pandora-kernel.git
11 years ago[SCSI] prevent stack buffer overflow in host_reset
Sasha Levin [Thu, 15 Nov 2012 20:51:46 +0000 (15:51 -0500)]
[SCSI] prevent stack buffer overflow in host_reset

store_host_reset() has tried to re-invent the wheel to compare sysfs strings.
Unfortunately it did so poorly and never bothered to check the input from
userspace before overwriting stack with it, so something simple as:

echo "WoopsieWoopsie" >
/sys/devices/pseudo_0/adapter0/host0/scsi_host/host0/host_reset

would result in:

[  316.310101] Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: ffffffff81f5bac7
[  316.310101]
[  316.320051] Pid: 6655, comm: sh Tainted: G        W    3.7.0-rc5-next-20121114-sasha-00016-g5c9d68d-dirty #129
[  316.320051] Call Trace:
[  316.340058] pps pps0: PPS event at 1352918752.620355751
[  316.340062] pps pps0: capture assert seq #303
[  316.320051]  [<ffffffff83b3856b>] panic+0xcd/0x1f4
[  316.320051]  [<ffffffff81f5bac7>] ? store_host_reset+0xd7/0x100
[  316.320051]  [<ffffffff8110b996>] __stack_chk_fail+0x16/0x20
[  316.320051]  [<ffffffff81f5bac7>] store_host_reset+0xd7/0x100
[  316.320051]  [<ffffffff81e55bb3>] dev_attr_store+0x13/0x30
[  316.320051]  [<ffffffff812f7db1>] sysfs_write_file+0x101/0x170
[  316.320051]  [<ffffffff8127acc8>] vfs_write+0xb8/0x180
[  316.320051]  [<ffffffff8127ae80>] sys_write+0x50/0xa0
[  316.320051]  [<ffffffff83c03418>] tracesys+0xe1/0xe6

Fix this by uninventing whatever was going on there and just use sysfs_streq.

Bug introduced by 29443691 ("[SCSI] scsi: Added support for adapter and
firmware reset").

[jejb: added necessary const to prevent compile warnings]
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
Cc: <stable@vger.kernel.org> #3.2+
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] csiostor: Chelsio FCoE offload driver
Naresh Kumar Inna [Thu, 15 Nov 2012 17:11:18 +0000 (22:41 +0530)]
[SCSI] csiostor: Chelsio FCoE offload driver

Signed-off-by: Naresh Kumar Inna <naresh@chelsio.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] cxgb4/cxgb4vf: Chelsio FCoE offload driver submission (common header updates).
Naresh Kumar Inna [Thu, 15 Nov 2012 17:11:17 +0000 (22:41 +0530)]
[SCSI] cxgb4/cxgb4vf: Chelsio FCoE offload driver submission (common header updates).

This patch contains updates to firmware/hardware header files shared
between csiostor and cxgb4/cxgb4vf, and the resulting changes to the
cxgb4/cxgb4vf source files.

Signed-off-by: Naresh Kumar Inna <naresh@chelsio.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] sd: Use SCSI read/write(16) with > 32-bit LBA drives
Jason J. Herne [Wed, 14 Nov 2012 22:03:22 +0000 (17:03 -0500)]
[SCSI] sd: Use SCSI read/write(16) with > 32-bit LBA drives

Force large capacity (> 0xFFFFFFFF blocks) drives to use READ/WRITE(16) instead
of READ/WRITE(10). Some(most/all?) USB enclosures do not like READ(10) commands
when a large capacity drive is installed. This issue was reported and discussed
here: http://marc.info/?l=linux-usb&m=135247705222324

Signed-off-by: Jason J. Herne <hernejj@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] aacraid: SCSI dma mapping failure case handling
Mahesh Rajashekhara [Fri, 19 Oct 2012 13:51:07 +0000 (19:21 +0530)]
[SCSI] aacraid: SCSI dma mapping failure case handling

This patch handles SCSI dma mapping failure case. Reporting error code to the
upper layer instead of BUG_ON().

Signed-off-by: Mahesh Rajashekhara <Mahesh_Rajashekhara@pmc-sierra.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] virtio_scsi: fix memory leak on full queue condition.
Eric Northup [Thu, 8 Nov 2012 09:55:50 +0000 (01:55 -0800)]
[SCSI] virtio_scsi: fix memory leak on full queue condition.

virtscsi_queuecommand was leaking memory when the virtio queue was full.

Tested: Guest operates correctly even with very small queue sizes, validated
we're not leaking kmalloc-192 sized allocations anymore.

Signed-off-by: Eric Northup <digitaleric@google.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Update lpfc version for 8.3.36 driver release
James Smart [Wed, 31 Oct 2012 18:45:29 +0000 (14:45 -0400)]
[SCSI] lpfc 8.3.36: Update lpfc version for 8.3.36 driver release

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Correct mask error
James Smart [Wed, 31 Oct 2012 18:45:21 +0000 (14:45 -0400)]
[SCSI] lpfc 8.3.36: Correct mask error

Reported via: https://bugzilla.kernel.org/show_bug.cgi?id=48891

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Correct buffer length overrun
James Smart [Wed, 31 Oct 2012 18:45:09 +0000 (14:45 -0400)]
[SCSI] lpfc 8.3.36: Correct buffer length overrun

Reported via: https://bugzilla.kernel.org/show_bug.cgi?id=48871

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Fixed setting sequential delivery bit
James Smart [Wed, 31 Oct 2012 18:45:01 +0000 (14:45 -0400)]
[SCSI] lpfc 8.3.36: Fixed setting sequential delivery bit

Fixed setting sequential delivery bit in a service class that is not valid

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Update DIF support for passthru/strip/insert
James Smart [Wed, 31 Oct 2012 18:44:49 +0000 (14:44 -0400)]
[SCSI] lpfc 8.3.36: Update DIF support for passthru/strip/insert

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Fix bug with Target Resets and FCP2 devices
James Smart [Wed, 31 Oct 2012 18:44:42 +0000 (14:44 -0400)]
[SCSI] lpfc 8.3.36: Fix bug with Target Resets and FCP2 devices

Fix bug with Target Resets and FCP2 devices

Create module parameter to disable Target Reset on FCP-Tape devices
when a "bus reset" is requested. Default is to reset all devices on
bus reset.

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] lpfc 8.3.36: Fixed boot from san failure
James Smart [Wed, 31 Oct 2012 18:44:33 +0000 (14:44 -0400)]
[SCSI] lpfc 8.3.36: Fixed boot from san failure

Fixed boot from san failure when SLI4 FC device presented on the same PCI bus

The request_firmware interface can induce delays while looking
for firmware files, even if no fw file is present. In some situations
the delays exceeded scan_wait timeouts, resulting in situations in which
the boot device had not been discovered in time.  Boot Device does not
need to be on a lpfc device.

Change request_firmware use to be module paramater driven. Default is to
not attempt firmware download on boot. Add sysfs parameter to invoke
firmware update.

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] hptiop: Support HighPoint RR4520/RR4522 HBA
HighPoint Linux Team [Thu, 25 Oct 2012 00:41:52 +0000 (08:41 +0800)]
[SCSI] hptiop: Support HighPoint RR4520/RR4522 HBA

Support IOP RR4520/RR4522 which are based on Marvell frey.

