List of all AoE devices */
static LIST_HEAD ( aoe_devices );
/** List of active AoE commands */
static LIST_HEAD ( aoe_commands );
/** An AoE device */
struct aoe_device {
/** Reference counter */
struct refcnt refcnt;
/** Network device */
struct net_device *netdev;
/** ATA command issuing interface */
struct interface ata;
/** Major number */
uint16_t major;
/** Minor number */
uint8_t minor;
/** Target MAC address */
uint8_t target[MAX_LL_ADDR_LEN];
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/** Saved timeout value */
unsigned long timeout;
/** Configuration command interface */
struct interface config;
/** Device is configued */
int configured;
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
/** ACPI descriptor */
struct acpi_descriptor desc;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
};
/** An AoE command */
struct aoe_command {
/** Reference count */
struct refcnt refcnt;
/** AOE device */
struct aoe_device *aoedev;
/** List of active commands */
struct list_head list;
/** ATA command interface */
struct interface ata;
/** ATA command */
struct ata_cmd command;
/** Command type */
struct aoe_command_type *type;
/** Command tag */
uint32_t tag;
/** Retransmission timer */
struct retry_timer timer;
};
/** An AoE command type */
struct aoe_command_type {
/**
* Calculate length of AoE command IU
*
* @v aoecmd AoE command
* @ret len Length of command IU
*/
size_t ( * cmd_len ) ( struct aoe_command *aoecmd );
/**
* Build AoE command IU
*
* @v aoecmd AoE command
* @v data Command IU
* @v len Length of command IU
*/
void ( * cmd ) ( struct aoe_command *aoecmd, void *data, size_t len );
/**
* Handle AoE response IU
*
* @v aoecmd AoE command
* @v data Response IU
* @v len Length of response IU
* @v ll_source Link-layer source address
* @ret rc Return status code
*/
int ( * rsp ) ( struct aoe_command *aoecmd, const void *data,
size_t len, const void *ll_source );
};
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/**
* Get reference to AoE device
*
* @v aoedev AoE device
* @ret aoedev AoE device
*/
static inline __attribute__ (( always_inline )) struct aoe_device *
aoedev_get ( struct aoe_device *aoedev ) {
ref_get ( &aoedev->refcnt );
return aoedev;
}
/**
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* Drop reference to AoE device
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* @v aoedev AoE device
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
static inline __attribute__ (( always_inline )) void
aoedev_put ( struct aoe_device *aoedev ) {
ref_put ( &aoedev->refcnt );
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/**
* Get reference to AoE command
*
* @v aoecmd AoE command
* @ret aoecmd AoE command
*/
static inline __attribute__ (( always_inline )) struct aoe_command *
aoecmd_get ( struct aoe_command *aoecmd ) {
ref_get ( &aoecmd->refcnt );
return aoecmd;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/**
* Drop reference to AoE command
*
* @v aoecmd AoE command
*/
static inline __attribute__ (( always_inline )) void
aoecmd_put ( struct aoe_command *aoecmd ) {
ref_put ( &aoecmd->refcnt );
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/**
* Name AoE device
*
* @v aoedev AoE device
* @ret name AoE device name
*/
static const char * aoedev_name ( struct aoe_device *aoedev ) {
static char buf[16];
snprintf ( buf, sizeof ( buf ), "%s/e%d.%d", aoedev->netdev->name,
aoedev->major, aoedev->minor );
return buf;
}
/**
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* Free AoE command
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* @v refcnt Reference counter
*/
static void aoecmd_free ( struct refcnt *refcnt ) {
struct aoe_command *aoecmd =
container_of ( refcnt, struct aoe_command, refcnt );
assert ( ! timer_running ( &aoecmd->timer ) );
assert ( list_empty ( &aoecmd->list ) );
aoedev_put ( aoecmd->aoedev );
free ( aoecmd );
}
/**
* Close AoE command
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* @v aoecmd AoE command
* @v rc Reason for close
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
static void aoecmd_close ( struct aoe_command *aoecmd, int rc ) {
struct aoe_device *aoedev = aoecmd->aoedev;
/* Stop timer */
stop_timer ( &aoecmd->timer );
/* Preserve the timeout value for subsequent commands */
aoedev->timeout = aoecmd->timer.timeout;
/* Remove from list of commands */
if ( ! list_empty ( &aoecmd->list ) ) {
list_del ( &aoecmd->list );
INIT_LIST_HEAD ( &aoecmd->list );
aoecmd_put ( aoecmd );
}
/* Shut down interfaces */
intf_shutdown ( &aoecmd->ata, rc );
}
/**
* Transmit AoE command request
*
* @v aoecmd AoE command
* @ret rc Return status code
*/
static int aoecmd_tx ( struct aoe_command *aoecmd ) {
struct aoe_device *aoedev = aoecmd->aoedev;
struct net_device *netdev = aoedev->netdev;
struct io_buffer *iobuf;
struct aoehdr *aoehdr;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
size_t cmd_len;
int rc;
/* Sanity check */
assert ( netdev != NULL );
/* If we are transmitting anything that requires a response,
* start the retransmission timer. Do this before attempting
* to allocate the I/O buffer, in case allocation itself
* fails.
