firefly-linux-kernel-4.4.55.git
14 years agodrm/i915: Fix sync to vblank when VGA output is turned off
Li Peng [Wed, 16 Dec 2009 15:33:25 +0000 (16:33 +0100)]
drm/i915: Fix sync to vblank when VGA output is turned off

commit 778c902640530371a169ad1c03566e7c51b09874 upstream

In current vblank-wait implementation, if we turn off VGA output,
drm_wait_vblank will still wait on the disabled pipe until timeout,
because vblank on the pipe is assumed be enabled. This would cause
slow system response on some system such as moblin.

This patch resolve the issue by adding a drm helper function
drm_vblank_off which explicitly clear vblank_enabled[crtc], wake up
any waiting queue and save last vblank counter before turning off
crtc. It also slightly change drm_vblank_get to ensure that we will
will return immediately if trying to wait on a disabled pipe.

Signed-off-by: Li Peng <peng.li@intel.com>
Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>
[anholt: hand-applied for conflicts with overlay changes]
Signed-off-by: Eric Anholt <eric@anholt.net>
Cc: Stefan Bader <stefan.bader@canonical.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomac80211: Fix dynamic power save for scanning.
Vivek Natarajan [Wed, 16 Dec 2009 16:51:45 +0000 (11:51 -0500)]
mac80211: Fix dynamic power save for scanning.

Upstream commit: 7c3f4bbedc241ddcd3abe1f419c356e625231da1

Not only ps_sdata but also IEEE80211_CONF_PS is to be considered
before restoring PS in scan_ps_disable(). For instance, when ps_sdata
is set but CONF_PS is not set just because the dynamic timer is still
running, a sw scan leads to setting of CONF_PS in scan_ps_disable
instead of restarting the dynamic PS timer.
Also for the above case, a null data frame is to be sent after
returning to operating channel which was not happening with the
current implementation. This patch fixes this too.

Signed-off-by: Vivek Natarajan <vnatarajan@atheros.com>
Reviewed-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoath9k: fix tx status reporting
Felix Fietkau [Wed, 16 Dec 2009 16:51:44 +0000 (11:51 -0500)]
ath9k: fix tx status reporting

This is a backport of upstream commit: e8c6342d989e241513baeba4b05a04b6b1f3bc8b

This patch fixes a bug in ath9k's tx status check, which
caused mac80211 to consider regularly transmitted unicast frames
as un-acked.

When checking the ts_status field for errors, it needs to be masked
with ATH9K_TXERR_FILT, because this field also contains other fields
like ATH9K_TX_ACKED.

Without this patch, AP mode is pretty much unusable, as hostapd
checks the ACK status for the frames that it injects.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoath9k: Fix maximum tx fifo settings for single stream devices
Luis R. Rodriguez [Wed, 16 Dec 2009 16:51:43 +0000 (11:51 -0500)]
ath9k: Fix maximum tx fifo settings for single stream devices

This is a backport of upstream commit: f4709fdf683e1ed37b321c258b614ebe39752bf3

Atheros single stream AR9285 and AR9271 have half the PCU TX FIFO
buffer size of that of dual stream devices. Dual stream devices
have a max PCU TX FIFO size of 8 KB while single stream devices
have 4 KB. Single stream devices have an issue though and require
hardware only to use half of the amount of its capable PCU TX FIFO
size, 2 KB and this requires a change in software.

Technically a change would not have been required (except for frame
burst considerations of 128 bytes) if these devices would have been
able to use the full 4 KB of the PCU TX FIFO size but our systems
engineers recommend 2 KB to be used only. We enforce this through
software by reducing the max frame triggger level to 2 KB.

Fixing the max frame trigger level should then have a few benefits:

  * The PER will now be adjusted as designed for underruns when the
    max trigger level is reached. This should help alleviate the
    bus as the rate control algorithm chooses a slower rate which
    should ensure frames are transmitted properly under high system
    bus load.

