7 years agocrypto: cryptd - Use subsys_initcall to prevent races with aesni
Herbert Xu [Fri, 19 Aug 2011 08:11:23 +0000]
crypto: cryptd - Use subsys_initcall to prevent races with aesni

commit b2bac6acf86d05d8af0499f37d91ecac15722803 upstream.

As cryptd is depeneded on by other algorithms such as aesni-intel,
it needs to be registered before them.  When everything is built
as modules, this occurs naturally.  However, for this to work when
they are built-in, we need to use subsys_initcall in cryptd.

Tested-by: Josh Boyer <jwboyer@redhat.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Kerin Millar <kerframil@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoPM / Suspend: Off by one in pm_suspend()
Dan Carpenter [Wed, 21 Sep 2011 18:55:04 +0000]
PM / Suspend: Off by one in pm_suspend()

commit 528f7ce6e439edeac38f6b3f8561f1be129b5e91 upstream.

In enter_state() we use "state" as an offset for the pm_states[]
array.  The pm_states[] array only has PM_SUSPEND_MAX elements so
this test is off by one.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomtd: pxa3xx_nand: Fix blank page ECC mismatch
Daniel Mack [Tue, 7 Jun 2011 10:01:07 +0000]
mtd: pxa3xx_nand: Fix blank page ECC mismatch

commit 543e32d5ff165d0d68deedb0e3557478c7c36a4a upstream.

This bug was introduced in f8155a40 ("mtd: pxa3xx_nand: rework irq
logic") and causes the PXA3xx NAND controller fail to operate with NAND
flash that has empty pages. According to the comment in this block, the
hardware controller will report a double-bit error for empty pages,
which can and must be ignored.

This patch restores the original behaviour of the driver.

Signed-off-by: Daniel Mack <zonque@gmail.com>
Acked-by: Lei Wen <leiwen@marvell.com>
Cc: Haojian Zhuang <haojian.zhuang@marvell.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomtd: pxa3xx_nand: fix nand detection issue
Lei Wen [Tue, 7 Jun 2011 10:01:06 +0000]
mtd: pxa3xx_nand: fix nand detection issue

commit 0fab028b77d714ad302404b23306cf7adb885223 upstream.

When keep_config is set, the detection would goes different routine.
That the driver would read out the setting which is set previously
by bootloader. While most bootloader keep the irq mask as off, and
current driver need all irq default open, keep_config behavior would
lead to no irq at all.

Signed-off-by: Lei Wen <leiwen@marvell.com>
Tested-by: Daniel Mack <zonque@gmail.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomtd: provide an alias for the redboot module name
Andres Salomon [Fri, 14 Oct 2011 14:33:20 +0000]
mtd: provide an alias for the redboot module name

commit d5de1907d0af22e1a02de2b16a624148517a39c2 upstream.

parse_mtd_partitions takes a list of partition types; if the driver
isn't loaded, it attempts to load it, and then it grabs the partition
parser.  For redboot, the module name is "redboot.ko", while the parser
name is "RedBoot".  Since modprobe is case-sensitive, attempting to
modprobe "RedBoot" will never work.  I suspect the embedded systems that
make use of redboot just always manually loaded redboot prior to loading
their specific nand chip drivers (or statically compiled it in).

Signed-off-by: Andres Salomon <dilinger@queued.net>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomtd: mtdchar: add missing initializer on raw write
Peter Wippich [Mon, 6 Jun 2011 13:50:58 +0000]
mtd: mtdchar: add missing initializer on raw write

commit bf5140817b2d65faac9b32fc9057a097044ac35b upstream.

On writes in MODE_RAW the mtd_oob_ops struct is not sufficiently
initialized which may cause nandwrite to fail. With this patch
it is possible to write raw nand/oob data without additional ECC
(either for testing or when some sectors need different oob layout
e.g. bootloader) like
nandwrite  -n -r -o  /dev/mtd0 <myfile>

Signed-off-by: Peter Wippich <pewi@gw-instruments.de>
Tested-by: Ricard Wanderlof <ricardw@axis.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomac80211: disable powersave for broken APs
Johannes Berg [Fri, 28 Oct 2011 09:59:47 +0000]
mac80211: disable powersave for broken APs

commit 05cb91085760ca378f28fc274fbf77fc4fd9886c upstream.

Only AID values 1-2007 are valid, but some APs have been
found to send random bogus values, in the reported case an
AP that was sending the AID field value 0xffff, an AID of
0x3fff (16383).

There isn't much we can do but disable powersave since
there's no way it can work properly in this case.

Reported-by: Bill C Riemers <briemers@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomac80211: config hw when going back on-channel
Eliad Peller [Thu, 20 Oct 2011 17:05:50 +0000]
mac80211: config hw when going back on-channel

commit 6911bf0453e0d6ea8eb694a4ce67a68d071c538e upstream.

When going back on-channel, we should reconfigure
the hw iff the hardware is not already configured
to the operational channel.

Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomac80211: fix remain_off_channel regression
Eliad Peller [Thu, 20 Oct 2011 17:05:49 +0000]
mac80211: fix remain_off_channel regression

commit eaa7af2ae582c9a8c51b374c48d5970b748a5ce2 upstream.

The offchannel code is currently broken - we should
remain_off_channel if the work was started, and
the work's channel and channel_type are the same
as local->tmp_channel and local->tmp_channel_type.

However, if wk->chan_type and local->tmp_channel_type
coexist (e.g. have the same channel type), we won't
remain_off_channel.

This behavior was introduced by commit da2fd1f
("mac80211: Allow work items to use existing
channel type.")

Tested-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoath9k_hw: Update AR9485 initvals to fix system hang issue
Rajkumar Manoharan [Mon, 24 Oct 2011 12:43:40 +0000]
ath9k_hw: Update AR9485 initvals to fix system hang issue

commit 98fb2cc115b4ef1ea0a2d87a170c183bd395dd6c upstream.

This patch fixes system hang when resuming from S3 state
and lower rate sens failure issue.

Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agonetlink: validate NLA_MSECS length
Johannes Berg [Thu, 3 Nov 2011 00:07:32 +0000]
netlink: validate NLA_MSECS length

commit c30bc94758ae2a38a5eb31767c1985c0aae0950b upstream.

L2TP for example uses NLA_MSECS like this:
policy:
        [L2TP_ATTR_RECV_TIMEOUT]        = { .type = NLA_MSECS, },
code:
        if (info->attrs[L2TP_ATTR_RECV_TIMEOUT])
                cfg.reorder_timeout = nla_get_msecs(info->attrs[L2TP_ATTR_RECV_TIMEOUT]);

As nla_get_msecs() is essentially nla_get_u64() plus the
conversion to a HZ-based value, this will not properly
reject attributes from userspace that aren't long enough
and might overrun the message.

Add NLA_MSECS to the attribute minlen array to check the
size properly.

Cc: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoACPI: Fix CONFIG_ACPI_DOCK=n compiler warning
Bart Van Assche [Wed, 2 Nov 2011 19:16:05 +0000]
ACPI: Fix CONFIG_ACPI_DOCK=n compiler warning

commit c1056b42a87b59375f8f81a92ef029165f44fcce upstream.

Recently the ACPI ops structs were constified but the inline version
of register_hotplug_dock_device() was overlooked (see also commit
9c8b04b, June 25 2011). Update the inline function
register_hotplug_dock_device() that is enabled with
CONFIG_ACPI_DOCK=n too. This patch fixes at least the following
compiler warnings:

drivers/ata/libata-acpi.c: In function .ata_acpi_associate.:
drivers/ata/libata-acpi.c:266:11: warning: passing argument 2 of .register_hotplug_dock_device. discards qualifiers from pointer target type
include/acpi/acpi_drivers.h:146:19: note: expected .struct acpi_dock_ops *. but argument is of type .const struct acpi_dock_ops *.
drivers/ata/libata-acpi.c:275:11: warning: passing argument 2 of .register_hotplug_dock_device. discards qualifiers from pointer target type
include/acpi/acpi_drivers.h:146:19: note: expected .struct acpi_dock_ops *. but argument is of type .const struct acpi_dock_ops *.

Signed-off-by: Len Brown <len.brown@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoACPI atomicio: Convert width in bits to bytes in __acpi_ioremap_fast()
Luck, Tony [Fri, 21 Oct 2011 21:42:55 +0000]
ACPI atomicio: Convert width in bits to bytes in __acpi_ioremap_fast()

commit 3bf3f8b19d2bfccc40f13c456bf339fd8f535ebc upstream.

Callers to __acpi_ioremap_fast() pass the bit_width that they found in the
acpi_generic_address structure. Convert from bits to bytes when passing to
__acpi_find_iomap() - as it wants to see bytes, not bits.

Signed-off-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: Fix deadlock in icswx code
Anton Blanchard [Wed, 14 Sep 2011 09:43:15 +0000]
powerpc: Fix deadlock in icswx code

commit 8bdafa39a47265bc029838b35cc6585f69224afa upstream.