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
start_timer ( &aoecmd->timer );
/* Create outgoing I/O buffer */
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
cmd_len = aoecmd->type->cmd_len ( aoecmd );
iobuf = alloc_iob ( MAX_LL_HEADER_LEN + cmd_len );
if ( ! iobuf )
return -ENOMEM;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
iob_reserve ( iobuf, MAX_LL_HEADER_LEN );
aoehdr = iob_put ( iobuf, cmd_len );
/* Fill AoE header */
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
memset ( aoehdr, 0, sizeof ( *aoehdr ) );
aoehdr->ver_flags = AOE_VERSION;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
aoehdr->major = htons ( aoedev->major );
aoehdr->minor = aoedev->minor;
aoehdr->tag = htonl ( aoecmd->tag );
aoecmd->type->cmd ( aoecmd, iobuf->data, iob_len ( iobuf ) );
/* Send packet */
if ( ( rc = net_tx ( iobuf, netdev, &aoe_protocol, aoedev->target,
netdev->ll_addr ) ) != 0 ) {
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
DBGC ( aoedev, "AoE %s/%08x could not transmit: %s\n",
aoedev_name ( aoedev ), aoecmd->tag,
strerror ( rc ) );
return rc;
}
return 0;
}
/**
* Receive AoE command response
*
* @v aoecmd AoE command
* @v iobuf I/O buffer
* @v ll_source Link-layer source address
* @ret rc Return status code
*/
static int aoecmd_rx ( struct aoe_command *aoecmd, struct io_buffer *iobuf,
const void *ll_source ) {
struct aoe_device *aoedev = aoecmd->aoedev;
struct aoehdr *aoehdr = iobuf->data;
int rc;
/* Sanity check */
if ( iob_len ( iobuf ) < sizeof ( *aoehdr ) ) {
DBGC ( aoedev, "AoE %s/%08x received underlength response "
"(%zd bytes)\n", aoedev_name ( aoedev ),
aoecmd->tag, iob_len ( iobuf ) );
rc = -EINVAL;
goto done;
}
if ( ( ntohs ( aoehdr->major ) != aoedev->major ) ||
( aoehdr->minor != aoedev->minor ) ) {
DBGC ( aoedev, "AoE %s/%08x received response for incorrect "
"device e%d.%d\n", aoedev_name ( aoedev ), aoecmd->tag,
ntohs ( aoehdr->major ), aoehdr->minor );
rc = -EINVAL;
goto done;
}
/* Catch command failures */
if ( aoehdr->ver_flags & AOE_FL_ERROR ) {
DBGC ( aoedev, "AoE %s/%08x terminated in error\n",
aoedev_name ( aoedev ), aoecmd->tag );
aoecmd_close ( aoecmd, -EIO );
rc = -EIO;
goto done;
}
/* Hand off to command completion handler */
if ( ( rc = aoecmd->type->rsp ( aoecmd, iobuf->data, iob_len ( iobuf ),
ll_source ) ) != 0 )
goto done;
done:
/* Free I/O buffer */
free_iob ( iobuf );
/* Terminate command */
aoecmd_close ( aoecmd, rc );
return rc;
}
/**
* Handle AoE retry timer expiry
*
* @v timer AoE retry timer
* @v fail Failure indicator
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
static void aoecmd_expired ( struct retry_timer *timer, int fail ) {
struct aoe_command *aoecmd =
container_of ( timer, struct aoe_command, timer );
if ( fail ) {
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
aoecmd_close ( aoecmd, -ETIMEDOUT );
} else {
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
aoecmd_tx ( aoecmd );
}
}
/**
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* Calculate length of AoE ATA command IU
*
* @v aoecmd AoE command
* @ret len Length of command IU
*/