Signed-off-by: HighPoint Linux Team <linux@highpoint-tech.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] bnx2i: removed the individual PCI DEVICE ID checking
Eddie Wai [Tue, 16 Oct 2012 00:31:37 +0000 (17:31 -0700)]
[SCSI] bnx2i: removed the individual PCI DEVICE ID checking

Removed the individual PCI DEVICE ID checking inside bnx2i.  The device
type can easily be read from the corresponding cnic->flags.  This will
free bnx2i from having to get updated for every new device ID that gets
added.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Acked-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Bump the driver version
John Soni Jose [Fri, 19 Oct 2012 23:16:04 +0000 (04:46 +0530)]
[SCSI] be2iscsi: Bump the driver version

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix Unrecoverable Error Detection
John Soni Jose [Fri, 19 Oct 2012 23:15:51 +0000 (04:45 +0530)]
[SCSI] be2iscsi: Fix Unrecoverable Error Detection

Driver periodically checks adapter state,is up fine or not.
Based on the value updates the internal structures of driver.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix for MBX timeout issue
John Soni Jose [Fri, 19 Oct 2012 23:15:40 +0000 (04:45 +0530)]
[SCSI] be2iscsi: Fix for MBX timeout issue

The MBX timeout value set to 100 and if adapter doesn;t
return response in that time driver will return from waiting
for completion with an error to the caller. In the earlier code
driver use to wait until MBX  response comes from adapter.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Update the copyright information
John Soni Jose [Fri, 19 Oct 2012 23:15:27 +0000 (04:45 +0530)]
[SCSI] be2iscsi: Update the copyright information

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix issue of displaying adapter family.
John Soni Jose [Fri, 19 Oct 2012 23:15:06 +0000 (04:45 +0530)]
[SCSI] be2iscsi: Fix issue of displaying adapter family.

Fix issue of displaying adapter family through the sysfs
entry for each Scsi_Host created for the adapter.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix Task Completion Event handling
John Soni Jose [Fri, 19 Oct 2012 23:14:49 +0000 (04:44 +0530)]
[SCSI] be2iscsi: Fix Task Completion Event handling

The completion events returned by adapter differs based on the
adapter. This fix checks for the adapter type and process the
completion event.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix session update context with V2 version.
John Soni Jose [Fri, 19 Oct 2012 23:14:35 +0000 (04:44 +0530)]
[SCSI] be2iscsi: Fix session update context with V2 version.

For updating session context on adapter, V2 version is to
be used with the latest adapter. This fix checks for the adapter type
and uses correct version of session context.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix support for V2 version of WRB.
John Soni Jose [Fri, 19 Oct 2012 23:14:23 +0000 (04:44 +0530)]
[SCSI] be2iscsi: Fix support for V2 version of WRB.

Latest adapters use the V2 version of WRB. This fix checks for the
adapter type and uses appropriate version of WRB.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix support for handling CQ_CREATE V2 version.
John Soni Jose [Fri, 19 Oct 2012 23:13:44 +0000 (04:43 +0530)]
[SCSI] be2iscsi: Fix support for handling CQ_CREATE V2 version.

For latest adapters, V2 version of cq_create MBX_CMD is to be used. When
driver is loaded depending on the adapter type appropriate cq_create command
will be called.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix max EQ supported by the driver.
John Soni Jose [Fri, 19 Oct 2012 23:13:32 +0000 (04:43 +0530)]
[SCSI] be2iscsi: Fix max EQ supported by the driver.

Fix the max EQ created when driver is loaded. Max EQ for
for new adapters will be 64.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix driver support for Skyhawk-R adapter.
John Soni Jose [Fri, 19 Oct 2012 23:13:20 +0000 (04:43 +0530)]
[SCSI] be2iscsi: Fix driver support for Skyhawk-R adapter.

Fix support for Skyhawk-R adapter by populating the pci_id_table

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix return value and typo.
John Soni Jose [Fri, 19 Oct 2012 23:13:03 +0000 (04:43 +0530)]
[SCSI] be2iscsi: Fix return value and typo.

Fix return value and typo in the message displayed.
Fix the goto label when wrb_hanlde allocation fails.
Fix the error message display in beiscsi_alloc_pdu

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix kernel panic in blk_iopoll disable mode.
John Soni Jose [Fri, 19 Oct 2012 23:12:49 +0000 (04:42 +0530)]
[SCSI] be2iscsi: Fix kernel panic in blk_iopoll disable mode.

Kernel used to panic while running IO is disable mode, as there
was an issue with getting the correct EQ on which completion has come.
Fix done is create workqueue per hba and work item for each EQ created.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Issue an function level reset when driver is loaded
John Soni Jose [Fri, 19 Oct 2012 23:12:37 +0000 (04:42 +0530)]
[SCSI] be2iscsi: Issue an function level reset when driver is loaded

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Display driver name and version in device attribute
John Soni Jose [Fri, 19 Oct 2012 23:12:25 +0000 (04:42 +0530)]
[SCSI] be2iscsi: Display driver name and version in device attribute

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix max supported EQ count to 8.
John Soni Jose [Fri, 19 Oct 2012 23:12:13 +0000 (04:42 +0530)]
[SCSI] be2iscsi: Fix max supported EQ count to 8.

The maximum EQ that can be created for a function is 8. Check the
CPU online count and create only 8 EQ if CPU_Count >= 8

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix memory leak in control path of driver
John Soni Jose [Fri, 19 Oct 2012 23:12:00 +0000 (04:42 +0530)]
[SCSI] be2iscsi: Fix memory leak in control path of driver

In contorl path of the driver the task was mapped using
pci_map_single which was not unmapped when the completion
for the task had come.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Display Completion Event string instead of Opcode
John Soni Jose [Fri, 19 Oct 2012 23:11:45 +0000 (04:41 +0530)]
[SCSI] be2iscsi: Display Completion Event string instead of Opcode

Display the event string along with the opcode and CID on which
an event has occured.

Signed-off-by: John Soni Jose <sony.john-n@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] be2iscsi: Fix the issue with soft reset.
Minh Tran [Fri, 19 Oct 2012 23:11:24 +0000 (04:41 +0530)]
[SCSI] be2iscsi: Fix the issue with soft reset.

Fixed soft_reset problem which driver modified
all 32bit before a write on second pass.

Signed-off-by: Minh Tran <minhduc.tran@emulex.com>
Signed-off-by: Jayamohan Kallickal <jayamohan.kallickal@emulex.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] osd_uld: Add osdname & systemid sysfs at scsi_osd class
Boaz Harrosh [Wed, 24 Oct 2012 21:51:41 +0000 (14:51 -0700)]
[SCSI] osd_uld: Add osdname & systemid sysfs at scsi_osd class

This patch adds the support for the following two read-only sysfs attributes
to scsi_osd class members : osdname & systemid

These attributes will show up as below in sysfs class hierarchy:
/sys/class/scsi_osd/osdX/osdname
/sys/class/scsi_osd/osdX/systemid

The osdname & systemid are OSD device attributes which uniquely
identify a device on the network, while it's IP and certainly
it's /dev/osdX device path might change.
Userspace utilities (e.g. mkfs.exofs) can parse these attributes to
identify the correct OSD in safer and faster way.

(Today osd apps open each device in the system and send a
 attributes query for these, in order to access the user
 requested device)

Signed-off-by: Sachin Bhamare <sbhamare@panasas.com>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] sd: Reshuffle init_sd to avoid crash
Joel D. Diaz [Wed, 10 Oct 2012 08:36:11 +0000 (10:36 +0200)]
[SCSI] sd: Reshuffle init_sd to avoid crash

scsi_register_driver will register a prep_fn() function, which
in turn migh need to use the sd_cdp_pool for DIF.
Which hasn't been initialised at this point, leading to
a crash. So reshuffle the init_sd() and exit_sd() paths
to have the driver registered last.

Signed-off-by: Joel D. Diaz <joeldiaz@us.ibm.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years ago[SCSI] MAINTAINERS: update LSILOGIC MPT FUSION DRIVERS (FC/SAS/SPI) maintainers
sreekanth.reddy@lsi.com [Tue, 9 Oct 2012 05:34:22 +0000 (11:04 +0530)]
[SCSI] MAINTAINERS: update LSILOGIC MPT FUSION DRIVERS (FC/SAS/SPI) maintainers

Signed-off-by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years agoLinux 3.7-rc7 v3.7-rc7
Linus Torvalds [Mon, 26 Nov 2012 01:59:19 +0000 (17:59 -0800)]
Linux 3.7-rc7

11 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Mon, 26 Nov 2012 01:57:01 +0000 (17:57 -0800)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

Pull powerpc EEH bugfixes from Benjamin Herrenschmidt.