The icswx code introduced an A-B B-A deadlock:

     CPU0                    CPU1
     ----                    ----
lock(&anon_vma->mutex);
                             lock(&mm->mmap_sem);
                             lock(&anon_vma->mutex);
lock(&mm->mmap_sem);

Instead of using the mmap_sem to keep mm_users constant, take the
page table spinlock.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc/eeh: Fix /proc/ppc64/eeh creation
Thadeu Lima de Souza Cascardo [Fri, 26 Aug 2011 10:36:31 +0000]
powerpc/eeh: Fix /proc/ppc64/eeh creation

commit 8feaa43494cee5e938fd5a57b9e9bf1c827e6ccd upstream.

Since commit 188917e183cf9ad0374b571006d0fc6d48a7f447, /proc/ppc64 is a
symlink to /proc/powerpc/. That means that creating /proc/ppc64/eeh will
end up with a unaccessible file, that is not listed under /proc/powerpc/
and, then, not listed under /proc/ppc64/.

Creating /proc/powerpc/eeh fixes that problem and maintain the
compatibility intended with the ppc64 symlink.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc/pseries: Avoid spurious error during hotplug CPU add
Anton Blanchard [Sun, 14 Aug 2011 14:30:30 +0000]
powerpc/pseries: Avoid spurious error during hotplug CPU add

commit 9c740025c51a26ab00192cfc464064d4ccbfe3fc upstream.

During hotplug CPU add we get the following error:

Unexpected Error (0) returned from configure-connector

ibm,configure-connector returns 0 for configuration complete, so
catch this and avoid the error.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: Fix oops when echoing bad values to /sys/devices/system/memory/probe
Anton Blanchard [Wed, 10 Aug 2011 20:44:24 +0000]
powerpc: Fix oops when echoing bad values to /sys/devices/system/memory/probe

commit a11940978bd598e65996b4f807cf4904793f7025 upstream.

If we echo an address the hypervisor doesn't like to
/sys/devices/system/memory/probe we oops the box:

# echo 0x10000000000 > /sys/devices/system/memory/probe

kernel BUG at arch/powerpc/mm/hash_utils_64.c:541!

The backtrace is:

create_section_mapping
arch_add_memory
add_memory
memory_probe_store
sysdev_class_store
sysfs_write_file
vfs_write
SyS_write

In create_section_mapping we BUG if htab_bolt_mapping returned
an error. A better approach is to return an error which will
propagate back to userspace.

Rerunning the test with this patch applied:

# echo 0x10000000000 > /sys/devices/system/memory/probe
-bash: echo: write error: Invalid argument

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc/numa: Remove double of_node_put in hot_add_node_scn_to_nid
Anton Blanchard [Wed, 10 Aug 2011 20:44:21 +0000]
powerpc/numa: Remove double of_node_put in hot_add_node_scn_to_nid

commit 6083184269fd723affca4f6340e491950267622a upstream.

During memory hotplug testing, I got the following warning:

ERROR: Bad of_node_put() on /memory@0

of_node_release
kref_put
of_node_put
of_find_node_by_type
hot_add_node_scn_to_nid
hot_add_scn_to_nid
memory_add_physaddr_to_nid
...

of_find_node_by_type() loop does the of_node_put for us so we only
need the handle the case where we terminate the loop early.

As suggested by Stephen Rothwell we can do the of_node_put
unconditionally outside of the loop since of_node_put handles a
NULL argument fine.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agotarget: Fix wrong se_tmr being added to drain_tmr_list
Joern Engel [Wed, 26 Oct 2011 20:37:56 +0000]
target: Fix wrong se_tmr being added to drain_tmr_list

commit 6eb40b2af4908e9aee71e43e7a384243128c56dd upstream.

This patch fixes another bug from LUN_RESET re-org fallout in
core_tmr_drain_tmr_list() that was adding the wrong se_tmr_req
into the local drain_tmr_list to be walked + released.

Signed-off-by: Joern Engel <joern@logfs.org>
Cc: Joern Engel <joern@logfs.org>
Reviewed-by: Roland Dreier <roland@purestorage.com>
Cc: Roland Dreier <roland@purestorage.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agotarget: Fix incorrect se_cmd assignment in core_tmr_drain_tmr_list
Joern Engel [Wed, 26 Oct 2011 05:08:43 +0000]
target: Fix incorrect se_cmd assignment in core_tmr_drain_tmr_list

commit 80ccbc8e00f7001d79dd503c2781487906b98611 upstream.

This patch fixes a bug in core_tmr_drain_tmr_list() where drain_tmr_list
was using the wrong se_tmr_req for cmd assignment due to a typo during the
LUN_RESET re-org.  This was resulting in general protection faults while
using the leftover bogus *tmr_p pointer from list_for_each_entry_safe().

Signed-off-by: Joern Engel <joern@logfs.org>
Cc: Joern Engel <joern@logfs.org>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoVFS: we need to set LOOKUP_JUMPED on mountpoint crossing
Al Viro [Mon, 7 Nov 2011 21:21:26 +0000]
VFS: we need to set LOOKUP_JUMPED on mountpoint crossing

commit a3fbbde70a0cec017f2431e8f8de208708c76acc upstream.

Mountpoint crossing is similar to following procfs symlinks - we do
not get ->d_revalidate() called for dentry we have arrived at, with
unpleasant consequences for NFS4.

Simple way to reproduce the problem in mainline:

    cat >/tmp/a.c <<'EOF'
    #include <unistd.h>
    #include <fcntl.h>
    #include <stdio.h>
    main()
    {
            struct flock fl = {.l_type = F_RDLCK, .l_whence = SEEK_SET, .l_len = 1};
            if (fcntl(0, F_SETLK, &fl))
                    perror("setlk");
    }
    EOF
cc /tmp/a.c -o /tmp/test

then on nfs4:

    mount --bind file1 file2
    /tmp/test < file1 # ok
    /tmp/test < file2 # spews "setlk: No locks available"...

What happens is the missing call of ->d_revalidate() after mountpoint
crossing and that's where NFS4 would issue OPEN request to server.

The fix is simple - treat mountpoint crossing the same way we deal with
following procfs-style symlinks.  I.e.  set LOOKUP_JUMPED...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoblk-flush: move the queue kick into
Jeff Moyer [Mon, 17 Oct 2011 10:57:23 +0000]
blk-flush: move the queue kick into

commit e67b77c791ca2778198c9e7088f3266ed2da7a55 upstream.

A dm-multipath user reported[1] a problem when trying to boot
a kernel with commit 4853abaae7e4a2af938115ce9071ef8684fb7af4
(block: fix flush machinery for stacking drivers with differring
flush flags) applied.  It turns out that an empty flush request
can be sent into blk_insert_flush.  When the BUG_ON was fixed
to allow for this, I/O on the underlying device would stall.  The
reason is that blk_insert_cloned_request does not kick the queue.
In the aforementioned commit, I had added a special case to
kick the queue if data was sent down but the queue flags did
not require a flush.  A better solution is to push the queue
kick up into blk_insert_cloned_request.

This patch, along with a follow-on which fixes the BUG_ON, fixes
the issue reported.

[1] http://www.redhat.com/archives/dm-devel/2011-September/msg00154.html

Reported-by: Christophe Saout <christophe@saout.de>
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoblk-flush: fix invalid BUG_ON in blk_insert_flush
Jeff Moyer [Mon, 17 Oct 2011 10:57:22 +0000]
blk-flush: fix invalid BUG_ON in blk_insert_flush

commit 834f9f61a525d2f6d3d0c93894e26326c8d3ceed upstream.

A user reported a regression due to commit
4853abaae7e4a2af938115ce9071ef8684fb7af4 (block: fix flush
machinery for stacking drivers with differring flush flags).
Part of the problem is that blk_insert_flush required a
single bio be attached to the request.  In reality, having
no attached bio is also a valid case, as can be observed with
an empty flush.

[1] http://www.redhat.com/archives/dm-devel/2011-September/msg00154.html

Reported-by: Christophe Saout <christophe@saout.de>
Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agohpsa: add small delay when using PCI Power Management to reset for kump
Mike Miller [Fri, 21 Oct 2011 06:19:43 +0000]
hpsa: add small delay when using PCI Power Management to reset for kump

commit c4853efec665134b2e6fc9c13447323240980351 upstream.

The P600 requires a small delay when changing states. Otherwise we may think
the board did not reset and we bail. This for kdump only and is particular
to the P600.

Signed-off-by: Mike Miller <mike.miller@hp.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoVFS: fix statfs() automounter semantics regression
Dan McGee [Tue, 1 Nov 2011 23:23:10 +0000]
VFS: fix statfs() automounter semantics regression

commit 5c8a0fbba543d9428a486f0d1282bbcf3cf1d95a upstream.

No one in their right mind would expect statfs() to not work on a
automounter managed mount point. Fix it.

[ I'm not sure about the "no one in their right mind" part.  It's not
  mounted, and you didn't ask for it to be mounted.  But nobody will
  really care, and this probably makes it match previous semantics, so..
      - Linus ]

This mirrors the fix made to the quota code in 815d405ceff0d69646.