  * The poll we use on our TX queues should now trigger and work
    as designed for single stream devices. The hardware passes
    data from each TX queue on the PCU TX FIFO queue respecting each
    queue's priority. The new trigger level ensures this seeding of
    the PCU TX FIFO queue occurs as designed which could mean avoiding
    false resets and actually reseting hw correctly when a TX queue
    is indeed stuck.

  * Some undocumented / unsupported behaviour could have been triggered
    when the max trigger level level was being set to 4 KB on single
    stream devices. Its not clear what this issue was to me yet.

Cc: Kyungwan Nam <kyungwan.nam@atheros.com>
Cc: Bennyam Malavazi <bennyam.malavazi@atheros.com>
Cc: Stephen Chen <stephen.chen@atheros.com>
Cc: Shan Palanisamy <shan.palanisamy@atheros.com>
Cc: Paul Shaw <paul.shaw@atheros.com>
Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoath9k: fix processing of TX PS null data frames
Luis R. Rodriguez [Wed, 16 Dec 2009 16:51:42 +0000 (11:51 -0500)]
ath9k: fix processing of TX PS null data frames

This is a backport of upstream commit: e7824a50662f7f79b1a739f705b4d906c31cf221

When mac80211 was telling us to go into Powersave we listened
and immediately turned RX off. This meant hardware would not
see the ACKs from the AP we're associated with and hardware
we'd end up retransmiting the null data frame in a loop
helplessly.

Fix this by keeping track of the transmitted nullfunc frames
and only when we are sure the AP has sent back an ACK do we
go ahead and shut RX off.

Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: Vivek Natarajan <Vivek.Natarajan@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoath9k: Fix TX hang poll routine
Sujith [Wed, 16 Dec 2009 16:51:41 +0000 (11:51 -0500)]
ath9k: Fix TX hang poll routine

This is a backport of upstream commit: 332c556633b8c5fb4e890b1783122f2315526590

When TX is hung, the chip is reset. Ensure that
the chip is awake by using the PS wrappers.

Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotracing: Fix event format export
Johannes Berg [Fri, 13 Nov 2009 22:40:09 +0000 (23:40 +0100)]
tracing: Fix event format export

commit 811cb50baf63461ce0bdb234927046131fc7fa8b upstream.

For some reason the export of the event print format to userspace
uses '#fmt' which breaks if the format string is anything but a plain
string, for example if it is built with macros then the macro names
are exported instead of their contents.

Use
"\"%s\"", fmt
instead of
"%s", #fmt
to export the string and not the way it is built.

For example, in net/mac80211/driver-trace.h for the trace event drv_start
there is:

        TP_printk(
                LOCAL_PR_FMT, LOCAL_PR_ARG
        )

Which use to produce:

 print fmt: LOCAL_PR_FMT, REC->wiphy_name

Now produces:

 print fmt: "%s", REC->wiphy_name

Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
LKML-Reference: <20091113224009.GB23942@elte.hu>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agob43legacy: avoid PPC fault during resume
Larry Finger [Tue, 24 Nov 2009 00:42:36 +0000 (18:42 -0600)]
b43legacy: avoid PPC fault during resume

commit 316a4d966cae3c2dec83ebb1ee1a3515f97b30ff upstream.

For PPC architecture with PHY Revision < 3, a read of the register
B43_MMIO_HWENABLED_LO will cause a CPU fault unless b43legacy_status()
returns a value of 2 (B43legacy_STAT_STARTED); however, one finds that
the driver is unable to associate after resuming from hibernation unless
this routine returns 1. To satisfy both conditions, the routine is rewritten
to return TRUE whenever b43legacy_status() returns a value < 2.

This patch fixes the second problem listed in the postings for Red Hat
Bugzilla #538523.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosparc: Set UTS_MACHINE correctly.
David S. Miller [Sun, 6 Dec 2009 01:17:55 +0000 (17:17 -0800)]
sparc: Set UTS_MACHINE correctly.