static size_t aoecmd_ata_cmd_len ( struct aoe_command *aoecmd ) {
struct ata_cmd *command = &aoecmd->command;
return ( sizeof ( struct aoehdr ) + sizeof ( struct aoeata ) +
command->data_out_len );
}
/**
* Build AoE ATA command IU
*
* @v aoecmd AoE command
* @v data Command IU
* @v len Length of command IU
*/
static void aoecmd_ata_cmd ( struct aoe_command *aoecmd,
void *data, size_t len ) {
struct aoe_device *aoedev = aoecmd->aoedev;
struct ata_cmd *command = &aoecmd->command;
struct aoehdr *aoehdr = data;
struct aoeata *aoeata = &aoehdr->payload[0].ata;
/* Sanity check */
linker_assert ( AOE_FL_DEV_HEAD == ATA_DEV_SLAVE, __fix_ata_h__ );
assert ( len == ( sizeof ( *aoehdr ) + sizeof ( *aoeata ) +
command->data_out_len ) );
/* Build IU */
aoehdr->command = AOE_CMD_ATA;
memset ( aoeata, 0, sizeof ( *aoeata ) );
aoeata->aflags = ( ( command->cb.lba48 ? AOE_FL_EXTENDED : 0 ) |
( command->cb.device & ATA_DEV_SLAVE ) |
( command->data_out_len ? AOE_FL_WRITE : 0 ) );
aoeata->err_feat = command->cb.err_feat.bytes.cur;
aoeata->count = command->cb.count.native;
aoeata->cmd_stat = command->cb.cmd_stat;
aoeata->lba.u64 = cpu_to_le64 ( command->cb.lba.native );
if ( ! command->cb.lba48 )
aoeata->lba.bytes[3] |=
( command->cb.device & ATA_DEV_MASK );
copy_from_user ( aoeata->data, command->data_out, 0,
command->data_out_len );
DBGC2 ( aoedev, "AoE %s/%08x ATA cmd %02x:%02x:%02x:%02x:%08llx",
aoedev_name ( aoedev ), aoecmd->tag, aoeata->aflags,
aoeata->err_feat, aoeata->count, aoeata->cmd_stat,
aoeata->lba.u64 );
if ( command->data_out_len )
DBGC2 ( aoedev, " out %04zx", command->data_out_len );
if ( command->data_in_len )
DBGC2 ( aoedev, " in %04zx", command->data_in_len );
DBGC2 ( aoedev, "\n" );
}
/**
* Handle AoE ATA response IU
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* @v aoecmd AoE command
* @v data Response IU
* @v len Length of response IU
* @v ll_source Link-layer source address
* @ret rc Return status code
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
static int aoecmd_ata_rsp ( struct aoe_command *aoecmd, const void *data,
size_t len, const void *ll_source __unused ) {
struct aoe_device *aoedev = aoecmd->aoedev;
struct ata_cmd *command = &aoecmd->command;
const struct aoehdr *aoehdr = data;
const struct aoeata *aoeata = &aoehdr->payload[0].ata;
size_t data_len;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Sanity check */
if ( len < ( sizeof ( *aoehdr ) + sizeof ( *aoeata ) ) ) {
DBGC ( aoedev, "AoE %s/%08x received underlength ATA response "
"(%zd bytes)\n", aoedev_name ( aoedev ),
aoecmd->tag, len );
return -EINVAL;
}
data_len = ( len - ( sizeof ( *aoehdr ) + sizeof ( *aoeata ) ) );
DBGC2 ( aoedev, "AoE %s/%08x ATA rsp %02x in %04zx\n",
aoedev_name ( aoedev ), aoecmd->tag, aoeata->cmd_stat,
data_len );
/* Check for command failure */
if ( aoeata->cmd_stat & ATA_STAT_ERR ) {
DBGC ( aoedev, "AoE %s/%08x status %02x\n",
aoedev_name ( aoedev ), aoecmd->tag, aoeata->cmd_stat );
return -EIO;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Check data-in length is sufficient. (There may be trailing
* garbage due to Ethernet minimum-frame-size padding.)