Signed-off-by: Dan McGee <dpmcgee@gmail.com>
Cc: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoxen/blkback: Report VBD_WSECT (wr_sect) properly.
Konrad Rzeszutek Wilk [Mon, 10 Oct 2011 16:33:21 +0000]
xen/blkback: Report VBD_WSECT (wr_sect) properly.

commit 5c62cb48602dba95159c81ffeca179d3852e25be upstream.

We did not increment the amount of sectors written to disk
b/c we tested for the == WRITE which is incorrect - as the
operations are more of WRITE_FLUSH, WRITE_ODIRECT. This patch
fixes it by doing a & WRITE check.

Reported-by: Andy Burns <xen.lists@burns.me.uk>
Suggested-by: Ian Campbell <Ian.Campbell@citrix.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoblock: make gendisk hold a reference to its queue
Tejun Heo [Mon, 17 Oct 2011 11:42:43 +0000]
block: make gendisk hold a reference to its queue

commit f992ae801a7dec34a4ed99a6598bbbbfb82af4fb upstream.

The following command sequence triggers an oops.

# mount /dev/sdb1 /mnt
# echo 1 > /sys/class/scsi_device/0\:0\:1\:0/device/delete
# umount /mnt

 general protection fault: 0000 [#1] PREEMPT SMP
 CPU 2
 Modules linked in:

 Pid: 791, comm: umount Not tainted 3.1.0-rc3-work+ #8 Bochs Bochs
 RIP: 0010:[<ffffffff810d0879>]  [<ffffffff810d0879>] __lock_acquire+0x389/0x1d60
...
 Call Trace:
  [<ffffffff810d2845>] lock_acquire+0x95/0x140
  [<ffffffff81aed87b>] _raw_spin_lock+0x3b/0x50
  [<ffffffff811573bc>] bdi_lock_two+0x5c/0x70
  [<ffffffff811c2f6c>] bdev_inode_switch_bdi+0x4c/0xf0
  [<ffffffff811c3fcb>] __blkdev_put+0x11b/0x1d0
  [<ffffffff811c4010>] __blkdev_put+0x160/0x1d0
  [<ffffffff811c40df>] blkdev_put+0x5f/0x190
  [<ffffffff8118f18d>] kill_block_super+0x4d/0x80
  [<ffffffff8118f4a5>] deactivate_locked_super+0x45/0x70
  [<ffffffff8119003a>] deactivate_super+0x4a/0x70
  [<ffffffff811ac4ad>] mntput_no_expire+0xed/0x130
  [<ffffffff811acf2e>] sys_umount+0x7e/0x3a0
  [<ffffffff81aeeeab>] system_call_fastpath+0x16/0x1b

This is because bdev holds on to disk but disk doesn't pin the
associated queue.  If a SCSI device is removed while the device is
still open, the sdev puts the base reference to the queue on release.
When the bdev is finally released, the associated queue is already
gone along with the bdi and bdev_inode_switch_bdi() ends up
dereferencing already freed bdi.

Even if it were not for this bug, disk not holding onto the associated
queue is very unusual and error-prone.

Fix it by making add_disk() take an extra reference to its queue and
put it on disk_release() and ensuring that disk and its fops owner are
put in that order after all accesses to the disk and queue are
complete.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoNFS/sunrpc: don't use a credential with extra groups.
NeilBrown [Mon, 24 Oct 2011 23:25:49 +0000]
NFS/sunrpc: don't use a credential with extra groups.

commit dc6f55e9f8dac4b6479be67c5c9128ad37bb491f upstream.

The sunrpc layer keeps a cache of recently used credentials and
'unx_match' is used to find the credential which matches the current
process.

However unx_match allows a match when the cached credential has extra
groups at the end of uc_gids list which are not in the process group list.

So if a process with a list of (say) 4 group accesses a file and gains
access because of the last group in the list, then another process
with the same uid and gid, and a gid list being the first tree of the
gids of the original process tries to access the file, it will be
granted access even though it shouldn't as the wrong rpc credential
will be used.

Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoASoC: Ensure we always delay for WM8962 FLL when starting from SYSCLK
Mark Brown [Tue, 1 Nov 2011 13:53:54 +0000]
ASoC: Ensure we always delay for WM8962 FLL when starting from SYSCLK

commit 0b7dd6ad92b6cace35dc5d06d6e236c2751c85dc upstream.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoASoC: Ensure the WM8962 oscillator and PLLs start up disabled
Mark Brown [Tue, 1 Nov 2011 13:53:37 +0000]
ASoC: Ensure the WM8962 oscillator and PLLs start up disabled

commit 2af8de8c39cf58e5a5e40a9d5d71332da98e6ba7 upstream.

Since there is no current software control for these they would otherwise
be left enabled, consuming power.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoASoC: Ensure WM8962 PLL registers are reset
Mark Brown [Tue, 1 Nov 2011 13:36:10 +0000]
ASoC: Ensure WM8962 PLL registers are reset

commit 4f4488abc97c1c27ff029f887944e6a6da1f5733 upstream.

The WM8962 has a separate software reset for the PLL registers. Ensure that
these are reset also on startup.

Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoASoC: WM8904: Set `invert' bit for Capture Switch
Hong Xu [Fri, 28 Oct 2011 07:36:39 +0000]
ASoC: WM8904: Set `invert' bit for Capture Switch

commit 5a7c5f26df3c0122814dfa1c13ef6dfbdbffdb86 upstream.

Set `invert' bit for Capture Switch. Otherwise analogue is muted when
Capture Switch is ON.

Signed-off-by: Hong Xu <hong.xu@atmel.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoASoC: Leave input audio data bit length settings untouched in wm8711_set_dai_fmt
Axel Lin [Fri, 28 Oct 2011 07:17:56 +0000]
ASoC: Leave input audio data bit length settings untouched in wm8711_set_dai_fmt

commit d558cfc30064a97c2c65dbd2b3a4f5a1dea7ec1b upstream.

Current implementation in wm8711_set_dai_fmt always clear BIT[3:2]
(the Input Audio Data Bit Length Select) of WM8711_IFACE(07h) register.
Input Audio Data Bit Length Select bits are set by wm8711_hw_params,
we should leave BIT[3:2] untouched in wm8711_set_dai_fmt.

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoASoC: wm8711: Fix wrong mask for setting input audio data bit length select
Axel Lin [Fri, 28 Oct 2011 03:08:01 +0000]
ASoC: wm8711: Fix wrong mask for setting input audio data bit length select

commit 04c57163c8edfbc50e022737014069998ba4fc5f upstream.

The Input Audio Data Bit Length Select is controlled by BIT[3:2] of
WM8711_IFACE(07h) register.
Current code incorrectly masks BIT[1:0] which is for Audio Data Format Select.

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agompt2sas: Fix for system hang when discovery in progress
nagalakshmi.nandigama@lsi.com [Fri, 21 Oct 2011 04:36:33 +0000]
mpt2sas: Fix for system hang when discovery in progress

commit 0167ac67ff6f35bf2364f7672c8012b0cd40277f upstream.

Fix for issue : While discovery is in progress, hot unplug and hot plug of
enclosure connected to the controller card is causing system to hang.

When a device is in the process of being detected at driver load time then
if it is removed, the device that is no longer present will not be added
to the list. So the code in _scsih_probe_sas() is rearranged as such so
the devices that failed to be detected are not added to the list.

Signed-off-by: Nagalakshmi Nandigama <nagalakshmi.nandigama@lsi.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoFix block queue and elevator memory leak in scsi_alloc_sdev
Anton Blanchard [Wed, 2 Nov 2011 21:56:22 +0000]
Fix block queue and elevator memory leak in scsi_alloc_sdev

commit f7c9c6bb14f3104608a3a83cadea10a6943d2804 upstream.

When looking at memory consumption issues I noticed quite a
lot of memory in the kmalloc-2048 bucket:

  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
  6561   6471  98%    2.30K    243       27     15552K kmalloc-2048

Over 15MB. slub debug shows that cfq is responsible for almost
all of it:

# sort -nr /sys/kernel/slab/kmalloc-2048/alloc_calls
6402 .cfq_init_queue+0xec/0x460 age=43423/43564/43655 pid=1 cpus=4,11,13

In scsi_alloc_sdev we do scsi_alloc_queue but if slave_alloc
fails we don't free it with scsi_free_queue.

The patch below fixes the issue:

  OBJS ACTIVE  USE OBJ SIZE  SLABS OBJ/SLAB CACHE SIZE NAME
   135     72  53%    2.30K      5       27       320K kmalloc-2048

# cat /sys/kernel/slab/kmalloc-2048/alloc_calls
3 .cfq_init_queue+0xec/0x460 age=3811/3876/3925 pid=1 cpus=4,11,13

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoMake scsi_free_queue() kill pending SCSI commands
Bart Van Assche [Fri, 23 Sep 2011 17:48:18 +0000]
Make scsi_free_queue() kill pending SCSI commands

commit 3308511c93e6ad0d3c58984ecd6e5e57f96b12c8 upstream.