[ Upstream commit 7f5620a5fcd658f219e85831d3691908f1eccbde ]

"ARCH" can be just about anything, so we shouldn't end up
with UTS_MACHINE of "sparc" in a 64-bit kernel build just
because someone set the personality using 'sparc32' or
similar.  CONFIG_SPARC64 drives the compilation and
therefore provides the definitive value, not "ARCH".

This mirrors commit 8c6531f7a99f29ba8817ffb12cc9ecf190049bd6
(x86: correctly set UTS_MACHINE for "make ARCH=x86")

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosparc64: Fix stack debugging IRQ stack regression.
David S. Miller [Wed, 9 Dec 2009 09:43:45 +0000 (01:43 -0800)]
sparc64: Fix stack debugging IRQ stack regression.

[ Upstream commit 166e553a575f09485f6d0df8a1ef3c5991f7d953 ]

Commit 4f70f7a91bffdcc39f088748dc678953eb9a3fbd
(sparc64: Implement IRQ stacks.) has two bugs.

First, the softirq range check forgets to subtract STACK_BIAS
before comparing with %sp.  Next, on failure the wrong label
is jumped to, resulting in a bogus stack being loaded.

Reported-by: Igor Kovalenko <igor.v.kovalenko@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosparc64: Fix overly strict range type matching for PCI devices.
David S. Miller [Wed, 9 Dec 2009 09:39:09 +0000 (01:39 -0800)]
sparc64: Fix overly strict range type matching for PCI devices.

[ Upstream commit 4230fa3b89ea1c413766bd411a8315a3d05aa6c7 ]

When we are trying to see if a range property entry applies
to a given address, we are overly strict about the type.

We should only allow I/O ranges for I/O addresses, and only allow
CONFIG space ranges for CONFIG space address.

However for MEM ranges, they come in 32-bit and 64-bit flavors.
And a lack of an exact match is OK if the range is 32-bit and
the address is 64-bit.  We can assign a 64-bit address properly
into a 32-bit parent range just fine.

So allow it.

Reported-by: Patrick Finnegan <pat@computer-refuge.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agosparc64: Don't specify IRQF_SHARED for LDC interrupts.
David S. Miller [Thu, 10 Dec 2009 07:44:43 +0000 (23:44 -0800)]
sparc64: Don't specify IRQF_SHARED for LDC interrupts.

[ Upstream commit 08a036d583409e3517e3d15b7478d029b25f2cf2 ]

IRQF_SHARED and IRQF_DISABLED don't mix.

Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agob44 WOL setup: one-bit-off stack corruption kernel panic fix
Stanislav Brabec [Wed, 9 Dec 2009 05:00:22 +0000 (21:00 -0800)]
b44 WOL setup: one-bit-off stack corruption kernel panic fix

[ Upstream commit: e0188829cb724e7d12a2d4e343b368ff1d6e1471 ]

About 50% of shutdowns of b44 Ethernet adapter ends by kernel panic
with kernels compiled with stack-protector.

Checking b44_magic_pattern() return values, one call of
b44_magic_pattern() returns 127. It means, that set_bit(128, pmask)
was called on line 1509. It means that bit 0 of 17th byte of pmask was
overwritten. But pmask has only 16 bytes. Stack corruption happens.

It seems that set_bit() on line 1509 always writes one bit off.

The fix does not only solve the stack corruption, but also makes Wake
On LAN working on my onboard B44 on Asus A7V-333X mainboard.

It seems that this problem affects all kernel versions since commit
725ad800 ([PATCH] b44: add wol for old nic) on 2006-06-20.