Two one-liner fixes for the new EEH code.

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/eeh: Do not invalidate PE properly
  powerpc/pseries: Fix oops with MSIs when missing EEH PEs

11 years agoMerge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus
Linus Torvalds [Mon, 26 Nov 2012 01:55:04 +0000 (17:55 -0800)]
Merge branch 'upstream' of git://git.linux-mips.org/ralf/upstream-linus

Pull MIPS fixes from Ralf Baechle:
 "Three issues fixed accross the field:

   - Some functions that were recently outlined as part of a preemption
     fix were causing problems with function tracing.
   - The recently merged in-kernel MPI library uses very outdated
     headers that contain MIPS-specific code which won't build on with
     gcc 4.4 or newer.
   - The MIPS non-NUMA memory initialization was making only a very
     half-baked attempt at merging adjacent memory ranges.  This kept
     the code simple enough but is now causing issues with kexec."

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
  MPI: Fix compilation on MIPS with GCC 4.4 and newer
  MIPS: Fix crash that occurs when function tracing is enabled
  MIPS: Merge overlapping bootmem ranges

11 years agopowerpc/eeh: Do not invalidate PE properly
Gavin Shan [Thu, 22 Nov 2012 21:58:26 +0000 (21:58 +0000)]
powerpc/eeh: Do not invalidate PE properly

While the EEH does recovery on the specific PE that has PCI errors,
the PCI devices belonging to the PE will be removed and the PE will
be marked as invalid since we still need the information stored in
the PE. We only invalidate the PE when it doesn't have associated
EEH devices and valid child PEs. However, the code used to check
that is wrong. The patch fixes that.

Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agoMerge tag 'sound-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Linus Torvalds [Sat, 24 Nov 2012 18:32:11 +0000 (08:32 -1000)]
Merge tag 'sound-3.7' of git://git./linux/kernel/git/tiwai/sound

Pull sound build error fix from Takashi Iwai:
 "Only a single commit for fixing the build error without CONFIG_PM in
  hda driver."

* tag 'sound-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: hda - Fix build without CONFIG_PM

11 years agoALSA: hda - Fix build without CONFIG_PM
Takashi Iwai [Sat, 24 Nov 2012 10:58:24 +0000 (11:58 +0100)]
ALSA: hda - Fix build without CONFIG_PM

I forgot this again...  codec->in_pm is in #ifdef CONFIG_PM

Reported-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
11 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 24 Nov 2012 06:03:14 +0000 (20:03 -1000)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 arch fixes from Peter Anvin:
 "Here is a collection of fixes for 3.7-rc7.  This is a superset of
  tglx' earlier pull request."

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86-64: Fix ordering of CFI directives and recent ASM_CLAC additions
  x86, microcode, AMD: Add support for family 16h processors
  x86-32: Export kernel_stack_pointer() for modules
  x86-32: Fix invalid stack address while in softirq
  x86, efi: Fix processor-specific memcpy() build error
  x86: remove dummy long from EFI stub
  x86, mm: Correct vmflag test for checking VM_HUGETLB
  x86, amd: Disable way access filter on Piledriver CPUs
  x86/mce: Do not change worker's running cpu in cmci_rediscover().
  x86/ce4100: Fix PCI configuration register access for devices without interrupts
  x86/ce4100: Fix reboot by forcing the reboot method to be KBD
  x86/ce4100: Fix pm_poweroff
  MAINTAINERS: Update email address for Robert Richter
  x86, microcode_amd: Change email addresses, MAINTAINERS entry
  MAINTAINERS: Change Boris' email address
  EDAC: Change Boris' email address
  x86, AMD: Change Boris' email address

11 years agoMerge tag 'for-linus-20121123' of git://git.infradead.org/mtd-2.6
Linus Torvalds [Sat, 24 Nov 2012 01:12:17 +0000 (15:12 -1000)]
Merge tag 'for-linus-20121123' of git://git.infradead.org/mtd-2.6

Pull MTD fixes from David Woodhouse:
 "The most important part of this is that it fixes a regression in
  Samsung NAND chip detection, introduced by some rework which went into
  3.7.  The initial fix wasn't quite complete, so it's in two parts.  In
  fact the first part is committed twice (Artem committed his own copy
  of the same patch) and I've merged Artem's tree into mine which
  already had that fix.

  I'd have recommitted that to make it somewhat cleaner, but figured by
  this point in the release cycle it was better to merge *exactly* the
  commits which have been in linux-next.

  If I'd recommitted, I'd also omit the sparse warning fix.  But it's
  there, and it's harmless â€” just marking one function as 'static' in
  onenand code.

  This also includes a couple more fixes for stable: an AB-BA deadlock
  in JFFS2, and an invalid range check in slram."

* tag 'for-linus-20121123' of git://git.infradead.org/mtd-2.6:
  mtd: nand: fix Samsung SLC detection regression
  mtd: nand: fix Samsung SLC NAND identification regression
  jffs2: Fix lock acquisition order bug in jffs2_write_begin
  mtd: onenand: Make flexonenand_set_boundary static
  mtd: slram: invalid checking of absolute end address
  mtd: ofpart: Fix incorrect NULL check in parse_ofoldpart_partitions()
  mtd: nand: fix Samsung SLC NAND identification regression

11 years agoMerge tag 'devicetree-for-linus' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Fri, 23 Nov 2012 22:36:06 +0000 (12:36 -1000)]
Merge tag 'devicetree-for-linus' of git://git.secretlab.ca/git/linux-2.6

Pull device tree regression fix from Grant Likely:
 "Simple build regression fix for DT device drivers on Sparc.  An
  earlier change had masked out the of_iomap() helper on SPARC."

* tag 'devicetree-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  of/address: sparc: Declare of_iomap as an extern function for sparc again

11 years agoMerge tag 'pm-for-3.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 23 Nov 2012 22:16:43 +0000 (12:16 -1000)]
Merge tag 'pm-for-3.7-rc7' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management update from Rafael Wysocki:
 "Fix for an incorrect error condition check in device PM QoS code that
  may lead to an Oops from Guennadi Liakhovetski."

* tag 'pm-for-3.7-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  PM / QoS: fix wrong error-checking condition

11 years agoMerge tag 'md-3.7-fixes' of git://neil.brown.name/md
Linus Torvalds [Fri, 23 Nov 2012 22:11:13 +0000 (12:11 -1000)]
Merge tag 'md-3.7-fixes' of git://neil.brown.name/md

Pull md fixes from NeilBrown:
 "Several bug fixes for md in 3.7:

   - raid5 discard has problems
   - raid10 replacement devices have problems
   - bad block lock seqlock usage has problems
   - dm-raid doesn't free everything"

* tag 'md-3.7-fixes' of git://neil.brown.name/md:
  md/raid10: decrement correct pending counter when writing to replacement.
  md/raid10: close race that lose writes lost when replacement completes.
  md/raid5: Make sure we clear R5_Discard when discard is finished.
  md/raid5: move resolving of reconstruct_state earlier in stripe_handle.
  md/raid5: round discard alignment up to power of 2.
  md: make sure everything is freed when dm-raid stops an array.
  md: Avoid write invalid address if read_seqretry returned true.
  md: Reassigned the parameters if read_seqretry returned true in func md_is_badblock.

11 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 23 Nov 2012 22:06:05 +0000 (12:06 -1000)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-block

Pull block layer fixes from Jens Axboe:
 "Distilled down version of bug fixes for 3.7.  The patches have been
  well tested.  If you notice that commit dates are from today, it's
  because I pulled less important bits out and shuffled them into the
  3.8 mix.  Apart from that, no changes, base still the same.

  It contains:

   - Fix for aoe, don't run request_fn while it's plugged.

   - Fix for a regression in floppy since 3.6, which causes problems if
     no floppy is found.

   - Stable fix for blk_exec(), don't touch a request after it has been
     sent to the scheduler (and the device as well).

   - Five fixes for various nasties in mtip32xx."