Make sure that SCSI device removal via scsi_remove_host() does finish
all pending SCSI commands. Currently that's not the case and hence
removal of a SCSI host during I/O can cause a deadlock. See also
"blkdev_issue_discard() hangs forever if underlying storage device is
removed" (http://bugzilla.kernel.org/show_bug.cgi?id=40472). See also
http://lkml.org/lkml/2011/8/27/6.

Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoscsi_dh: check queuedata pointer before proceeding further
Moger, Babu [Wed, 26 Oct 2011 18:29:38 +0000]
scsi_dh: check queuedata pointer before proceeding further

commit a18a920c70d48a8e4a2b750d8a183b3c1a4be514 upstream.

This patch validates sdev pointer in scsi_dh_activate before proceeding further.

Without this check we might see the panic as below. I have seen this
panic multiple times..

Call trace:

 #0 [ffff88007d647b50] machine_kexec at ffffffff81020902
 #1 [ffff88007d647ba0] crash_kexec at ffffffff810875b0
 #2 [ffff88007d647c70] oops_end at ffffffff8139c650
 #3 [ffff88007d647c90] __bad_area_nosemaphore at ffffffff8102dd15
 #4 [ffff88007d647d50] page_fault at ffffffff8139b8cf
    [exception RIP: scsi_dh_activate+0x82]
    RIP: ffffffffa0041922  RSP: ffff88007d647e00  RFLAGS: 00010046
    RAX: 0000000000000000  RBX: 0000000000000000  RCX: 00000000000093c5
    RDX: 00000000000093c5  RSI: ffffffffa02e6640  RDI: ffff88007cc88988
    RBP: 000000000000000f   R8: ffff88007d646000   R9: 0000000000000000
    R10: ffff880082293790  R11: 00000000ffffffff  R12: ffff88007cc88988
    R13: 0000000000000000  R14: 0000000000000286  R15: ffff880037b845e0
    ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0000
 #5 [ffff88007d647e38] run_workqueue at ffffffff81060268
 #6 [ffff88007d647e78] worker_thread at ffffffff81060386
 #7 [ffff88007d647ee8] kthread at ffffffff81064436
 #8 [ffff88007d647f48] kernel_thread at ffffffff81003fba

Signed-off-by: Babu Moger <babu.moger@netapp.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agost: fix race in st_scsi_execute_end
Petr Uzel [Fri, 21 Oct 2011 11:31:09 +0000]
st: fix race in st_scsi_execute_end

commit c68bf8eeaa57c852e74adcf597237be149eef830 upstream.

The call to complete() in st_scsi_execute_end() wakes up sleeping thread
in write_behind_check(), which frees the st_request, thus invalidating
the pointer to the associated bio structure, which is then passed to the
blk_rq_unmap_user(). Fix by storing pointer to bio structure into
temporary local variable.

This bug is present since at least linux-2.6.32.

Signed-off-by: Petr Uzel <petr.uzel@suse.cz>
Reported-by: Juergen Groß <juergen.gross@ts.fujitsu.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Acked-by: Kai Mäkisara <kai.makisara@kolumbus.fi>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoiscsi-target: Fix non-immediate TMR handling
Nicholas Bellinger [Fri, 28 Oct 2011 20:37:19 +0000]
iscsi-target: Fix non-immediate TMR handling

commit 5a4c8666c6d576f076a7c6824589cdbb984c0f84 upstream.

This patch addresses two issues with non immediate TMR handling in
iscsit_handle_task_mgt_cmd().  The first involves breakage due to
v3.1-rc conversion of iscsit_sequence_cmd(), which upon good status
would hit the iscsit_add_reject_from_cmd() block of code.  This patch
adds an explict check for CMDSN_ERROR_CANNOT_RECOVER.

The second adds a check to return when non immediate TMR operation is
detected after iscsit_ack_from_expstatsn(), as iscsit_sequence_cmd()
-> iscsit_execute_cmd() will have called transport_generic_handle_tmr()
for the non immediate TMR case already.

Cc: Andy Grover <agrover@redhat.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoiscsi-target: Add missing CMDSN_LOWER_THAN_EXP check in iscsit_handle_scsi_cmd
Nicholas Bellinger [Fri, 28 Oct 2011 20:32:35 +0000]
iscsi-target: Add missing CMDSN_LOWER_THAN_EXP check in iscsit_handle_scsi_cmd

commit 7e32da55e26b80d57961681e53aa67938268db3f upstream.

This patch adds a missing CMDSN_LOWER_THAN_EXP return check for
iscsit_sequence_cmd() in iscsit_handle_scsi_cmd() that was incorrectly
dropped during the v3.1-rc cleanups to use iscsit_sequence_cmd().

Cc: Andy Grover <agrover@redhat.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agotcm_loop: Add explict read buffer memset for SCF_SCSI_CONTROL_SG_IO_CDB
Nicholas Bellinger [Mon, 24 Oct 2011 20:35:37 +0000]
tcm_loop: Add explict read buffer memset for SCF_SCSI_CONTROL_SG_IO_CDB

commit 8cd79f24350826b81e16990d9e12bc878e67d385 upstream.

This patch addresses an issue with buggy userspace code sending I/O
via scsi-generic that does not explictly clear their associated read
buffers.  It adds an explict memset of the first SGL entry within
tcm_loop_new_cmd_map() for SCF_SCSI_CONTROL_SG_IO_CDB payloads that
are currently guaranteed to be a single SGL by target-core code.

This issue is a side effect of the v3.1-rc1 merge to remove the
extra memcpy between certain control CDB types using a contigious
+ cleared buffer in target-core, and performing a memcpy into the
SGL list within tcm_loop.

It was originally mainfesting itself by udev + scsi_id + scsi-generic
not properly setting up the expected /dev/disk/by-id/ symlinks because
the INQUIRY payload was containing extra bogus data preventing the
proper NAA IEEE WWN from being parsed by userspace.

Cc: Christoph Hellwig <hch@lst.de>
Cc: Andy Grover <agrover@redhat.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agohwmon: (w83627ehf) Fix broken driver init
Guenter Roeck [Sun, 6 Nov 2011 19:25:18 +0000]
hwmon: (w83627ehf) Fix broken driver init

commit bfa02b0da66965caf46e441270af87edda4fea14 upstream.

Commit 2265cef2 (hwmon: (w83627ehf) Properly report PECI and AMD-SI
sensor types) results in kernel panic if data->temp_label was not
initialized.
The problem was found with chip W83627DHG-P.

Add check if data->temp->label was set before use.

Based on incomplete patch by Alexander Beregalov.

Reported-by: Alexander Beregalov <a.beregalov@gmail.com>
Tested-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agohwmon: (w83627ehf) Properly report PECI and AMD-SI sensor types
Jean Delvare [Fri, 4 Nov 2011 11:00:47 +0000]
hwmon: (w83627ehf) Properly report PECI and AMD-SI sensor types

commit 2265cef2751b3441df91f85e0107f9f549e5b711 upstream.

When temperature sources are PECI or AMD-SI agents, it makes no sense
to report their type as diode or thermistor. Instead we must report
their digital nature.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agohwmon: (coretemp) Fix for non-SMP builds
Jean Delvare [Sun, 6 Nov 2011 19:25:18 +0000]
hwmon: (coretemp) Fix for non-SMP builds

commit 2aba6cac2a84f3b80e11a680c34d55e7739b474d upstream.

The definition of TO_ATTR_NO in the non-SMP case is wrong. As the SMP
definition resolves to the correct value, just use this for both
cases.

Without this fix the temperature attributes are named temp0_* instead
of temp2_*, so libsensors won't pick them. Broken since kernel 3.0.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Tested-by: Phil Sutter <phil@nwl.cc>
Acked-by: Durgadoss R <Durgadoss.r@intel.com>
Acked-by: Guenter Roeck <guenter.roeck@ericsson.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agocciss: add small delay when using PCI Power Management to reset for kump
Mike Miller [Thu, 20 Oct 2011 20:19:17 +0000]
cciss: add small delay when using PCI Power Management to reset for kump

commit ab5dbebe33e0c353e8545f09c34553ac3351dad6 upstream.

The P600 requires a small delay when changing states. Otherwise we may think
the board did not reset and we bail. This for kdump only and is particular
to the P600.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoUSB: Update last_busy time after autosuspend fails
Alan Stern [Thu, 3 Nov 2011 23:52:46 +0000]
USB: Update last_busy time after autosuspend fails

commit b2c0a863e14676fa5760c6d828fd373288e2f64a upstream.

Originally, the runtime PM core would send an idle notification
whenever a suspend attempt failed.  The idle callback routine could
then schedule a delayed suspend for some time later.

However this behavior was changed by commit
f71648d73c1650b8b4aceb3856bebbde6daa3b86 (PM / Runtime: Remove idle
notification after failing suspend).  No notifications were sent, and
there was no clear mechanism to retry failed suspends.

