- 19 May, 2014 40 commits
-
-
Quinn Tran authored
Fix double free problem within qla2xxx driver where current code prematurely free qla_tgt_cmd while firmware still has the command. When firmware release the command after abort, the code attempt a second free as part of command completion processing. When TCM start the free process, NULL pointer was hit. ------ WARNING: CPU: 8 PID: 43613 at lib/list_debug.c:62 __list_del_entry+0x82/0xd0() list_del corruption. next->prev should be ffff88082b5cfb08, but was 6b6b6b6b6b6b6b6b CPU: 8 PID: 43613 Comm: kworker/8:0 Tainted: GF W O 3.13.0-rc3-nab_t10dif+ #6 Hardware name: HP ProLiant DL380p Gen8, BIOS P70 08/20/2012 Workqueue: events cache_reap 000000000000003e ffff88081b2e3c78 ffffffff815a051f 000000000000003e ffff88081b2e3cc8 ffff88081b2e3cb8 ffffffff8104fc2c 0000000000000000 ffff88082b5cfb00 ffff88081c788d00 ffff88082b5d7200 ffff88082b5d3080 Call Trace: [<ffffffff815a051f>] dump_stack+0x49/0x62 [<ffffffff8104fc2c>] warn_slowpath_common+0x8c/0xc0 [<ffffffff8104fd16>] warn_slowpath_fmt+0x46/0x50 [<ffffffff812b6592>] __list_del_entry+0x82/0xd0 [<ffffffff8106d48c>] process_one_work+0x12c/0x510 [<ffffffff8106d4d3>] ? process_one_work+0x173/0x510 [<ffffffff8106ebdf>] worker_thread+0x11f/0x3a0 [<ffffffff8106eac0>] ? manage_workers+0x170/0x170 [<ffffffff81074f26>] kthread+0xf6/0x120 [<ffffffff8109f103>] ? __lock_release+0x133/0x1b0 [<ffffffff81074e30>] ? __init_kthread_worker+0x70/0x70 [<ffffffff815aec2c>] ret_from_fork+0x7c/0xb0 [<ffffffff81074e30>] ? __init_kthread_worker+0x70/0x70 ---[ end trace dfc05c3f7caf8ebe ]--- BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 IP: [<ffffffff8106d391>] process_one_work+0x31/0x510 ------- Signed-off-by:
Quinn Tran <quinn.tran@qlogic.com> Signed-off-by:
Giridhar Malavali <giridhar.malavali@qlogic.com> Signed-off-by:
Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by:
Christoph Hellwig <hch@lst.de>
-
Quinn Tran authored
Add support for T10-Dif for Target Mode to qla driver. The driver will look for firmware attribute that support this feature. When the feature is present, the capabilities will be report to TCM layer. Add CTIO CRC2 iocb to build T10-Dif commands. Add support routines to process good & error cases. Signed-off-by:
Nicholas Bellinger <nab@linux-iscsi.org> Signed-off-by:
-
Himanshu Madhani authored
Signed-off-by:
Himanshu Madhani <himanshu.madhani@qlogic.com> Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
Joe Carnuccio <joe.carnuccio@qlogic.com> Signed-off-by:
-
Himanshu Madani authored
Signed-off-by:
Himanshu Madani <himanshu.madani@qlogic.com> Signed-off-by:
-
Tej Prakash authored
Signed-off-by:
Tej Prakash <tej.prakash@qlogic.com> Signed-off-by:
-
Sawan Chandak authored
Signed-off-by:
Sawan Chandak <sawan.chandak@qlogic.com> Signed-off-by:
-
Joe Carnuccio authored
For ISP27xx use the request/response queue index shadow registers to avoid directly access them on the PCI bus. Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Chad Dupuis authored
Signed-off-by:
Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by:
-
Hiral Patel authored
Signed-off-by:
Hiral Patel <hiral.patel@qlogic.com> Signed-off-by:
-
Chad Dupuis authored
Signed-off-by:
-
Armen Baloyan authored
Signed-off-by:
Armen Baloyan <armen.baloyan@qlogic.com> Signed-off-by:
-
Chad Dupuis authored
If the firmware cannot find the command specified then return SUCCESS to the error handler so as not to needlessly escalate. Also cleanup the resources for the command since we cannot expect the original command to returned in interrupt context. Signed-off-by:
-
Atul Deshmukh authored
Signed-off-by:
Atul Deshmukh <atul.deshmukh@qlogic.com> Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Chad Dupuis authored
Signed-off-by:
-
Atul Deshmukh authored
Signed-off-by:
-
Atul Deshmukh authored
Signed-off-by:
-
Armen Baloyan authored
Signed-off-by:
-
Armen Baloyan authored
Signed-off-by:
-
Joe Carnuccio authored
Increment the address by 4 on each iteration. Signed-off-by:
-
Hiral Patel authored
Signed-off-by:
-
Hiral Patel authored
Signed-off-by:
-
Hiral Patel authored
Signed-off-by:
-
Hiral Patel authored
Signed-off-by:
-
Joe Carnuccio authored
The window address inserted into the dump should be have been incremented by 1 rather than 4. Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Giridhar Malavali authored
Signed-off-by:
-
Pratik Mohanty authored
qla2xxx: Support of new firmware dump opcodes QLA8044_RDDFE(38), QLA8044_RDMDIO(39),QLA8044_POLLWR(40). Signed-off-by:
Pratik Mohanty <pratik.mohanty@qlogic.com> Signed-off-by:
-
Saurav Kashyap authored
Signed-off-by:
Giridhar Malavali <giridhar.malvali@qlogic.com> Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Giridhar Malavali authored
Signed-off-by:
-
Giridhar Malavali authored
Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Joe Carnuccio authored
Signed-off-by:
-
Chad Dupuis authored
If an adapter reset is already in progress, do not schedule another one when we receive an invalid status handle. Signed-off-by:
-
Saurav Kashyap authored
Signed-off-by:
-
Tomas Henzl authored
When the driver reads state values from the hw it might happen that different values are read in subsequent reads and this can cause problems, this may lead to a timeout in this function and a non working adapter. Cc: Adam Radford <aradford@gmail.com> Signed-off-by:
Tomas Henzl <thenzl@redhat.com> Reviewed-by:
Shintaro Minemoto <fj3207hq@aa.jp.fujitsu.com> Acked-by:
Sumit Saxena <sumit.saxena@lsi.com> Signed-off-by:
-