*/
if ( data_len < command->data_in_len ) {
DBGC ( aoedev, "AoE %s/%08x data-in underrun (received %zd, "
"expected %zd)\n", aoedev_name ( aoedev ), aoecmd->tag,
data_len, command->data_in_len );
return -ERANGE;
}
/* Copy out data payload */
copy_to_user ( command->data_in, 0, aoeata->data,
command->data_in_len );
return 0;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/** AoE ATA command */
static struct aoe_command_type aoecmd_ata = {
.cmd_len = aoecmd_ata_cmd_len,
.cmd = aoecmd_ata_cmd,
.rsp = aoecmd_ata_rsp,
};
/**
* Calculate length of AoE configuration command IU
*
* @v aoecmd AoE command
* @ret len Length of command IU
*/
static size_t aoecmd_cfg_cmd_len ( struct aoe_command *aoecmd __unused ) {
return ( sizeof ( struct aoehdr ) + sizeof ( struct aoecfg ) );
}
/**
* Build AoE configuration command IU
*
* @v aoecmd AoE command
* @v data Command IU
* @v len Length of command IU
*/
static void aoecmd_cfg_cmd ( struct aoe_command *aoecmd,
void *data, size_t len ) {
struct aoe_device *aoedev = aoecmd->aoedev;
struct aoehdr *aoehdr = data;
struct aoecfg *aoecfg = &aoehdr->payload[0].cfg;
/* Sanity check */
assert ( len == ( sizeof ( *aoehdr ) + sizeof ( *aoecfg ) ) );
/* Build IU */
aoehdr->command = AOE_CMD_CONFIG;
memset ( aoecfg, 0, sizeof ( *aoecfg ) );
DBGC ( aoedev, "AoE %s/%08x CONFIG cmd\n",
aoedev_name ( aoedev ), aoecmd->tag );
}
/**
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* Handle AoE configuration response IU
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* @v aoecmd AoE command
* @v data Response IU
* @v len Length of response IU
* @v ll_source Link-layer source address
* @ret rc Return status code
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
static int aoecmd_cfg_rsp ( struct aoe_command *aoecmd, const void *data,
size_t len, const void *ll_source ) {
struct aoe_device *aoedev = aoecmd->aoedev;
struct ll_protocol *ll_protocol = aoedev->netdev->ll_protocol;
const struct aoehdr *aoehdr = data;
const struct aoecfg *aoecfg = &aoehdr->payload[0].cfg;
/* Sanity check */
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
if ( len < ( sizeof ( *aoehdr ) + sizeof ( *aoecfg ) ) ) {
DBGC ( aoedev, "AoE %s/%08x received underlength "
"configuration response (%zd bytes)\n",
aoedev_name ( aoedev ), aoecmd->tag, len );
return -EINVAL;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
DBGC ( aoedev, "AoE %s/%08x CONFIG rsp buf %04x fw %04x scnt %02x\n",
aoedev_name ( aoedev ), aoecmd->tag, ntohs ( aoecfg->bufcnt ),
aoecfg->fwver, aoecfg->scnt );
/* Record target MAC address */
memcpy ( aoedev->target, ll_source, ll_protocol->ll_addr_len );
DBGC ( aoedev, "AoE %s has MAC address %s\n",
aoedev_name ( aoedev ), ll_protocol->ntoa ( aoedev->target ) );
return 0;
}
/** AoE configuration command */
static struct aoe_command_type aoecmd_cfg = {
.cmd_len = aoecmd_cfg_cmd_len,
.cmd = aoecmd_cfg_cmd,
.rsp = aoecmd_cfg_rsp,
};
/** AoE command ATA interface operations */
static struct interface_operation aoecmd_ata_op[] = {
INTF_OP ( intf_close, struct aoe_command *, aoecmd_close ),
};
/** AoE command ATA interface descriptor */
static struct interface_descriptor aoecmd_ata_desc =