This caused problems for the usbhid driver, because it fails
autosuspend attempts as long as a key is being held down.  A companion
patch changes the PM core's behavior, but we also need to change the
USB core.  In particular, this patch (as1493) updates the device's
last_busy time when an autosuspend fails, so that the PM core will
retry the autosuspend in the future when the delay time expires
again.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Henrik Rydberg <rydberg@euromail.se>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoPM / Runtime: Automatically retry failed autosuspends
Alan Stern [Thu, 3 Nov 2011 22:39:18 +0000]
PM / Runtime: Automatically retry failed autosuspends

commit 886486b792e4f6f96d4fbe8ec5bf20811cab7d6a upstream.

Originally, the runtime PM core would send an idle notification
whenever a suspend attempt failed.  The idle callback routine could
then schedule a delayed suspend for some time later.

However this behavior was changed by commit
f71648d73c1650b8b4aceb3856bebbde6daa3b86 (PM / Runtime: Remove idle
notification after failing suspend).  No notifications were sent, and
there was no clear mechanism to retry failed suspends.

This caused problems for the usbhid driver, because it fails
autosuspend attempts as long as a key is being held down.  Therefore
this patch (as1492) adds a mechanism for retrying failed
autosuspends.  If the callback routine updates the last_busy field so
that the next autosuspend expiration time is in the future, the
autosuspend will automatically be rescheduled.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Henrik Rydberg <rydberg@euromail.se>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agokbuild: Fix help text not displayed in choice option.
Srinivas Kandagatla [Tue, 2 Aug 2011 17:49:52 +0000]
kbuild: Fix help text not displayed in choice option.

commit 3f198dfee49d2e9c30583c62b0c79286c78c7b44 upstream.

Help text under choice menu is never displayed because it does not have
symbol name associated with it, however many kconfigs have help text
under choice, assuming that it will be displayed when user selects help.
for example in Kconfig if we have:
choice
        prompt "Choice"
        ---help---
           HELP TEXT ...

config A
        bool "A"

config B
        bool "B"

endchoice

Without this patch "HELP TEXT" is not displayed when user selects help
option when "Choice" is highlighted from menuconfig or xconfig or
gconfig.

This patch changes the logic in menu_get_ext_help to display help for
cases which dont have symbol names like choice.

Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@st.com>
Reviewed-by: Stuart Menefy <stuart.menefy@st.com>
Reviewed-by: Arnaud Lacombe <lacombar@gmail.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon/kms: set HPD polarity in hpd_init()
Alex Deucher [Thu, 3 Nov 2011 15:21:39 +0000]
drm/radeon/kms: set HPD polarity in hpd_init()

commit 64912e997f0fe13512e4c7b90e4f7c11cb922ab5 upstream.

Polarity needs to be set accordingly to connector status (connected
or disconnected). Set it up in hpd_init() so first hotplug works
reliably no matter what is the initial set of connector. hpd_init()
also covers resume so HPD will work correctly after resume as well.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Jerome Glisse <j.glisse@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon/kms: add MSI module parameter
Alex Deucher [Tue, 1 Nov 2011 18:20:30 +0000]
drm/radeon/kms: add MSI module parameter

commit a18cee15ed4c8b6a35f96b7b26a46bac32e04bd9 upstream.

Allow the user to override whether MSIs are enabled
or not on supported ASICs.  MSIs are disabled by default
on IGP chips as they tend not to work.  However certain
IGP chips only seem to work with MSIs enabled.

I suspect this is a chipset or bios issue, but I'm not sure
what the proper fix is.  This will at least make diagnosing
and working around the problem much easier.

See:
https://bugs.freedesktop.org/show_bug.cgi?id=37679

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon/kms: Add MSI quirk for Dell RS690
Alex Deucher [Tue, 1 Nov 2011 18:14:18 +0000]
drm/radeon/kms: Add MSI quirk for Dell RS690

commit 01e718ec194e30b3e8eb3858c742c13649757efc upstream.

Some Dell laptops only seem to work with MSIs.  This
looks like a platform/bios bug.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=37679

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon/kms: properly set panel mode for eDP
Alex Deucher [Mon, 31 Oct 2011 12:54:41 +0000]
drm/radeon/kms: properly set panel mode for eDP

commit 00dfb8df5bf8c3afe4c0bb8361133156b06b7a2c upstream.

This should make eDP more reliable.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon: set hpd polarity at init time so hotplug detect works
Jerome Glisse [Fri, 28 Oct 2011 21:52:34 +0000]
drm/radeon: set hpd polarity at init time so hotplug detect works

commit 8ab250d4484b72ccc78e34276c5ffa84c1d41303 upstream.

Polarity needs to be set accordingly to connector status (connected
or disconnected). Set it up at module init so first hotplug works
reliably no matter what is the initial set of connector.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon/kms: Add MSI quirk for HP RS690
Alex Deucher [Tue, 25 Oct 2011 19:11:08 +0000]
drm/radeon/kms: Add MSI quirk for HP RS690

commit b362105f7f5223fa4d2e03ceeea0e51da754ccc6 upstream.

Some HP laptops only seem to work with MSIs.  This
looks like a platform/bios bug.

Fixes:
https://bugs.freedesktop.org/show_bug.cgi?id=37679

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon/kms: split MSI check into a separate function
Alex Deucher [Tue, 25 Oct 2011 18:58:49 +0000]
drm/radeon/kms: split MSI check into a separate function

commit 8f6c25c59b0c895c68cae59d1b34e9a7b36971bc upstream.

This makes it easier to add quirks for certain systems.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrm/radeon: avoid bouncing connector status btw disconnected & unknown
Jerome Glisse [Mon, 24 Oct 2011 22:16:34 +0000]
drm/radeon: avoid bouncing connector status btw disconnected & unknown

commit 340764465aa4a586ca332e61ae64883e5ad6f183 upstream.

Since force handling rework of d0d0a225e6ad43314c9aa7ea081f76adc5098ad4
we could end up bouncing connector status btw disconnected and unknown.
When connector status change a call to output_poll_changed happen which
in turn ask again for detect but with force set.

So set the load detect flags whenever we report the connector as
connected or unknown this avoid bouncing btw disconnected and unknown.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Cc: Stefan Lippers-Hollmann <s.L-H@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda - Fix a regression for DMA-position check with CA0110
Takashi Iwai [Sun, 6 Nov 2011 12:49:13 +0000]
ALSA: hda - Fix a regression for DMA-position check with CA0110

commit 69f9ba9b0cad67bc03f0a096f7f274de795ca844 upstream.

The regression-fix in 3.1 for the check of DMA-position validity caused
yet another regression for CA0110.  As usual, this hardware seems working
only with LPIB properly.  Adding the appropriate driver-caps bit to force
LPIB fixes the problem.

Reported-and-tested-by: Andres Freund <andres@anarazel.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda - Fix silent output regression with ALC861
Takashi Iwai [Sun, 6 Nov 2011 10:25:34 +0000]
ALSA: hda - Fix silent output regression with ALC861

commit 43dea228a3ba5463392281535dfb3d3fe56f4c2c upstream.

The 3.1 kernel has a regression for ALC861 codec where no sound output
is heard with the default setup.  It's because the amps in DACs aren't
properly unmuted while the output mixers are assigned only to pins.

This patch fixes the missing initialization of DACs when no mixer is
assigned to them.

Tested-by: Andrea Iob <andrea_iob@yahoo.it>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda/realtek - Skip invalid digital out pins
Takashi Iwai [Thu, 3 Nov 2011 15:54:06 +0000]
ALSA: hda/realtek - Skip invalid digital out pins

commit 51e4152a969aa6d2306492ebf143932dcb535c9b upstream.

Some BIOS report invalid pins as digital output pins.  The driver checks
the connection but it doesn't do it fully correctly, and it leaves some
undefined value as the audio-out widget, which makes the driver spewing
warnings.  This patch fixes the issue.

Reference: https://bugzilla.novell.com/show_bug.cgi?id=727348

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda/realtek - Don't create alt-stream for capture when unnecessary
Takashi Iwai [Wed, 2 Nov 2011 20:30:51 +0000]
ALSA: hda/realtek - Don't create alt-stream for capture when unnecessary

commit 1fa1757366783fb52e6e85c2d735db49b818d382 upstream.

When the driver finds multiple ADCs, it tries to create an alternative
capture PCM stream.  However, these secondary ADCs might be useless or
in uncontrolled paths in some cases, e.g. when auto-mic or dynamic
ADC-switching is enabled.  Also, when only a single capture source is
available, the multi-streams don't make sense, too.

With this patch, the driver checks such condition and skips the alt
stream appropriately.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda - Add support for 92HD65 / 92HD66 family of codecs
Charles Chin [Wed, 2 Nov 2011 06:56:58 +0000]
ALSA: hda - Add support for 92HD65 / 92HD66 family of codecs

commit ad5d8755116b431f0709c745ee17cb567a478d43 upstream.

These codecs have SPDIF-in, which is new to the 92HD83xxx compatible
families, so a bit of logic is added to support them.