* 'for-linus' of git://git.kernel.dk/linux-block:
  block: Don't access request after it might be freed
  mtip32xx: Fix padding issue
  aoe: avoid running request handler on plugged queue
  mtip32xx: fix potential NULL pointer dereference in mtip_timeout_function()
  mtip32xx: fix shift larger than type warning
  mtip32xx: Fix incorrect mask used for erase mode
  mtip32xx: Fix to make lba address correct in big-endian systems
  mtip32xx: fix potential crash on SEC_ERASE_UNIT
  dm: fix deadlock with request based dm and queue request_fn recursion
  floppy: destroy floppy workqueue before cleaning up the queue

11 years agoof/address: sparc: Declare of_iomap as an extern function for sparc again
Andreas Larsson [Fri, 23 Nov 2012 11:24:09 +0000 (12:24 +0100)]
of/address: sparc: Declare of_iomap as an extern function for sparc again

This bug-fix makes sure that of_iomap is defined extern for sparc so that the
sparc-specific implementation of_iomap is once again used when including
include/linux/of_address.h in a sparc context. OF_GPIO that is now available for
sparc relies on this.

The bug was inadvertently introduced in a850a75, "of/address: add empty static
inlines for !CONFIG_OF", that added a static dummy inline for of_iomap when
!CONFIG_OF_ADDRESS. However, CONFIG_OF_ADDRESS is never defined for sparc, but
there is a sparc-specific implementation /arch/sparc/kernel/of_device_common.c.

This fix takes the same approach as 0bce04b that solved the equivalent problem
for of_address_to_resource.

Signed-off-by: Andreas Larsson <andreas@gaisler.com>
Acked-by: David Miller <davem@davemloft.net>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
11 years agoMerge tag 'omapdss-for-3.7-rc' of git://gitorious.org/linux-omap-dss2/linux
Linus Torvalds [Fri, 23 Nov 2012 22:01:02 +0000 (12:01 -1000)]
Merge tag 'omapdss-for-3.7-rc' of git://gitorious.org/linux-omap-dss2/linux

Pull omapdss fixes from Tomi Valkeinen:
 "Here are a few OMAPDSS fixes for the next -rc.  I'm sending these
  directly to you, and quite late, as the fbdev tree maintainer
  (Florian) has been busy with his work and hasn't had time to manage
  the fb patches."

* tag 'omapdss-for-3.7-rc' of git://gitorious.org/linux-omap-dss2/linux:
  OMAPDSS: do not fail if dpll4_m4_ck is missing
  OMAPFB: Fix possible null pointer dereferencing
  OMAPDSS: HDMI: fix missing unlock on error in hdmi_dump_regs()
  omapdss: dss: Fix clocks on OMAP363x
  OMAPDSS: DSI: fix dsi_get_dsidev_from_id()

11 years agoMerge branch 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux
Linus Torvalds [Fri, 23 Nov 2012 21:59:26 +0000 (11:59 -1000)]
Merge branch 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "Bugfixes for the i2c subsystem.

  Except for a few one-liners, there is mainly one revert because of an
  overlooked dependency.  Since there is no linux-next at the moment, I
  did some extra testing, and all was fine for me."

* 'i2c-embedded/for-current' of git://git.pengutronix.de/git/wsa/linux:
  i2c: mxs: Handle i2c DMA failure properly
  i2c: s3c2410: Fix code to free gpios
  i2c: omap: ensure writes to dev->buf_len are ordered
  Revert "ARM: OMAP: convert I2C driver to PM QoS for MPU latency constraints"
  i2c: at91: fix SMBus quick command

11 years agoMerge tag 'sound-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound
Linus Torvalds [Fri, 23 Nov 2012 21:58:28 +0000 (11:58 -1000)]
Merge tag 'sound-3.7' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "The highlight of this update is the fixes for ASoC kirkwood by
  Russell.  In addition to that, a couple of regression fixes for
  HD-audio due to the runtime PM support on 3.7, and other driver-
  specific regression fixes like USB MIDI on non-standard USB audio
  drivers."

* tag 'sound-3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: snd-usb: properly initialize the sync endpoint
  ALSA: hda - Cirrus: Correctly clear line_out_pins when moving to speaker
  ALSA: hda - Add support for Realtek ALC292
  ASoC: kirkwood-i2s: more pause-mode fixes
  ASoC: kirkwood-i2s: fix DMA underruns
  ASoC: kirkwood-i2s: fix DCO lock detection
  ASoC: kirkwood-dma: don't ignore other irq causes on error
  ASoC: kirkwood-dma: fix use of virt_to_phys()
  ALSA: hda - Limit runtime PM support only to known Intel chips
  ALSA: hda - Fix recursive suspend/resume call
  ALSA: ua101, usx2y: fix broken MIDI output
  ASoC: arizona: Fix typo - Swap value in 48k_rates[] and 44k1_rates[]
  ASoC: bells: Fix up git patch application failure
  ASoC: cs4271: free allocated GPIO

11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Fri, 23 Nov 2012 21:55:49 +0000 (11:55 -1000)]
Merge git://git./linux/kernel/git/davem/net

Pull networkign fixes from David Miller:
 "Networking bug fixes, Cacio e Pepe edition:

  1) BNX2X accidently accesses chip rev specific registers without an
     appropriate guard, fix from Ariel Elior.

  2) When we removed the routing cache, we set ip_rt_max_size to ~0 just
     to keep reporting a value to userspace via sysfs.  But the ipv4
     IPSEC layer was using this to tune itself which is completely bogus
     to now do.  Fix from Steffen Klassert.

  3) Missing initialization in netfilter ipset code from Jozsef
     Kadlecsik.

  4) Check CTA_TIMEOUT_NAME length properly in netfilter cttimeout code,
     fix from Florian Westphal.

  5) After removing the routing cache, we inadvertantly are caching
     multicast routes that end up looping back locally, we cannot do
     that legitimately any more.  Fix from Julian Anastasov.

  6) Revert a race fix for 8139cp qemu/kvm that doesn't actually work
     properly on real hardware.  From Francois Romieu.

  7) Fixup errors in example command lines in VXLAN device docs."

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  bnx2x: remove redundant warning log
  vxlan: fix command usage in its doc
  8139cp: revert "set ring address before enabling receiver"
  ipv4: do not cache looped multicasts
  netfilter: cttimeout: fix buffer overflow
  netfilter: ipset: Fix range bug in hash:ip,port,net
  xfrm: Fix the gc threshold value for ipv4

11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc
Linus Torvalds [Fri, 23 Nov 2012 21:55:09 +0000 (11:55 -1000)]
Merge git://git./linux/kernel/git/davem/sparc

Pull sparc fix from David Miller:
 "Bug fix from Al Viro"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
  sparc64: not any error from do_sigaltstack() should fail rt_sigreturn()

11 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Fri, 23 Nov 2012 21:54:22 +0000 (11:54 -1000)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull one more ARM SoC fix from Olof Johansson:
 "I missed one pull request from Samsung with one fix in the previous
  batch.  Here it is -- a dma driver fix for an early version of silicon
  that they still support."

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: EXYNOS: PL330 MDMA1 fix for revision 0 of Exynos4210 SOC

11 years agoPM / QoS: fix wrong error-checking condition
Guennadi Liakhovetski [Fri, 23 Nov 2012 19:55:06 +0000 (20:55 +0100)]
PM / QoS: fix wrong error-checking condition

dev_pm_qos_add_request() can return 0, 1, or a negative error code,
therefore the correct error test is "if (error < 0)." Checking just for
non-zero return code leads to erroneous setting of the req->dev pointer
to NULL, which then leads to a repeated call to
dev_pm_qos_add_ancestor_request() in st1232_ts_irq_handler(). This in turn
leads to an Oops, when the I2C host adapter is unloaded and reloaded again
because of the inconsistent state of its QoS request list.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
11 years agobnx2x: remove redundant warning log
Ariel Elior [Thu, 22 Nov 2012 07:16:17 +0000 (07:16 +0000)]
bnx2x: remove redundant warning log

fix bug where a register which was only meant to be read in 578xx/57712
devices causes a bogus error message to be logged when read from other
devices.