INTF_DESC ( struct aoe_command, ata, aoecmd_ata_op );
/**
* Identify AoE command by tag
*
* @v tag Command tag
* @ret aoecmd AoE command, or NULL
*/
static struct aoe_command * aoecmd_find_tag ( uint32_t tag ) {
struct aoe_command *aoecmd;
list_for_each_entry ( aoecmd, &aoe_commands, list ) {
if ( aoecmd->tag == tag )
return aoecmd;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
return NULL;
}
/**
* Choose an AoE command tag
*
* @ret tag New tag, or negative error
*/
static int aoecmd_new_tag ( void ) {
static uint16_t tag_idx;
unsigned int i;
for ( i = 0 ; i < 65536 ; i++ ) {
tag_idx++;
if ( aoecmd_find_tag ( tag_idx ) == NULL )
return ( AOE_TAG_MAGIC | tag_idx );
}
return -EADDRINUSE;
}
/**
* Create AoE command
*
* @v aoedev AoE device
* @v type AoE command type
* @ret aoecmd AoE command
*/
static struct aoe_command * aoecmd_create ( struct aoe_device *aoedev,
struct aoe_command_type *type ) {
struct aoe_command *aoecmd;
int tag;
/* Allocate command tag */
tag = aoecmd_new_tag();
if ( tag < 0 )
return NULL;
/* Allocate and initialise structure */
aoecmd = zalloc ( sizeof ( *aoecmd ) );
if ( ! aoecmd )
return NULL;
ref_init ( &aoecmd->refcnt, aoecmd_free );
list_add ( &aoecmd->list, &aoe_commands );
intf_init ( &aoecmd->ata, &aoecmd_ata_desc, &aoecmd->refcnt );
timer_init ( &aoecmd->timer, aoecmd_expired, &aoecmd->refcnt );
aoecmd->aoedev = aoedev_get ( aoedev );
aoecmd->type = type;
aoecmd->tag = tag;
/* Preserve timeout from last completed command */
aoecmd->timer.timeout = aoedev->timeout;
/* Return already mortalised. (Reference is held by command list.) */
return aoecmd;
}
/**
* Issue AoE ATA command
*
* @v aoedev AoE device
* @v parent Parent interface
* @v command ATA command
* @ret tag Command tag, or negative error
*/
static int aoedev_ata_command ( struct aoe_device *aoedev,
struct interface *parent,
struct ata_cmd *command ) {
struct net_device *netdev = aoedev->netdev;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
struct aoe_command *aoecmd;
/* Fail immediately if net device is closed */
if ( ! netdev_is_open ( netdev ) ) {
DBGC ( aoedev, "AoE %s cannot issue command while net device "
"is closed\n", aoedev_name ( aoedev ) );
return -EWOULDBLOCK;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Create command */
aoecmd = aoecmd_create ( aoedev, &aoecmd_ata );
if ( ! aoecmd )
return -ENOMEM;
memcpy ( &aoecmd->command, command, sizeof ( aoecmd->command ) );
/* Attempt to send command. Allow failures to be handled by
* the retry timer.
*/
aoecmd_tx ( aoecmd );
/* Attach to parent interface, leave reference with command
* list, and return.
*/
intf_plug_plug ( &aoecmd->ata, parent );
return aoecmd->tag;
}
/**
* Issue AoE configuration command
*
* @v aoedev AoE device
* @v parent Parent interface
* @ret tag Command tag, or negative error
*/
static int aoedev_cfg_command ( struct aoe_device *aoedev,
struct interface *parent ) {
struct aoe_command *aoecmd;
/* Create command */
aoecmd = aoecmd_create ( aoedev, &aoecmd_cfg );
if ( ! aoecmd )
return -ENOMEM;
/* Attempt to send command. Allow failures to be handled by
* the retry timer.
*/
aoecmd_tx ( aoecmd );
/* Attach to parent interface, leave reference with command
* list, and return.