Signed-off-by: Charles Chin <Charles.Chin@idt.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda - Disable power-widget control for IDT 92HD83/93 as default
Charles Chin [Wed, 2 Nov 2011 06:53:30 +0000]
ALSA: hda - Disable power-widget control for IDT 92HD83/93 as default

commit 35c11777b906042eca9e6f1c03e464726c7faa07 upstream.

The power-widget control in patch_stac92hd83xxx() never worked properly,
thus it's safer to turn it off as default for now.

Signed-off-by: Charles Chin <Charles.Chin@idt.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: ua101: fix crash when unplugging
Clemens Ladisch [Sat, 15 Oct 2011 21:19:25 +0000]
ALSA: ua101: fix crash when unplugging

commit 862a6244eb9f9f5123fe819454fcfcae0ee1f2f9 upstream.

If the device is unplugged while running, it is possible for a PCM
device to be closed after the disconnect callback has returned.  This
means that kill_stream_urb() and disable_iso_interface() would try to
access already-invalid or freed USB data structures.

The function free_usb_related_resources() was intended to prevent this,
but forgot to clear the affected variables.

Reported-and-tested-by: Olivier Courtay <olivier@courtay.org>
Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agonet: Unlock sock before calling sk_free()
Thomas Gleixner [Tue, 25 Oct 2011 02:30:50 +0000]
net: Unlock sock before calling sk_free()

[ Upstream commit b0691c8ee7c28a72748ff32e91b165ec12ae4de6 ]

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agothp: share get_huge_page_tail()
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:36 +0000]
thp: share get_huge_page_tail()

commit b35a35b556f5e6b7993ad0baf20173e75c09ce8c upstream.

This avoids duplicating the function in every arch gup_fast.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agosparc: gup_pte_range() support THP based tail recounting
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:31 +0000]
sparc: gup_pte_range() support THP based tail recounting

commit e0d85a366c2300efd230ef82a9b22110b0658331 upstream.

Up to this point the code assumed old refcounting for hugepages (pre-thp).
 This updates the code directly to the thp mapcount tail page refcounting.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Acked-by: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agos390: gup_huge_pmd() return 0 if pte changes
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:28 +0000]
s390: gup_huge_pmd() return 0 if pte changes

commit 0693bc9ce2cc4f6a1b9c3c05790fc149a74c0b87 upstream.

s390 didn't return 0 in that case, if it's rolling back the *nr pointer it
should also return zero to avoid adding pages to the array at the wrong
offset.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agos390: gup_huge_pmd() support THP tail recounting
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:25 +0000]
s390: gup_huge_pmd() support THP tail recounting

commit 220a2eb228d032acde60e9fd044ca802706ff583 upstream.

Up to this point the code assumed old refcounting for hugepages (pre-thp).
This updates the code directly to the thp mapcount tail page refcounting.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: gup_huge_pmd() return 0 if pte changes
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:19 +0000]
powerpc: gup_huge_pmd() return 0 if pte changes

commit cf592bf768c4fa40282b8fce58a80820065de2cb upstream.

powerpc didn't return 0 in that case, if it's rolling back the *nr pointer
it should also return zero to avoid adding pages to the array at the wrong
offset.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: gup_hugepte() support THP based tail recounting
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:15 +0000]
powerpc: gup_hugepte() support THP based tail recounting

commit 3526741f0964c88bc2ce511e1078359052bf225b upstream.

Up to this point the code assumed old refcounting for hugepages (pre-thp).
This updates the code directly to the thp mapcount tail page refcounting.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: gup_hugepte() avoid freeing the head page too many times
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:11 +0000]
powerpc: gup_hugepte() avoid freeing the head page too many times

commit 8596468487e2062cae2aad56e973784e03959245 upstream.

We only taken "refs" pins on the head page not "*nr" pins.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: get_hugepte() don't put_page() the wrong page
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:08 +0000]
powerpc: get_hugepte() don't put_page() the wrong page

commit 405e44f2e312dd5dd63e5a9f459bffcbcd4368ef upstream.

"page" may have changed to point to the next hugepage after the loop
completed, The references have been taken on the head page, so the
put_page must happen there too.

This is a longstanding issue pre-thp inclusion.

It's totally unclear how these page_cache_add_speculative and
pte_val(pte) != pte_val(*ptep) checks are necessary across all the
powerpc gup_fast code, when x86 doesn't need any of that: there's no way
the page can be freed with irq disabled so we're guaranteed the
atomic_inc will happen on a page with page_count > 0 (so not needing the
speculative check).

The pte check is also meaningless on x86: no need to rollback on x86 if
the pte changed, because the pte can still change a CPU tick after the
check succeeded and it won't be rolled back in that case.  The important
thing is we got a reference on a valid page that was mapped there a CPU
tick ago.  So not knowing the soft tlb refill code of ppc64 in great
detail I'm not removing the "speculative" page_count increase and the
pte checks across all the code, but unless there's a strong reason for
it they should be later cleaned up too.

If a pte can change from huge to non-huge (like it could happen with
THP) passing a pte_t *ptep to gup_hugepte() would also require to repeat
the is_hugepd in gup_hugepte(), but that shouldn't happen with hugetlbfs
only so I'm not altering that.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agopowerpc: remove superfluous PageTail checks on the pte gup_fast
Andrea Arcangeli [Wed, 2 Nov 2011 20:37:03 +0000]
powerpc: remove superfluous PageTail checks on the pte gup_fast

commit 2839bdc1bfc0af76a2f0f11eca011590520a04fa upstream.

This part of gup_fast doesn't seem capable of handling hugetlbfs ptes,
those should be handled by gup_hugepd only, so these checks are
superfluous.

Plus if this wasn't a noop, it would have oopsed because, the insistence
of using the speculative refcounting would trigger a VM_BUG_ON if a tail
page was encountered in the page_cache_get_speculative().

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoxHCI: Clear PLC for USB2 root hub ports
Andiry Xu [Fri, 23 Sep 2011 21:19:50 +0000]
xHCI: Clear PLC for USB2 root hub ports

commit 6fd4562178508a0949c9fdecd8558d8b10d671bd upstream.

When the link state changes, xHC will report a port status change event
and set the PORT_PLC bit, for both USB3 and USB2 root hub ports.

The PLC will be cleared by usbcore for USB3 root hub ports, but not for
USB2 ports, because they do not report USB_PORT_STAT_C_LINK_STATE in
wPortChange.

Clear it for USB2 root hub ports in handle_port_status().

Signed-off-by: Andiry Xu <andiry.xu@amd.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoxHCI: test and clear RWC bit
Andiry Xu [Fri, 23 Sep 2011 21:19:49 +0000]
xHCI: test and clear RWC bit

commit d2f52c9e585bbb1a3c164e02b8dcd0d996c67353 upstream.

Introduce xhci_test_and_clear_bit() to clear RWC bit in PORTSC register.

Signed-off-by: Andiry Xu <andiry.xu@amd.com>
Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoxhci: If no endpoints changed, don't issue BW command.
Sarah Sharp [Fri, 2 Sep 2011 18:05:40 +0000]
xhci: If no endpoints changed, don't issue BW command.

commit 2dc3753997f3c80ce8b950242ab9bb3fb936acfd upstream.

Some alternate interface settings have no endpoints associated with them.
This shows up in some USB webcams, particularly the Logitech HD 1080p,
which uses the uvcvideo driver.  If a driver switches between two alt
settings with no endpoints, there is no need to issue a configure endpoint
command, because there is no endpoint information to update.

The only time a configure endpoint command with just the add slot flag set
makes sense is when the driver is updating hub characteristics in the slot
context.  However, that code never calls xhci_check_bandwidth, so we
should be safe not issuing a command if only the slot context add flag is
set.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agousb_storage: Don't freeze in usb-stor-scan
Seth Forshee [Tue, 26 Jul 2011 14:56:07 +0000]
usb_storage: Don't freeze in usb-stor-scan

commit f02fe890ece7d695a5744b20525d45312382e6e4 upstream.

Scanning cannot be run during suspend or hibernation, but if
usb-stor-scan freezes another thread waiting on scanning to
complete may fail to freeze.

However, if usb-stor-scan is left freezable without ever actually
freezing then the freezer will wait on it to exit, and threads
waiting for scanning to finish will no longer be blocked. One
problem with this approach is that usb-stor-scan has a delay to
wait for devices to settle (which is currently the only point where
it can freeze). To work around this we can request that the freezer
send a fake signal when freezing, then use interruptible sleep to
wake the thread early when freezing happens.

To make this happen, the following changes are made to
usb-stor-scan:

 * Use set_freezable_with_signal() instead of set_freezable() to
   request a fake signal when freezing

 * Use wait_event_interruptible_timeout() instead of
   wait_event_freezable_timeout() to avoid freezing

Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoext4: fix race in xattr block allocation path
Eric Sandeen [Sat, 29 Oct 2011 14:15:35 +0000]
ext4: fix race in xattr block allocation path

commit 6d6a435190bdf2e04c9465cde5bdc3ac68cf11a4 upstream.

Ceph users reported that when using Ceph on ext4, the filesystem
would often become corrupted, containing inodes with incorrect
i_blocks counters.

I managed to reproduce this with a very hacked-up "streamtest"
binary from the Ceph tree.