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agovxlan: fix command usage in its doc
Zhi Yong Wu [Thu, 22 Nov 2012 00:10:01 +0000 (00:10 +0000)]
vxlan: fix command usage in its doc

  Some commands don't work in its example doc. The patch will fix it.

Signed-off-by: Zhi Yong Wu <wuzhy@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years ago8139cp: revert "set ring address before enabling receiver"
françois romieu [Wed, 21 Nov 2012 10:07:29 +0000 (10:07 +0000)]
8139cp: revert "set ring address before enabling receiver"

This patch reverts b01af4579ec41f48e9b9c774e70bd6474ad210db.

The original patch was tested with emulated hardware. Real
hardware chokes.

Fixes https://bugzilla.kernel.org/show_bug.cgi?id=47041

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Acked-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMPI: Fix compilation on MIPS with GCC 4.4 and newer
Manuel Lauss [Thu, 22 Nov 2012 10:58:22 +0000 (11:58 +0100)]
MPI: Fix compilation on MIPS with GCC 4.4 and newer

Since 4.4 GCC on MIPS no longer recognizes the "h" constraint,
leading to this build failure:

  CC      lib/mpi/generic_mpih-mul1.o
lib/mpi/generic_mpih-mul1.c: In function 'mpihelp_mul_1':
lib/mpi/generic_mpih-mul1.c:50:3: error: impossible constraint in 'asm'

This patch updates MPI with the latest umul_ppm implementations for MIPS.

Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Cc: Linux-MIPS <linux-mips@linux-mips.org>
Cc: Dmitry Kasatkin <dmitry.kasatkin@intel.com>
Cc: James Morris <jmorris@namei.org>
Patchwork: https://patchwork.linux-mips.org/patch/4612/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
11 years agoMIPS: Fix crash that occurs when function tracing is enabled
Al Cooper [Thu, 15 Nov 2012 23:16:14 +0000 (18:16 -0500)]
MIPS: Fix crash that occurs when function tracing is enabled

A recent patch changed some irq routines from inlines to functions.
These routines are called by the tracer code. Now that they're functions,
if they are compiled for function tracing they will call the tracer
and crash the system due to infinite recursion. The fix disables
tracing in these functions by using "notrace" in the function
definition.

Signed-off-by: Al Cooper <alcooperx@gmail.com>
Reviewed-by: David Daney <david.daney@cavium.com>
Pathchwork: https://patchwork.linux-mips.org/patch/4564/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
11 years agoMIPS: Merge overlapping bootmem ranges
Ralf Baechle [Thu, 15 Nov 2012 11:53:59 +0000 (12:53 +0100)]
MIPS: Merge overlapping bootmem ranges

Without this, we may end up with something like this in /proc/iomem:

01100000-014fffff : System RAM
  01100000-013bf48f : Kernel code
  013bf490-0149e01f : Kernel data
01500000-0c0fffff : System RAM

but the two System RAM ranges should be one single range.  This particular
case will result in kexec failure on Octeon systems if the kernel being
loaded by kexec is bigger than the already running kernel.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
11 years agoblock: Don't access request after it might be freed
Roland Dreier [Thu, 22 Nov 2012 10:00:11 +0000 (02:00 -0800)]
block: Don't access request after it might be freed

After we've done __elv_add_request() and __blk_run_queue() in
blk_execute_rq_nowait(), the request might finish and be freed
immediately.  Therefore checking if the type is REQ_TYPE_PM_RESUME
isn't safe afterwards, because if it isn't, rq might be gone.
Instead, check beforehand and stash the result in a temporary.

This fixes crashes in blk_execute_rq_nowait() I get occasionally when
running with lots of memory debugging options enabled -- I think this
race is usually harmless because the window for rq to be reallocated
is so small.

Signed-off-by: Roland Dreier <roland@purestorage.com>
Cc: stable@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agomtip32xx: Fix padding issue
Selvan Mani [Wed, 14 Nov 2012 13:16:35 +0000 (06:16 -0700)]
mtip32xx: Fix padding issue

Hi Jens,

Another tiny patch.

Removed __packed before the struct smart_attr and added __packed at end of
the structure to fix padding issue.

Signed-off-by: Selvan Mani <smani@micron.com>
Signed-off-by: Asai Thambi S P <asamymuthupa@micron.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agoaoe: avoid running request handler on plugged queue
Ed Cashin [Fri, 9 Nov 2012 00:17:15 +0000 (19:17 -0500)]
aoe: avoid running request handler on plugged queue

Calling the request handler directly on a plugged queue defeats
the performance improvements provided by the plugging mechanism.
Use the __blk_run_queue function instead of calling the request
handler directly, so that we don't interfere with the block
layer's ability to plug the queue.

Signed-off-by: Ed Cashin <ecashin@coraid.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agomtip32xx: fix potential NULL pointer dereference in mtip_timeout_function()
Wei Yongjun [Thu, 8 Nov 2012 09:35:38 +0000 (17:35 +0800)]
mtip32xx: fix potential NULL pointer dereference in mtip_timeout_function()

The dereference to port should be moved below the NULL test.

dpatch engine is used to auto generate this patch.
(https://github.com/weiyj/dpatch)

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agomtip32xx: fix shift larger than type warning
Jens Axboe [Thu, 8 Nov 2012 06:58:53 +0000 (07:58 +0100)]
mtip32xx: fix shift larger than type warning

If we're building a 32-bit kernel and CONFIG_LBADF isn't set,
sector_t is 32-bits wide. The shifts by 32 and 40 are thus
larger than we support.

Cast the sector offset to a u64 to avoid these warnings.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agomtip32xx: Fix incorrect mask used for erase mode
Selvan Mani [Wed, 7 Nov 2012 13:03:56 +0000 (06:03 -0700)]
mtip32xx: Fix incorrect mask used for erase mode

Previous commit use value 3 for erasemode mask.
Changing the mask to correct value to 2

Signed-off-by: Selvan Mani <smani@micron.com>
Signed-off-by: Asai Thambi S P <asamymuthupa@micron.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agomtip32xx: Fix to make lba address correct in big-endian systems
Selvan Mani [Wed, 7 Nov 2012 13:03:53 +0000 (06:03 -0700)]
mtip32xx: Fix to make lba address correct in big-endian systems

Earlier lba address was assigned directly to lba_low and lba_low_ex,
which would result in a different number (bytes reversed) in
big-endian systems. Now assigning lba address byte-by-byte to fis.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Selvan Mani <smani@micron.com>
Signed-off-by: Asai Thambi S P <asamymuthupa@micron.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agomtip32xx: fix potential crash on SEC_ERASE_UNIT
Selvan Mani [Wed, 7 Nov 2012 13:03:37 +0000 (06:03 -0700)]
mtip32xx: fix potential crash on SEC_ERASE_UNIT

The mtip driver lifted this code from elsewhere and then added a special
handling check for SEC_ERASE_UNIT. If the caller tries to do a security
erase but passes no output data for the command then outbuf is not
allocated and the driver duly explodes.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Selvan Mani <smani@micron.com>
Signed-off-by: Asai Thambi S P <asamymuthupa@micron.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agodm: fix deadlock with request based dm and queue request_fn recursion
Jens Axboe [Tue, 6 Nov 2012 11:24:26 +0000 (12:24 +0100)]
dm: fix deadlock with request based dm and queue request_fn recursion

Request based dm attempts to re-run the request queue off the
request completion path. If used with a driver that potentially does
end_io from its request_fn, we could deadlock trying to recurse
back into request dispatch. Fix this by punting the request queue
run to kblockd.

Tested to fix a quickly reproducible deadlock in such a scenario.