*/
intf_plug_plug ( &aoecmd->ata, parent );
return aoecmd->tag;
}
/**
* Free AoE device
*
* @v refcnt Reference count
*/
static void aoedev_free ( struct refcnt *refcnt ) {
struct aoe_device *aoedev =
container_of ( refcnt, struct aoe_device, refcnt );
netdev_put ( aoedev->netdev );
free ( aoedev );
}
/**
* Close AoE device
*
* @v aoedev AoE device
* @v rc Reason for close
*/
static void aoedev_close ( struct aoe_device *aoedev, int rc ) {
struct aoe_command *aoecmd;
struct aoe_command *tmp;
/* Shut down interfaces */
intf_shutdown ( &aoedev->ata, rc );
intf_shutdown ( &aoedev->config, rc );
/* Shut down any active commands */
list_for_each_entry_safe ( aoecmd, tmp, &aoe_commands, list ) {
if ( aoecmd->aoedev != aoedev )
continue;
aoecmd_get ( aoecmd );
aoecmd_close ( aoecmd, rc );
aoecmd_put ( aoecmd );
}
}
/**
* Check AoE device flow-control window
*
* @v aoedev AoE device
* @ret len Length of window
*/
static size_t aoedev_window ( struct aoe_device *aoedev ) {
return ( aoedev->configured ? ~( ( size_t ) 0 ) : 0 );
}
/**
* Handle AoE device configuration completion
*
* @v aoedev AoE device
* @v rc Reason for completion
*/
static void aoedev_config_done ( struct aoe_device *aoedev, int rc ) {
/* Shut down interface */
intf_shutdown ( &aoedev->config, rc );
/* Close device on failure */
if ( rc != 0 ) {
aoedev_close ( aoedev, rc );
return;
}
/* Mark device as configured */
aoedev->configured = 1;
xfer_window_changed ( &aoedev->ata );
}
/**
* Identify device underlying AoE device
*
* @v aoedev AoE device
* @ret device Underlying device
*/
static struct device * aoedev_identify_device ( struct aoe_device *aoedev ) {
return aoedev->netdev->dev;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/**
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
* Get AoE ACPI descriptor
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
*
* @v aoedev AoE device
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
* @ret desc ACPI descriptor
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
*/
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
static struct acpi_descriptor * aoedev_describe ( struct aoe_device *aoedev ) {
return &aoedev->desc;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
}
/** AoE device ATA interface operations */
static struct interface_operation aoedev_ata_op[] = {
INTF_OP ( ata_command, struct aoe_device *, aoedev_ata_command ),
INTF_OP ( xfer_window, struct aoe_device *, aoedev_window ),
INTF_OP ( intf_close, struct aoe_device *, aoedev_close ),
INTF_OP ( acpi_describe, struct aoe_device *, aoedev_describe ),
INTF_OP ( identify_device, struct aoe_device *,
aoedev_identify_device ),
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
};
/** AoE device ATA interface descriptor */
static struct interface_descriptor aoedev_ata_desc =
INTF_DESC ( struct aoe_device, ata, aoedev_ata_op );
/** AoE device configuration interface operations */
static struct interface_operation aoedev_config_op[] = {
INTF_OP ( intf_close, struct aoe_device *, aoedev_config_done ),
};
/** AoE device configuration interface descriptor */
static struct interface_descriptor aoedev_config_desc =
INTF_DESC ( struct aoe_device, config, aoedev_config_op );
/**
* Open AoE device
*
* @v parent Parent interface
* @v netdev Network device
* @v major Device major number
* @v minor Device minor number
* @ret rc Return status code
*/
static int aoedev_open ( struct interface *parent, struct net_device *netdev,
unsigned int major, unsigned int minor ) {
struct aoe_device *aoedev;
int rc;
/* Allocate and initialise structure */
aoedev = zalloc ( sizeof ( *aoedev ) );
if ( ! aoedev ) {
rc = -ENOMEM;
goto err_zalloc;
}
ref_init ( &aoedev->refcnt, aoedev_free );
intf_init ( &aoedev->ata, &aoedev_ata_desc, &aoedev->refcnt );
intf_init ( &aoedev->config, &aoedev_config_desc, &aoedev->refcnt );
aoedev->netdev = netdev_get ( netdev );
aoedev->major = major;
aoedev->minor = minor;
memcpy ( aoedev->target, netdev->ll_broadcast,
netdev->ll_protocol->ll_addr_len );
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
acpi_init ( &aoedev->desc, &abft_model, &aoedev->refcnt );
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Initiate configuration */
if ( ( rc = aoedev_cfg_command ( aoedev, &aoedev->config ) ) < 0 ) {
DBGC ( aoedev, "AoE %s could not initiate configuration: %s\n",
aoedev_name ( aoedev ), strerror ( rc ) );
goto err_config;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Attach ATA device to parent interface */
if ( ( rc = ata_open ( parent, &aoedev->ata, ATA_DEV_MASTER,
AOE_MAX_COUNT ) ) != 0 ) {
DBGC ( aoedev, "AoE %s could not create ATA device: %s\n",
aoedev_name ( aoedev ), strerror ( rc ) );