Ceph is doing a lot of xattr writes, to out-of-inode blocks.
There is also another thread which does sync_file_range and close,
of the same files.  The problem appears to happen due to this race:

sync/flush thread               xattr-set thread
-----------------               ----------------

do_writepages                   ext4_xattr_set
ext4_da_writepages              ext4_xattr_set_handle
mpage_da_map_blocks             ext4_xattr_block_set
        set DELALLOC_RESERVE
                                ext4_new_meta_blocks
                                        ext4_mb_new_blocks
                                                if (!i_delalloc_reserved_flag)
                                                        vfs_dq_alloc_block
ext4_get_blocks
down_write(i_data_sem)
        set i_delalloc_reserved_flag
...
up_write(i_data_sem)
                                        if (i_delalloc_reserved_flag)
                                                vfs_dq_alloc_block_nofail

In other words, the sync/flush thread pops in and sets
i_delalloc_reserved_flag on the inode, which makes the xattr thread
think that it's in a delalloc path in ext4_new_meta_blocks(),
and add the block for a second time, after already having added
it once in the !i_delalloc_reserved_flag case in ext4_mb_new_blocks

The real problem is that we shouldn't be using the DELALLOC_RESERVED
state flag, and instead we should be passing
EXT4_GET_BLOCKS_DELALLOC_RESERVE down to ext4_map_blocks() instead of
using an inode state flag.  We'll fix this for now with using
i_data_sem to prevent this race, but this is really not the right way
to fix things.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoext4: let ext4_page_mkwrite stop started handle in failure
Yongqiang Yang [Wed, 26 Oct 2011 09:00:19 +0000]
ext4: let ext4_page_mkwrite stop started handle in failure

commit fcbb5515825f1bb20b7a6f75ec48bee61416f879 upstream.

The started journal handle should be stopped in failure case.

Signed-off-by: Yongqiang Yang <xiaoqiangnk@gmail.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoext4: call ext4_handle_dirty_metadata with correct inode in ext4_dx_add_entry
Theodore Ts'o [Wed, 31 Aug 2011 16:02:51 +0000]
ext4: call ext4_handle_dirty_metadata with correct inode in ext4_dx_add_entry

commit 5930ea643805feb50a2f8383ae12eb6f10935e49 upstream.

ext4_dx_add_entry manipulates bh2 and frames[0].bh, which are two buffer_heads
that point to directory blocks assigned to the directory inode.  However, the
function calls ext4_handle_dirty_metadata with the inode of the file that's
being added to the directory, not the directory inode itself.  Therefore,
correct the code to dirty the directory buffers with the directory inode, not
the file inode.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoext4: ext4_mkdir should dirty dir_block with newly created directory inode
Darrick J. Wong [Wed, 31 Aug 2011 16:00:51 +0000]
ext4: ext4_mkdir should dirty dir_block with newly created directory inode

commit f9287c1f2d329f4d78a3bbc9cf0db0ebae6f146a upstream.

ext4_mkdir calls ext4_handle_dirty_metadata with dir_block and the inode "dir".
Unfortunately, dir_block belongs to the newly created directory (which is
"inode"), not the parent directory (which is "dir").  Fix the incorrect
association.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoext4: ext4_rename should dirty dir_bh with the correct directory
Darrick J. Wong [Wed, 31 Aug 2011 15:58:51 +0000]
ext4: ext4_rename should dirty dir_bh with the correct directory

commit bcaa992975041e40449be8c010c26192b8c8b409 upstream.

When ext4_rename performs a directory rename (move), dir_bh is a
buffer that is modified to update the '..' link in the directory being
moved (old_inode).  However, ext4_handle_dirty_metadata is called with
the old parent directory inode (old_dir) and dir_bh, which is
incorrect because dir_bh does not belong to the parent inode.  Fix
this error.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoext2,ext3,ext4: don't inherit APPEND_FL or IMMUTABLE_FL for new inodes
Theodore Ts'o [Wed, 31 Aug 2011 15:54:51 +0000]
ext2,ext3,ext4: don't inherit APPEND_FL or IMMUTABLE_FL for new inodes

commit 1cd9f0976aa4606db8d6e3dc3edd0aca8019372a upstream.

This doesn't make much sense, and it exposes a bug in the kernel where
attempts to create a new file in an append-only directory using
O_CREAT will fail (but still leave a zero-length file).  This was
discovered when xfstests #79 was generalized so it could run on all
file systems.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrivers/power/ds2780_battery.c: fix deadlock upon insertion and removal
Clifton Barnes [Wed, 2 Nov 2011 20:39:55 +0000]
drivers/power/ds2780_battery.c: fix deadlock upon insertion and removal

commit 0e053fcbbbc4d945247cb32cad2767b483cb65f8 upstream.

Fixes the deadlock when inserting and removing the ds2780.

Signed-off-by: Clifton Barnes <cabarnes@indesign-llc.com>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrivers/power/ds2780_battery.c: add a nolock function to w1 interface
Clifton Barnes [Wed, 2 Nov 2011 20:39:52 +0000]
drivers/power/ds2780_battery.c: add a nolock function to w1 interface

commit 9fe678fa2feb4aaac0b4220de63e1b7f8ccebae6 upstream.

Adds a nolock function to the w1 interface to avoid locking the
mutex if needed.

Signed-off-by: Clifton Barnes <cabarnes@indesign-llc.com>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrivers/power/ds2780_battery.c: create central point for calling w1 interface
Clifton Barnes [Wed, 2 Nov 2011 20:39:50 +0000]
drivers/power/ds2780_battery.c: create central point for calling w1 interface

commit 853eee72f74f449797f0500ea19fc1bf497428d8 upstream.

Simply creates one point to call the w1 interface.

Signed-off-by: Clifton Barnes <cabarnes@indesign-llc.com>
Cc: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agohwspinlock/core: use a mutex to protect the radix tree
Juan Gutierrez [Tue, 6 Sep 2011 06:30:16 +0000]
hwspinlock/core: use a mutex to protect the radix tree

commit 93b465c2e186d96fb90012ba0f9372eb9952e732 upstream.

Since we're using non-atomic radix tree allocations, we
should be protecting the tree using a mutex and not a
spinlock.

Non-atomic allocations and process context locking is good enough,
as the tree is manipulated only when locks are registered/
unregistered/requested/freed.

The locks themselves are still protected by spinlocks of course,
and mutexes are not involved in the locking/unlocking paths.

Signed-off-by: Juan Gutierrez <jgutierrez@ti.com>
[ohad@wizery.com: rewrite the commit log, #include mutex.h, add minor
commentary]
[ohad@wizery.com: update register/unregister parts in hwspinlock.txt]
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agodrivers/net/rionet.c: fix ethernet address macros for LE platforms
Alexandre Bounine [Wed, 2 Nov 2011 20:39:15 +0000]
drivers/net/rionet.c: fix ethernet address macros for LE platforms

commit e0c87bd95e8dad455c23bc56513af8dcb1737e55 upstream.

Modify Ethernet addess macros to be compatible with BE/LE platforms

Signed-off-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Chul Kim <chul.kim@idt.com>
Cc: Kumar Gala <galak@kernel.crashing.org>
Cc: Matt Porter <mporter@kernel.crashing.org>
Cc: Li Yang <leoli@freescale.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoALSA: hda - Add missing static ADC tables for ALC269 quirks
Takashi Iwai [Thu, 3 Nov 2011 14:56:55 +0000]
ALSA: hda - Add missing static ADC tables for ALC269 quirks

[There is no upstream commit for this patch since the corresponding
code was removed from 3.2 kernel.  This is a regression found only in
3.1 kernel, so please apply this only to 3.1.x series.]

Some ALC269 quirks define their own .cap_mixer field but without the
static adc_nids[].  This resulted in the mismatch of ADC because ALC269
may have the widget 0x07 for another audio-in, and the auto-parser picks
this up instead.

This patch fixes the problem by adding the static adc_nids[] and co
again to these entries.

Tested-by: Chris Vine <chris@cvine.freeserve.co.uk>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agobinfmt_elf: fix PIE execution with randomization disabled
Jiri Kosina [Wed, 2 Nov 2011 20:37:41 +0000]
binfmt_elf: fix PIE execution with randomization disabled

commit a3defbe5c337dbc6da911f8cc49ae3cc3b49b453 upstream.

The case of address space randomization being disabled in runtime through
randomize_va_space sysctl is not treated properly in load_elf_binary(),
resulting in SIGKILL coming at exec() time for certain PIE-linked binaries
in case the randomization has been disabled at runtime prior to calling
exec().

Handle the randomize_va_space == 0 case the same way as if we were not
supporting .text randomization at all.

Based on original patch by H.J. Lu and Josh Boyer.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: H.J. Lu <hongjiu.lu@intel.com>
Cc: <stable@kernel.org>
Tested-by: Josh Boyer <jwboyer@redhat.com>
Acked-by: Nicolas Pitre <nicolas.pitre@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agomm: thp: tail page refcounting fix
Andrea Arcangeli [Wed, 2 Nov 2011 20:36:59 +0000]
mm: thp: tail page refcounting fix

commit 70b50f94f1644e2aa7cb374819cfd93f3c28d725 upstream.