Cc: stable@kernel.org
Acked-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agofloppy: destroy floppy workqueue before cleaning up the queue
Jiri Kosina [Tue, 6 Nov 2012 10:47:13 +0000 (11:47 +0100)]
floppy: destroy floppy workqueue before cleaning up the queue

We need to first destroy the floppy_wq workqueue before cleaning up
the queue. Otherwise we might race with still pending work with the
workqueue, but all the block queue already gone. This might lead to
various oopses, such as

 CPU 0
 Pid: 6, comm: kworker/u:0 Not tainted 3.7.0-rc4 #1 Bochs Bochs
 RIP: 0010:[<ffffffff8134eef5>]  [<ffffffff8134eef5>] blk_peek_request+0xd5/0x1c0
 RSP: 0000:ffff88000dc7dd88  EFLAGS: 00010092
 RAX: 0000000000000001 RBX: 0000000000000000 RCX: 0000000000000000
 RDX: ffff88000f602688 RSI: ffffffff81fd95d8 RDI: 6b6b6b6b6b6b6b6b
 RBP: ffff88000dc7dd98 R08: ffffffff81fd95c8 R09: 0000000000000000
 R10: ffffffff81fd9480 R11: 0000000000000001 R12: 6b6b6b6b6b6b6b6b
 R13: ffff88000dc7dfd8 R14: ffff88000dc7dfd8 R15: 0000000000000000
 FS:  0000000000000000(0000) GS:ffffffff81e21000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
 CR2: 0000000000000000 CR3: 0000000001e11000 CR4: 00000000000006f0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
 Process kworker/u:0 (pid: 6, threadinfo ffff88000dc7c000, task ffff88000dc5ecc0)
 Stack:
  0000000000000000 0000000000000000 ffff88000dc7ddb8 ffffffff8134efee
  ffff88000dc7ddb8 0000000000000000 ffff88000dc7dde8 ffffffff814aef3c
  ffffffff81e75d80 ffff88000dc0c640 ffff88000fbfb000 ffffffff814aed90
 Call Trace:
  [<ffffffff8134efee>] blk_fetch_request+0xe/0x30
  [<ffffffff814aef3c>] redo_fd_request+0x1ac/0x400
  [<ffffffff814aed90>] ? start_motor+0x130/0x130
  [<ffffffff8106b526>] process_one_work+0x136/0x450
  [<ffffffff8106af65>] ? manage_workers+0x205/0x2e0
  [<ffffffff8106bb6d>] worker_thread+0x14d/0x420
  [<ffffffff8106ba20>] ? rescuer_thread+0x1a0/0x1a0
  [<ffffffff8107075a>] kthread+0xba/0xc0
  [<ffffffff810706a0>] ? __kthread_parkme+0x80/0x80
  [<ffffffff818b553a>] ret_from_fork+0x7a/0xb0
  [<ffffffff810706a0>] ? __kthread_parkme+0x80/0x80
 Code: 0f 84 c0 00 00 00 83 f8 01 0f 85 e2 00 00 00 81 4b 40 00 00 80 00 48 89 df e8 58 f8 ff ff be fb ff ff ff
 fe ff ff <49> 8b 1c 24 49 39 dc 0f 85 2e ff ff ff 41 0f b6 84 24 28 04 00
 RIP  [<ffffffff8134eef5>] blk_peek_request+0xd5/0x1c0
  RSP <ffff88000dc7dd88>

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Tested-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
11 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Fri, 23 Nov 2012 07:45:34 +0000 (21:45 -1000)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

Pull input updates from Dmitry Torokhov:
 "This fixes recent regression where /dev/input/mice got assigned wrong
  device node which messed up setups with static /dev, and a regression
  in ads7846 GPIO debounce setup."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  ARM - OMAP: ads7846: fix pendown debounce setting
  Input: ads7846 - enable pendown GPIO debounce time setting
  Input: mousedev - move /dev/input/mice to the correct minor
  Input: MT - document new 'flags' argument of input_mt_init_slots()

11 years agoMerge branch 'v3.7-samsung-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Fri, 23 Nov 2012 04:43:20 +0000 (20:43 -0800)]
Merge branch 'v3.7-samsung-fixes-3' of git://git./linux/kernel/git/kgene/linux-samsung into fixes

From Kukjin Kim:

Here is Samsung fixes for v3.7 and it is for fixing of mdma1 address
for exynos4210 rev0 SoC.

* 'v3.7-samsung-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: EXYNOS: PL330 MDMA1 fix for revision 0 of Exynos4210 SOC

Signed-off-by: Olof Johansson <olof@lixom.net>
11 years agopowerpc/pseries: Fix oops with MSIs when missing EEH PEs
Alexey Kardashevskiy [Fri, 23 Nov 2012 02:25:39 +0000 (13:25 +1100)]
powerpc/pseries: Fix oops with MSIs when missing EEH PEs

The new EEH code introduced a small regression, if the EEH PEs
are missin (which happens currently in qemu for example), it
will deref a NULL pointer in the MSI code.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
11 years agoi2c: mxs: Handle i2c DMA failure properly
Marek Vasut [Sun, 18 Nov 2012 05:25:07 +0000 (06:25 +0100)]
i2c: mxs: Handle i2c DMA failure properly

Properly terminate the DMA transfer in case the DMA PIO transfer
or setup fails for any reason.

Signed-off-by: Marek Vasut <marex@denx.de>
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
11 years agoipv4: do not cache looped multicasts
Julian Anastasov [Thu, 22 Nov 2012 21:04:14 +0000 (23:04 +0200)]
ipv4: do not cache looped multicasts

Starting from 3.6 we cache output routes for
multicasts only when using route to 224/4. For local receivers
we can set RTCF_LOCAL flag depending on the membership but
in such case we use maddr and saddr which are not caching
keys as before. Additionally, we can not use same place to
cache routes that differ in RTCF_LOCAL flag value.

Fix it by caching only RTCF_MULTICAST entries
without RTCF_LOCAL (send-only, no loopback). As a side effect,
we avoid unneeded lookup for fnhe when not caching because
multicasts are not redirected and they do not learn PMTU.

Thanks to Maxime Bizon for showing the caching
problems in __mkroute_output for 3.6 kernels: different
RTCF_LOCAL flag in cache can lead to wrong ip_mc_output or
ip_output call and the visible problem is that traffic can
not reach local receivers via loopback.

Reported-by: Maxime Bizon <mbizon@freebox.fr>
Tested-by: Maxime Bizon <mbizon@freebox.fr>
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'master' of git://1984.lsi.us.es/nf
David S. Miller [Thu, 22 Nov 2012 20:27:18 +0000 (15:27 -0500)]
Merge branch 'master' of git://1984.lsi.us.es/nf

Pablo Neira Ayuso says:

====================
The following patchset contains two Netfilter fixes:

* Fix buffer overflow in the name of the timeout policy object
  in the cttimeout infrastructure, from Florian Westphal.

* Fix a bug in the hash set in case that IP ranges are
  specified, from Jozsef Kadlecsik.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec
David S. Miller [Thu, 22 Nov 2012 20:24:30 +0000 (15:24 -0500)]
Merge branch 'master' of git://git./linux/kernel/git/klassert/ipsec

Steffen Klassert says:

====================
This pull request is intended for 3.7 and contains a single patch to
fix the IPsec gc threshold value for ipv4.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 years agoALSA: snd-usb: properly initialize the sync endpoint
Daniel Mack [Thu, 22 Nov 2012 19:27:59 +0000 (20:27 +0100)]
ALSA: snd-usb: properly initialize the sync endpoint

Jeffrey Barish reported an obvious bug in the pcm part of the usb-audio
driver which causes the code to not initialize the sync endpoint from
configure_endpoint().

Reported-by: Jeffrey Barish <jeff_barish@earthlink.net>
Signed-off-by: Daniel Mack <zonque@gmail.com>
Cc: stable@kernel.org [3.5+]
Signed-off-by: Takashi Iwai <tiwai@suse.de>
11 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Thu, 22 Nov 2012 19:22:13 +0000 (09:22 -1000)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Olof Johansson:
 "A few more fixes for final 3.7.  Two dealing with pinmux setup on
  OMAP, and one dealing with TV output on DaVinci.  And one small
  MAINTAINER update."