goto err_ata_open;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Mortalise self and return */
ref_put ( &aoedev->refcnt );
return 0;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
err_ata_open:
err_config:
aoedev_close ( aoedev, rc );
ref_put ( &aoedev->refcnt );
err_zalloc:
return rc;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/******************************************************************************
*
* AoE network protocol
*
******************************************************************************
*/
/**
* Process incoming AoE packets
*
* @v iobuf I/O buffer
* @v netdev Network device
* @v ll_dest Link-layer destination address
* @v ll_source Link-layer source address
* @v flags Packet flags
* @ret rc Return status code
*/
static int aoe_rx ( struct io_buffer *iobuf,
struct net_device *netdev __unused,
const void *ll_dest __unused,
const void *ll_source,
unsigned int flags __unused ) {
struct aoehdr *aoehdr = iobuf->data;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
struct aoe_command *aoecmd;
int rc;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Sanity check */
if ( iob_len ( iobuf ) < sizeof ( *aoehdr ) ) {
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
DBG ( "AoE received underlength packet (%zd bytes)\n",
iob_len ( iobuf ) );
rc = -EINVAL;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
goto err_sanity;
}
if ( ( aoehdr->ver_flags & AOE_VERSION_MASK ) != AOE_VERSION ) {
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
DBG ( "AoE received packet for unsupported protocol version "
"%02x\n", ( aoehdr->ver_flags & AOE_VERSION_MASK ) );
rc = -EPROTONOSUPPORT;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
goto err_sanity;
}
if ( ! ( aoehdr->ver_flags & AOE_FL_RESPONSE ) ) {
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
DBG ( "AoE received request packet\n" );
rc = -EOPNOTSUPP;
goto err_sanity;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Demultiplex amongst active AoE commands */
aoecmd = aoecmd_find_tag ( ntohl ( aoehdr->tag ) );
if ( ! aoecmd ) {
DBG ( "AoE received packet for unused tag %08x\n",
ntohl ( aoehdr->tag ) );
rc = -ENOENT;
goto err_demux;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Pass received frame to command */
aoecmd_get ( aoecmd );
if ( ( rc = aoecmd_rx ( aoecmd, iob_disown ( iobuf ),
ll_source ) ) != 0 )
goto err_rx;
err_rx:
aoecmd_put ( aoecmd );
err_demux:
err_sanity:
free_iob ( iobuf );
return rc;
}
/** AoE protocol */
struct net_protocol aoe_protocol __net_protocol = {
.name = "AoE",
.net_proto = htons ( ETH_P_AOE ),
.rx = aoe_rx,
};
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/******************************************************************************
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* AoE URIs
*
******************************************************************************
*/
/**
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* Parse AoE URI
*
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
* @v uri URI
* @ret major Major device number
* @ret minor Minor device number
* @ret rc Return status code
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
*
* An AoE URI has the form "aoe:e<major>.<minor>".
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
static int aoe_parse_uri ( struct uri *uri, unsigned int *major,
unsigned int *minor ) {
const char *ptr;
char *end;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Check for URI with opaque portion */
if ( ! uri->opaque )
return -EINVAL;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
ptr = uri->opaque;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Check for initial 'e' */
if ( *ptr != 'e' )
return -EINVAL;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
ptr++;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Parse major device number */
*major = strtoul ( ptr, &end, 10 );
if ( *end != '.' )
return -EINVAL;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
ptr = ( end + 1 );
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Parse minor device number */
*minor = strtoul ( ptr, &end, 10 );
if ( *end )
return -EINVAL;
return 0;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/**
* Open AoE URI
*
* @v parent Parent interface
* @v uri URI
* @ret rc Return status code
*/
static int aoe_open ( struct interface *parent, struct uri *uri ) {
struct net_device *netdev;
unsigned int major;
unsigned int minor;
int rc;
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Identify network device. This is something of a hack, but
* the AoE URI scheme that has been in use for some time now
* provides no way to specify a particular device.
