linux-stable.git/drivers/target/target_core_device.c, branch v3.6

target: fail REPORT LUNS with less than 16 bytes of payload

2012-09-07T18:09:14+00:00

SPC says:

"The ALLOCATION LENGTH field is defined in 4.3.5.6. The allocation length
should be at least 16.  Device servers compliant with SPC return CHECK
CONDITION status, with the sense key set to ILLEGAL REQUEST, and the
additional sense code set to INVALID FIELD IN CDB when the allocation
length is less than 16 bytes".

Testcase: sg_raw -r8 /dev/sdb a0 00 00 00 00 00 00 00 00 08 00 00
    should fail with ILLEGAL REQUEST / INVALID FIELD IN CDB sense
    does not fail without the patch
    fails correctly with the patch

Signed-off-by: Paolo Bonzini 
Signed-off-by: Nicholas Bellinger

target: Make unnecessarily global se_dev_align_max_sectors() static

2012-07-17T00:35:35+00:00

Signed-off-by: Roland Dreier 
Signed-off-by: Nicholas Bellinger

target: Make core_disable_device_list_for_node use pre-refactoring lock ordering

2012-07-17T00:35:24+00:00

So after kicking around commit 547ac4c9c90 around a bit more, a tcm_qla2xxx LUN
unlink OP has generated the following warning:

[   50.386625] qla2xxx [0000:07:00.0]-00af:0: Performing ISP error recovery - ha=ffff880263774000.
[   70.572988] qla2xxx [0000:07:00.0]-8038:0: Cable is unplugged...
[  126.527531] ------------[ cut here ]------------
[  126.532677] WARNING: at kernel/softirq.c:159 local_bh_enable_ip+0x41/0x8c()
[  126.540433] Hardware name: S5520HC
[  126.544248] Modules linked in: tcm_vhost ib_srpt ib_cm ib_sa ib_mad ib_core tcm_qla2xxx tcm_loop tcm_fc libfc iscsi_target_mod target_core_pscsi target_core_file target_core_iblock target_core_mod configfs ipv6 iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi dm_round_robin dm_multipath scsi_dh loop i2c_i801 kvm_intel kvm crc32c_intel i2c_core microcode joydev button iomemory_vsl(O) pcspkr ext3 jbd uhci_hcd lpfc ata_piix libata ehci_hcd qla2xxx mlx4_core scsi_transport_fc scsi_tgt igb [last unloaded: scsi_wait_scan]
[  126.595567] Pid: 3283, comm: unlink Tainted: G           O 3.5.0-rc2+ #33
[  126.603128] Call Trace:
[  126.605853]  [] ? warn_slowpath_common+0x78/0x8c
[  126.612737]  [] ? local_bh_enable_ip+0x41/0x8c
[  126.619433]  [] ? core_disable_device_list_for_node+0x70/0xe3 [target_core_mod]
[  126.629323]  [] ? core_clear_lun_from_tpg+0x88/0xeb [target_core_mod]
[  126.638244]  [] ? core_tpg_post_dellun+0x17/0x48 [target_core_mod]
[  126.646873]  [] ? core_dev_del_lun+0x26/0x8c [target_core_mod]
[  126.655114]  [] ? dput+0x27/0x154
[  126.660549]  [] ? target_fabric_port_unlink+0x3b/0x41 [target_core_mod]
[  126.669661]  [] ? configfs_unlink+0xfc/0x14a [configfs]
[  126.677224]  [] ? vfs_unlink+0x58/0xb7
[  126.683141]  [] ? do_unlinkat+0xbb/0x142
[  126.689253]  [] ? page_fault+0x25/0x30
[  126.695170]  [] ? system_call_fastpath+0x16/0x1b
[  126.702053] ---[ end trace 2f8e5b0a9ec797ef ]---
[  126.756336] qla2xxx [0000:07:00.0]-00af:0: Performing ISP error recovery - ha=ffff880263774000.
[  146.942414] qla2xxx [0000:07:00.0]-8038:0: Cable is unplugged...