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: davinci: dm644x: fix out range signal for ED
  ARM: OMAP4: TWL: mux sys_drm_msecure as output for PMIC
  ARM: OMAP3: igep0020: Set WIFI/BT GPIO pins in correct mux mode
  ARM: OMAP: Add maintainer entry for IGEP machines

11 years agoMerge tag 'parisc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/parisc-2.6
Linus Torvalds [Thu, 22 Nov 2012 19:16:29 +0000 (09:16 -1000)]
Merge tag 'parisc-fixes' of git://git./linux/kernel/git/jejb/parisc-2.6

Pull PARISC fixes from James Bottomley:
 "This is two bug fixes: one fixes a loophole where rt_sigprocmask()
  with the wrong values panics the box (Denial of Service) and the other
  fixes an aliasing problem with get_shared_area() which could cause
  data corruption.

Signed-off-by: James Bottomley <JBottomley@Parallels.com>"
* tag 'parisc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/parisc-2.6:
  [PARISC] fix user-triggerable panic on parisc
  [PARISC] fix virtual aliasing issue in get_shared_area()

11 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Thu, 22 Nov 2012 19:14:54 +0000 (09:14 -1000)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is a set of four bug fixes.

  The isci one is an obvious thinko (using request buffer instead of
  response buffer) which causes a command to fail.

  The three others are DIF/DIX updates which are required because
  they're part of a series of ten patches, the other seven of which went
  into the block layer during the merge window meaning our current
  DIF/DIX implementation is broken without these three.

Signed-off-by: James Bottomley <JBottomley@Parallels.com>"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  [SCSI] sd: Implement support for WRITE SAME
  [SCSI] sd: Permit merged discard requests
  [SCSI] Add a report opcode helper
  [SCSI] isci: copy fis 0x34 response into proper buffer

11 years agoMerge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux
Linus Torvalds [Thu, 22 Nov 2012 19:14:24 +0000 (09:14 -1000)]
Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie.

Small fixes for (mostly Nouveau, some radeon) regressions.

* 'drm-fixes' of git://people.freedesktop.org/~airlied/linux:
  drm/nouveau: use the correct fence implementation for nv50
  drm/radeon: add new SI pci id
  radeon: add AGPMode 1 quirk for RV250
  drm/radeon: properly track the crtc not_enabled case evergreen_mc_stop()
  drm/nouveau/bios: fix DCB v1.5 parsing
  drm/nouveau: add missing pll_calc calls
  drm/nouveau: fix crash with noaccel=1
  drm/nv40: allocate ctxprog with kmalloc
  drm/nvc0/disp: fix thinko in vblank regression fix..

11 years agoOMAPDSS: do not fail if dpll4_m4_ck is missing
Aaro Koskinen [Wed, 21 Nov 2012 19:48:51 +0000 (21:48 +0200)]
OMAPDSS: do not fail if dpll4_m4_ck is missing

Do not fail if dpll4_m4_ck is missing. The clock is not there on omap24xx,
so this should not be a hard error.

The patch retains the functionality before the commit 185bae10 (OMAPDSS:
DSS: Cleanup cpu_is_xxxx checks).

Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Tomi Valkeinen <tomi.valkeinen@ti.com>
11 years ago[PARISC] fix user-triggerable panic on parisc
Al Viro [Wed, 21 Nov 2012 19:27:23 +0000 (19:27 +0000)]
[PARISC] fix user-triggerable panic on parisc

int sys32_rt_sigprocmask(int how, compat_sigset_t __user *set, compat_sigset_t __user *oset,
                                    unsigned int sigsetsize)
{
        sigset_t old_set, new_set;
        int ret;

        if (set && get_sigset32(set, &new_set, sigsetsize))

...
static int
get_sigset32(compat_sigset_t __user *up, sigset_t *set, size_t sz)
{
        compat_sigset_t s;
        int r;

        if (sz != sizeof *set) panic("put_sigset32()");

In other words, rt_sigprocmask(69, (void *)69, 69) done by 32bit process
will promptly panic the box.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: <stable@vger.kernel.org>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
11 years agomd/raid10: decrement correct pending counter when writing to replacement.
NeilBrown [Thu, 22 Nov 2012 04:12:09 +0000 (15:12 +1100)]
md/raid10: decrement correct pending counter when writing to replacement.

When a write to a replacement device completes, we carefully
and correctly found the rdev that the write actually went to
and the blithely called rdev_dec_pending on the primary rdev,
even if this write was to the replacement.

This means that any writes to an array while a replacement
was ongoing would cause the nr_pending count for the primary
device to go negative, so it could never be removed.

This bug has been present since replacement was introduced in
3.3, so it is suitable for any -stable kernel since then.

Reported-by: "George Spelvin" <linux@horizon.com>
Cc: stable@vger.kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
11 years agomd/raid10: close race that lose writes lost when replacement completes.
NeilBrown [Thu, 22 Nov 2012 03:42:49 +0000 (14:42 +1100)]
md/raid10: close race that lose writes lost when replacement completes.

When a replacement operation completes there is a small window
when the original device is marked 'faulty' and the replacement
still looks like a replacement.  The faulty should be removed and
the replacement moved in place very quickly, bit it isn't instant.

So the code write out to the array must handle the possibility that
the only working device for some slot in the replacement - but it
doesn't.  If the primary device is faulty it just gives up.  This
can lead to corruption.

So make the code more robust: if either  the primary or the
replacement is present and working, write to them.  Only when
neither are present do we give up.

This bug has been present since replacement was introduced in
3.3, so it is suitable for any -stable kernel since then.

Reported-by: "George Spelvin" <linux@horizon.com>
Cc: stable@vger.kernel.org
Signed-off-by: NeilBrown <neilb@suse.de>
11 years agoMerge branch 'drm-fixes-3.7' of git://people.freedesktop.org/~agd5f/linux into drm...
Dave Airlie [Thu, 22 Nov 2012 03:21:46 +0000 (13:21 +1000)]
Merge branch 'drm-fixes-3.7' of git://people.freedesktop.org/~agd5f/linux into drm-fixes

Alex writes:
A couple more small fixes for 3.7:
- another evergreen_mc fix
- add an AGP quirk for an old RV250
- new pci id.

* 'drm-fixes-3.7' of git://people.freedesktop.org/~agd5f/linux:
  drm/radeon: add new SI pci id
  radeon: add AGPMode 1 quirk for RV250
  drm/radeon: properly track the crtc not_enabled case evergreen_mc_stop()

11 years agoMerge branch 'drm-nouveau-fixes' of git://anongit.freedesktop.org/git/nouveau/linux...
Dave Airlie [Thu, 22 Nov 2012 03:20:45 +0000 (13:20 +1000)]
Merge branch 'drm-nouveau-fixes' of git://anongit.freedesktop.org/git/nouveau/linux-2.6 into drm-fixes

nouveau: one more regression fix.

* 'drm-nouveau-fixes' of git://anongit.freedesktop.org/git/nouveau/linux-2.6:
  drm/nouveau: use the correct fence implementation for nv50

11 years agoMerge branch 'drm-nouveau-fixes' of git://anongit.freedesktop.org/git/nouveau/linux...
Dave Airlie [Mon, 19 Nov 2012 04:03:27 +0000 (14:03 +1000)]
Merge branch 'drm-nouveau-fixes' of git://anongit.freedesktop.org/git/nouveau/linux-2.6 into drm-fixes

Some more misc fallout from nouveau rework.

* 'drm-nouveau-fixes' of git://anongit.freedesktop.org/git/nouveau/linux-2.6:
  drm/nouveau/bios: fix DCB v1.5 parsing
  drm/nouveau: add missing pll_calc calls
  drm/nouveau: fix crash with noaccel=1
  drm/nv40: allocate ctxprog with kmalloc
  drm/nvc0/disp: fix thinko in vblank regression fix..

11 years agodrm/nouveau: use the correct fence implementation for nv50
Maarten Lankhorst [Wed, 21 Nov 2012 12:21:12 +0000 (13:21 +0100)]
drm/nouveau: use the correct fence implementation for nv50

Only compile time tested, noticed nv50_fence_create was never used,
so fix this. This will probably fix vblank on nv50 cards.