Michel while working on the working set estimation code, noticed that
calling get_page_unless_zero() on a random pfn_to_page(random_pfn)
wasn't safe, if the pfn ended up being a tail page of a transparent
hugepage under splitting by __split_huge_page_refcount().

He then found the problem could also theoretically materialize with
page_cache_get_speculative() during the speculative radix tree lookups
that uses get_page_unless_zero() in SMP if the radix tree page is freed
and reallocated and get_user_pages is called on it before
page_cache_get_speculative has a chance to call get_page_unless_zero().

So the best way to fix the problem is to keep page_tail->_count zero at
all times.  This will guarantee that get_page_unless_zero() can never
succeed on any tail page.  page_tail->_mapcount is guaranteed zero and
is unused for all tail pages of a compound page, so we can simply
account the tail page references there and transfer them to
tail_page->_count in __split_huge_page_refcount() (in addition to the
head_page->_mapcount).

While debugging this s/_count/_mapcount/ change I also noticed get_page is
called by direct-io.c on pages returned by get_user_pages.  That wasn't
entirely safe because the two atomic_inc in get_page weren't atomic.  As
opposed to other get_user_page users like secondary-MMU page fault to
establish the shadow pagetables would never call any superflous get_page
after get_user_page returns.  It's safer to make get_page universally safe
for tail pages and to use get_page_foll() within follow_page (inside
get_user_pages()).  get_page_foll() is safe to do the refcounting for tail
pages without taking any locks because it is run within PT lock protected
critical sections (PT lock for pte and page_table_lock for
pmd_trans_huge).

The standard get_page() as invoked by direct-io instead will now take
the compound_lock but still only for tail pages.  The direct-io paths
are usually I/O bound and the compound_lock is per THP so very
finegrined, so there's no risk of scalability issues with it.  A simple
direct-io benchmarks with all lockdep prove locking and spinlock
debugging infrastructure enabled shows identical performance and no
overhead.  So it's worth it.  Ideally direct-io should stop calling
get_page() on pages returned by get_user_pages().  The spinlock in
get_page() is already optimized away for no-THP builds but doing
get_page() on tail pages returned by GUP is generally a rare operation
and usually only run in I/O paths.

This new refcounting on page_tail->_mapcount in addition to avoiding new
RCU critical sections will also allow the working set estimation code to
work without any further complexity associated to the tail page
refcounting with THP.

Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Reported-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Michel Lespinasse <walken@google.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <jweiner@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agotg3: fix tigon3_dma_hwbug_workaround()
Eric Dumazet [Sat, 22 Oct 2011 07:29:53 +0000]
tg3: fix tigon3_dma_hwbug_workaround()

[ Upstream commit f7ff19871bb4a3451e1ca2cf660bf633018cfbec ]

Ari got kernel panics using tg3 NIC, and bisected to 2669069aacc9 "tg3:
enable transmit time stamping."

This is because tigon3_dma_hwbug_workaround() might alloc a new skb and
free the original. We panic when skb_tx_timestamp() is called on freed
skb.

Reported-by: Ari Savolainen <ari.m.savolainen@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoipv6: fix route error binding peer in func icmp6_dst_alloc
Gao feng [Fri, 28 Oct 2011 02:46:57 +0000]
ipv6: fix route error binding peer in func icmp6_dst_alloc

[ Upstream commit 7011687f0f2f12cc348f6a693fafe63d89167eac ]

in func icmp6_dst_alloc,dst_metric_set call ipv6_cow_metrics to set metric.
ipv6_cow_metrics may will call rt6_bind_peer to set rt6_info->rt6i_peer.
So,we should move ipv6_addr_copy before dst_metric_set to make sure rt6_bind_peer success.

Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoipv4: fix ipsec forward performance regression
Yan, Zheng [Sat, 22 Oct 2011 21:58:20 +0000]
ipv4: fix ipsec forward performance regression

[ Upstream commit b73233960a59ee66e09d642f13d0592b13651e94 ]

There is bug in commit 5e2b61f(ipv4: Remove flowi from struct rtable).
It makes xfrm4_fill_dst() modify wrong data structure.

Signed-off-by: Zheng Yan <zheng.z.yan@intel.com>
Reported-by: Kim Phillips <kim.phillips@freescale.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agojsm: remove buggy write queue
Thadeu Lima de Souza Cascardo [Wed, 24 Aug 2011 16:14:22 +0000]
jsm: remove buggy write queue

commit 9d898966c4a07e4a5092215b5a2829d0ef02baa2 upstream.

jsm uses a write queue that copies from uart_core circular buffer. This
copying however has some bugs, like not wrapping the head counter. Since
this write queue is also a circular buffer, the consumer function is
ready to use the uart_core circular buffer directly.

This buggy copying function was making some bytes be dropped when
transmitting to a raw tty, doing something like this.

[root@hostname ~]$ cat /dev/ttyn1 > cascardo/dump &
[1] 2658
[root@hostname ~]$ cat /proc/tty/drivers > /dev/ttyn0
[root@hostname ~]$ cat /proc/tty/drivers
/dev/tty             /dev/tty        5       0 system:/dev/tty
/dev/console         /dev/console    5       1 system:console
/dev/ptmx            /dev/ptmx       5       2 system
/dev/vc/0            /dev/vc/0       4       0 system:vtmaster
jsm                  /dev/ttyn     250 0-31 serial
serial               /dev/ttyS       4 64-95 serial
hvc                  /dev/hvc      229 0-7 system
pty_slave            /dev/pts      136 0-1048575 pty:slave
pty_master           /dev/ptm      128 0-1048575 pty:master
unknown              /dev/tty        4 1-63 console
[root@hostname ~]$ cat cascardo/dump
/dev/tty             /dev/tty        5       0 system:/dev/tty
/dev/console         /dev/console    5       1 system:console
/dev/ptmx            /dev/ptmx       5       2 system
/dev/vc/0            /dev/vc/0       4       0 system:vtmaste[root@hostname ~]$

This patch drops the driver write queue entirely, using the circular
buffer from uart_core only.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoOMAP: SPI: Fix the trying to free nonexistent resource error
Shubhrajyoti D [Mon, 24 Oct 2011 10:24:24 +0000]
OMAP: SPI: Fix the trying to free nonexistent resource error

commit 1458d160de3f1862aeaac57447ba96e7857ac52b upstream.

Currently there is a request_mem_region(r->start, ..
followed by r->start += pdata->regs_offset;

And then in remove

   r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
   release_mem_region(r->start, resource_size(r));

Here the offset addition is not taken care. Fix the code for the
same.

Signed-off-by: Shubhrajyoti D <shubhrajyoti@ti.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoum: Fix kmalloc argument order in um/vdso/vma.c
Dave Jones [Mon, 24 Oct 2011 22:15:32 +0000]
um: Fix kmalloc argument order in um/vdso/vma.c

commit 0d65ede0a605d6252acc5c8a9c536c4cd0211f3c upstream.

kmalloc size is 1st arg, not second.

Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoreadlinkat: ensure we return ENOENT for the empty pathname for normal lookups
Andy Whitcroft [Wed, 2 Nov 2011 08:44:39 +0000]
readlinkat: ensure we return ENOENT for the empty pathname for normal lookups

commit 1fa1e7f615f4d3ae436fa319af6e4eebdd4026a8 upstream.

Since the commit below which added O_PATH support to the *at() calls, the
error return for readlink/readlinkat for the empty pathname has switched
from ENOENT to EINVAL:

  commit 65cfc6722361570bfe255698d9cd4dccaf47570d
  Author: Al Viro <viro@zeniv.linux.org.uk>
  Date:   Sun Mar 13 15:56:26 2011 -0400

    readlinkat(), fchownat() and fstatat() with empty relative pathnames

This is both unexpected for userspace and makes readlink/readlinkat
inconsistant with all other interfaces; and inconsistant with our stated
return for these pathnames.

As the readlinkat call does not have a flags parameter we cannot use the
AT_EMPTY_PATH approach used in the other calls.  Therefore expose whether
the original path is infact entry via a new user_path_at_empty() path
lookup function.  Use this to determine whether to default to EINVAL or
ENOENT for failures.

Addresses http://bugs.launchpad.net/bugs/817187

[akpm@linux-foundation.org: remove unused getname_flags()]
Signed-off-by: Andy Whitcroft <apw@canonical.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

7 years agoum: fix ubd cow size
Richard Weinberger [Wed, 2 Nov 2011 12:17:27 +0000]
um: fix ubd cow size

commit 8535639810e578960233ad39def3ac2157b0c3ec upstream.

ubd_file_size() cannot use ubd_dev->cow.file because at this time
ubd_dev->cow.file is not initialized.
Therefore, ubd_file_size() will always report a wrong disk size when
COW files are used.
Reading from /dev/ubd* would crash the kernel.

We have to read the correct disk size from the COW file's backing
file.

Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>