So this warning triggered because device_list disable logic is now
holding nacl->device_list_lock w/ spin_lock_irqsave before obtaining
port->sep_alua_lock with only spin_lock_bh..

The original disable logic obtains *deve ahead of dropping the entry
from deve->alua_port_list and then obtains ->device_list_lock to do the
remaining work.  Also, I'm pretty sure this particular warning is being
generated by a demo-mode session in tcm_qla2xxx, and not by explicit
NodeACL MappedLUNs.  The Initiator MappedLUNs are already protected by a
seperate configfs symlink reference back se_lun->lun_group, and the
demo-mode se_node_acl (and associated ->device_list[]) is released
during se_portal_group->tpg_group shutdown.

The following patch drops the extra functional change to disable logic
in commit 547ac4c9c90

Cc: Andy Grover 
Signed-off-by: Nicholas Bellinger

target: refactor core_update_device_list_for_node()

2012-07-17T00:35:24+00:00

Code was almost entirely divided based on value of bool param "enable".

Split it into two functions.

Signed-off-by: Andy Grover 
Signed-off-by: Nicholas Bellinger

target: Remove hba param from core_dev_add_lun

2012-07-17T00:35:22+00:00

Only used in a debugprint, and function signature is cleaner now.

Signed-off-by: Andy Grover 
Signed-off-by: Nicholas Bellinger

target: replace the processing thread with a TMR work queue

2012-07-17T00:35:21+00:00

The last functionality of the target processing thread is offloading possibly
long running task management requests from the submitter context.  To keep
TMR semantics the same we need a single threaded ordered queue, which can
be provided by a per-device workqueue with the right flags.

Signed-off-by: Christoph Hellwig 
Signed-off-by: Nicholas Bellinger

target: Remove max_sectors device attribute for modern se_task less code

2012-05-09T22:08:47+00:00

This patch removes the original usage of dev_attr->max_sectors in favor of
dev_attr->hw_max_sectors that is now being enforced by target core from
within transport_generic_cmd_sequencer() for SCF_SCSI_DATA_SG_IO_CDB ops.

After the recent se_task removal patches from hch, this value for IBLOCK
backends being set via configfs by userspace from an saved max_sectors
value that is turning out to be problematic, so it makes sense to go ahead
and remove this now legacy attribute all-together.

This patch also continues to make se_dev_set_default_attribs() do
(sectors / block_size) alignment for what actually get used by
target_core_mod to be safe here, following the same alignment currently
used by fabric_max_sectors.

Reported-by: Andy Grover 
Cc: Christoph Hellwig 
Cc: Roland Dreier 
Signed-off-by: Nicholas Bellinger

target: replace ->execute_task with ->execute_cmd

2012-05-06T22:11:14+00:00

Make CDB emulation work on commands instead of tasks again as a preparation
of removing tasks completely.

Signed-off-by: Christoph Hellwig 
Signed-off-by: Nicholas Bellinger

target: Remove obsolete DF_READ_ONLY usage

2012-04-14T23:04:58+00:00

This was used at one time as a hack by FILEIO backend registration to
allow a struct block_device that was claimed with blkdev_get (by a local
filesystem mount for example) to be exported as read-only (SCSI WP=1).

Since FILEIO backend registration will no longer attempt to obtain
exclusive access to an underlying struct block_device here, this flag is
now obsolete.

Reported-by: Roland Dreier 
Signed-off-by: Nicholas Bellinger

target: remove obvious warnings

2012-03-16T02:16:09+00:00

Get rid of a bunch of write-only variables.  In a number of cases I
suspect actual bugs to be present, so I left all of those for a second
look.

(nab: fix lio-core patch fuzz)

Signed-off-by: Joern Engel 
Signed-off-by: Nicholas Bellinger