Hopefully this is still in time for 3.7 final release.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
11 years agodrm/radeon: add new SI pci id
Alex Deucher [Wed, 21 Nov 2012 23:37:38 +0000 (18:37 -0500)]
drm/radeon: add new SI pci id

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
11 years agonetfilter: cttimeout: fix buffer overflow
Florian Westphal [Wed, 21 Nov 2012 01:37:38 +0000 (01:37 +0000)]
netfilter: cttimeout: fix buffer overflow

Chen Gang reports:
the length of nla_data(cda[CTA_TIMEOUT_NAME]) is not limited in server side.

And indeed, its used to strcpy to a fixed-sized buffer.

Fortunately, nfnetlink users need CAP_NET_ADMIN.

Reported-by: Chen Gang <gang.chen@asianux.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
11 years agonetfilter: ipset: Fix range bug in hash:ip,port,net
Jozsef Kadlecsik [Mon, 19 Nov 2012 06:42:21 +0000 (06:42 +0000)]
netfilter: ipset: Fix range bug in hash:ip,port,net

Due to the missing ininitalization at adding/deleting entries, when
a plain_ip,port,net element was the object, multiple elements were
added/deleted instead. The bug came from the missing dangling
default initialization.

The error-prone default initialization is corrected in all hash:* types.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
11 years agofix incorrect NR_FREE_PAGES accounting (appears like memory leak)
Dave Hansen [Wed, 21 Nov 2012 19:21:51 +0000 (14:21 -0500)]
fix incorrect NR_FREE_PAGES accounting (appears like memory leak)

There have been some 3.7-rc reports of vm issues, including some kswapd
bugs and, more importantly, some memory "leaks":

http://www.spinics.net/lists/linux-mm/msg46187.html
https://bugzilla.kernel.org/show_bug.cgi?id=50181

Commit 1fb3f8ca0e92 ("mm: compaction: capture a suitable high-order page
immediately when it is made available") took split_free_page() and
reused it for the compaction code.  It does something curious with
capture_free_page() (previously known as split_free_page()):

  int capture_free_page(struct page *page, int alloc_order,
  ...
          __mod_zone_page_state(zone, NR_FREE_PAGES, -(1UL << order));

  -       /* Split into individual pages */
  -       set_page_refcounted(page);
  -       split_page(page, order);
  +       if (alloc_order != order)
  +               expand(zone, page, alloc_order, order,
  +                       &zone->free_area[order], migratetype);

Note that expand() puts the pages _back_ in the allocator, but it does
not bump NR_FREE_PAGES.  We "return" 'alloc_order' worth of pages, but
we accounted for removing 'order' in the __mod_zone_page_state() call.

For the old split_page()-style use (order==alloc_order) the bug will not
trigger.  But, when called from the compaction code where we
occasionally get a larger page out of the buddy allocator than we need,
we will run in to this.

This patch simply changes the NR_FREE_PAGES manipulation to the correct
'alloc_order' instead of 'order'.

I've been able to repeatedly trigger this in my testing environment.
The amount "leaked" very closely tracks the imbalance I see in buddy
pages vs.  NR_FREE_PAGES.  I have confirmed that this patch fixes the
imbalance

Signed-off-by: Dave Hansen <dave@linux.vnet.ibm.com>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
11 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Wed, 21 Nov 2012 22:28:19 +0000 (12:28 -1000)]
Merge git://git./linux/kernel/git/davem/net

Pull networking updates from David Miller:

 1) inet6_csk_update_pmtu() must return NULL or non-NULL, so translate
    ERR_PTR to NULL, as needed.  Fix from Eric Dumazet.

 2) Fix copy&paste error in IRDA sir_dev ->set_speed method invocation,
    it was testing the NULL'ness of a different method to guard the
    call.  Fix from Alexander Shiyan.

 3) Fix build regression of xilinx driver, from Jeff Mahoney.

 4) Make XEN netfront (like XEN netback) handle compound pages in SKBs
    properly.  From Ian Campbell.

 5) Fix inverted logic of team_dev_queue_xmit() return value checks,
    from Jiri Pirko and Dan Carpenter.

 6) dma_poll_create() no longer allows a NULL device argument, breaking
    both ixp4xx drivers.  Fix from Xi Wang.

 7) ne2000 driver doesn't hook up the parent device properly, breaking
    udev matching.  Fix from Alan Cox.

 8) Locking and memory leak fixes in Near Field Communications layer.
    From Thierry Escande, Szymon Janc, and Waldemar Rymarkiewicz.

 9) sis900 resume regression, sis900_set_mode() is being called with the
    iomem pointer instead of the expected device private.  Fix from
    Francois Romieu.

10) Fix IBSS regression caused by uninitializing the ibss-internals
    before performing an emptyness check, from Simon WUnderlich.

11) Fix SNIFFER mode regression in iwlwifi driver, from Johannes Berg.

12) Fix task wedges in mwifiex_cmd_timeout_func(), from Bing Zhao.

13) Add back wireless sysfs directory, too much stuff depends upon it
    being there (actually I'd say it never should have been removed to
    begin with).  From Johannes Berg.

14) Fix hang introduced by suspend/resume changes in ath9k.  Fix from
    Sujith Manoharan.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (24 commits)
  team: bcast: convert return value of team_dev_queue_xmit() to bool correctly
  bonding: Bonding driver does not consider the gso_max_size/gso_max_segs setting of slave devices.
  xen/netfront: handle compound page fragments on transmit
  net: fix build failure in xilinx
  irda: sir_dev: Fix copy/paste typo
  ipv6: fix inet6_csk_update_pmtu() return value
  ixp4xx_hss: avoid calling dma_pool_create() with NULL dev
  ixp4xx_eth: avoid calling dma_pool_create() with NULL dev
  ne2000: add the right platform device
  of/net/mdio-gpio: Fix pdev->id issue when using devicetrees.
  NFC: Fix pn533 target mode memory leak
  NFC: pn533: Fix mem leak in pn533_in_dep_link_up
  NFC: pn533: Fix use after free
  NFC: pn533: Fix missing lock while operating on commands list
  NFC: Fix nfc_llcp_local chained list insertion
  ath9k_hw: Fix regression in device reset
  sis900: fix sis900_set_mode call parameters.
  iwlwifi: don't WARN when a non empty queue is disabled
  wireless: add back sysfs directory
  mwifiex: report error to MMC core if we cannot suspend
  ...

11 years agomd/raid5: Make sure we clear R5_Discard when discard is finished.
NeilBrown [Wed, 21 Nov 2012 05:33:40 +0000 (16:33 +1100)]
md/raid5: Make sure we clear R5_Discard when discard is finished.

commit 9e44476851e91c86c98eb92b9bc27fb801f89072
    MD: raid5 avoid unnecessary zero page for trim

change raid5 to clear R5_Discard when the complete request is
handled rather than when submitting the per-device discard request.
However it did not clear R5_Discard for the parity device.

This means that if the stripe_head was reused before it expired from
the cache, the setting would be wrong and a hang would result.

Also if the R5_Uptodate bit happens to be set, R5_Discard again
won't be cleared.  But R5_Uptodate really should be clear at this point.

So make sure R5_Discard is cleared in all cases, and clear
R5_Uptodate when a 'discard' completes.

Signed-off-by: NeilBrown <neilb@suse.de>
11 years agomd/raid5: move resolving of reconstruct_state earlier in
NeilBrown [Wed, 21 Nov 2012 22:13:36 +0000 (09:13 +1100)]
md/raid5: move resolving of reconstruct_state earlier in
stripe_handle.

The chunk of code in stripe_handle which responds to a
*_result value in reconstruct_state is really the completion
of some processing that happened outside of handle_stripe
(possibly asynchronously) and so should be one of the first
things done in handle_stripe().

After the next patch it will be important that it happens before
handle_stripe_clean_event(), as that will clear some dev->flags
bit that this code tests.

Signed-off-by: NeilBrown <neilb@suse.de>