*/
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
netdev = last_opened_netdev();
if ( ! netdev ) {
DBG ( "AoE cannot identify network device\n" );
return -ENODEV;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Parse URI */
if ( ( rc = aoe_parse_uri ( uri, &major, &minor ) ) != 0 ) {
DBG ( "AoE cannot parse URI\n" );
return rc;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/* Open AoE device */
if ( ( rc = aoedev_open ( parent, netdev, major, minor ) ) != 0 )
return rc;
return 0;
}
[block] Replace gPXE block-device API with an iPXE asynchronous interface The block device interface used in gPXE predates the invention of even the old gPXE data-transfer interface, let alone the current iPXE generic asynchronous interface mechanism. Bring this old code up to date, with the following benefits: o Block device commands can be cancelled by the requestor. The INT 13 layer uses this to provide a global timeout on all INT 13 calls, with the result that an unexpected passive failure mode (such as an iSCSI target ACKing the request but never sending a response) will lead to a timeout that gets reported back to the INT 13 user, rather than simply freezing the system. o INT 13,00 (reset drive) is now able to reset the underlying block device. INT 13 users, such as DOS, that use INT 13,00 as a method for error recovery now have a chance of recovering. o All block device commands are tagged, with a numerical tag that will show up in debugging output and in packet captures; this will allow easier interpretation of bug reports that include both sources of information. o The extremely ugly hacks used to generate the boot firmware tables have been eradicated and replaced with a generic acpi_describe() method (exploiting the ability of iPXE interfaces to pass through methods to an underlying interface). The ACPI tables are now built in a shared data block within .bss16, rather than each requiring dedicated space in .data16. o The architecture-independent concept of a SAN device has been exposed to the iPXE core through the sanboot API, which provides calls to hook, unhook, boot, and describe SAN devices. This allows for much more flexible usage patterns (such as hooking an empty SAN device and then running an OS installer via TFTP). Signed-off-by: Michael Brown <mcb30@ipxe.org>
2010-09-03 17:11:51 +02:00
/** AoE URI opener */
struct uri_opener aoe_uri_opener __uri_opener = {
.scheme = "aoe",
.open = aoe_open,
};
[block] Describe all SAN devices via ACPI tables Describe all SAN devices via ACPI tables such as the iBFT. For tables that can describe only a single device (i.e. the aBFT and sBFT), one table is installed per device. For multi-device tables (i.e. the iBFT), all devices are described in a single table. An underlying SAN device connection may be closed at the time that we need to construct an ACPI table. We therefore introduce the concept of an "ACPI descriptor" which enables the SAN boot code to maintain an opaque pointer to the underlying object, and an "ACPI model" which can build tables from a list of such descriptors. This separates the lifecycles of ACPI descriptions from the lifecycles of the block device interfaces, and allows for construction of the ACPI tables even if the block device interface has been closed. For a multipath SAN device, iPXE will wait until sufficient information is available to describe all devices but will not wait for all paths to connect successfully. For example: with a multipath iSCSI boot iPXE will wait until at least one path has become available and name resolution has completed on all other paths. We do this since the iBFT has to include IP addresses rather than DNS names. We will commence booting without waiting for the inactive paths to either become available or close; this avoids unnecessary boot delays. Note that the Linux kernel will refuse to accept an iBFT with more than two NIC or target structures. We therefore describe only the NICs that are actually required in order to reach the described targets. Any iBFT with at most two targets is therefore guaranteed to describe at most two NICs. Signed-off-by: Michael Brown <mcb30@ipxe.org>
2017-03-27 17:20:34 +02:00
/******************************************************************************
*
* AoE boot firmware table (aBFT)
*
******************************************************************************
*/
/**
* Check if AoE boot firmware table descriptor is complete
*
* @v desc ACPI descriptor
* @ret rc Return status code
*/
static int abft_complete ( struct acpi_descriptor *desc __unused ) {
return 0;
}
/**
* Install AoE boot firmware table(s)
*
* @v install Installation method
* @ret rc Return status code
*/
static int abft_install ( int ( * install ) ( struct acpi_header *acpi ) ) {
struct aoe_device *aoedev;
struct abft_table abft;
int rc;
list_for_each_entry ( aoedev, &abft_model.descs, desc.list ) {
/* Populate table */
memset ( &abft, 0, sizeof ( abft ) );
abft.acpi.signature = cpu_to_le32 ( ABFT_SIG );
abft.acpi.length = cpu_to_le32 ( sizeof ( abft ) );
abft.acpi.revision = 1;
abft.shelf = cpu_to_le16 ( aoedev->major );
abft.slot = aoedev->minor;
memcpy ( abft.mac, aoedev->netdev->ll_addr,
sizeof ( abft.mac ) );
/* Install table */
if ( ( rc = install ( &abft.acpi ) ) != 0 ) {
DBGC ( aoedev, "AoE %s could not install aBFT: %s\n",
aoedev_name ( aoedev ), strerror ( rc ) );
return rc;
}
}
return 0;
}
/** aBFT model */
struct acpi_model abft_model __acpi_model = {
.descs = LIST_HEAD_INIT ( abft_model.descs ),
.complete = abft_complete,
.install = abft_install,
};