Signed-off-by: Stanislav Brabec <sbrabec@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoip_fragment: also adjust skb->truesize for packets not owned by a socket
Patrick McHardy [Tue, 1 Dec 2009 23:53:57 +0000 (15:53 -0800)]
ip_fragment: also adjust skb->truesize for packets not owned by a socket

[ Upstream commit b2722b1c3a893ec6021508da15b32282ec79f4da ]

When a large packet gets reassembled by ip_defrag(), the head skb
accounts for all the fragments in skb->truesize. If this packet is
refragmented again, skb->truesize is not re-adjusted to reflect only
the head size since its not owned by a socket. If the head fragment
then gets recycled and reused for another received fragment, it might
exceed the defragmentation limits due to its large truesize value.

skb_recycle_check() explicitly checks for linear skbs, so any recycled
skb should reflect its true size in skb->truesize. Change ip_fragment()
to also adjust the truesize value of skbs not owned by a socket.

Reported-and-tested-by: Ben Menchaca <ben@bigfootnetworks.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agotcp: Stalling connections: Fix timeout calculation routine
Damian Lukowski [Mon, 7 Dec 2009 06:06:15 +0000 (06:06 +0000)]
tcp: Stalling connections: Fix timeout calculation routine

[ Upstream commit 07f29bc5bbae4e53e982ab956fed7207990a7786 ]

This patch fixes a problem in the TCP connection timeout calculation.
Currently, timeout decisions are made on the basis of the current
tcp_time_stamp and retrans_stamp, which is usually set at the first
retransmission.
However, if the retransmission fails in tcp_retransmit_skb(),
retrans_stamp is not updated and remains zero. This leads to wrong
decisions in retransmits_timed_out() if tcp_time_stamp is larger than
the specified timeout, which is very likely.
In this case, the TCP connection dies after the first attempted
(and unsuccessful) retransmission.

With this patch, tcp_skb_cb->when is used instead, when retrans_stamp
is not available.

This bug has been introduced together with retransmits_timed_out() in
2.6.32, as the number of retransmissions has been used for timeout
decisions before. The corresponding commit was
6fa12c85031485dff38ce550c24f10da23b0adaa (Revert Backoff [v3]:
Calculate TCP's connection close threshold as a time value.).

Thanks to Ilpo Järvinen for code suggestions and Frederic Leroy for
testing.

Reported-by: Frederic Leroy <fredo@starox.org>
Signed-off-by: Damian Lukowski <damian@tvk.rwth-aachen.de>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoslc90e66: fix UDMA handling
Bartlomiej Zolnierkiewicz [Mon, 30 Nov 2009 08:55:18 +0000 (08:55 +0000)]
slc90e66: fix UDMA handling

[ Upstream commit ee31527a02b0a8e1aa4a5e4084d2db5fa34737ed ]

Fix checking of the currently programmed UDMA mode.

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
Acked-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm crypt: make wipe message also wipe essiv key
Milan Broz [Thu, 10 Dec 2009 23:51:57 +0000 (23:51 +0000)]
dm crypt: make wipe message also wipe essiv key

commit 542da317668c35036e8471822a564b609d05af66 upstream.

The "wipe key" message is used to wipe the volume key from memory
temporarily, for example when suspending to RAM.

But the initialisation vector in ESSIV mode is calculated from the
hashed volume key, so the wipe message should wipe this IV key too and
reinitialise it when the volume key is reinstated.

This patch adds an IV wipe method called from a wipe message callback.
ESSIV is then reinitialised using the init function added by the
last patch.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm crypt: separate essiv allocation from initialisation
Milan Broz [Thu, 10 Dec 2009 23:51:56 +0000 (23:51 +0000)]
dm crypt: separate essiv allocation from initialisation

commit b95bf2d3d5a48b095bffe2a0cd8c40453cf59557 upstream.

This patch separates the construction of IV from its initialisation.
(For ESSIV it is a hash calculation based on volume key.)

Constructor code now preallocates hash tfm and salt array
and saves it in a private IV structure.

The next patch requires this to reinitialise the wiped IV
without reallocating memory when resuming a suspended device.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm snapshot: cope with chunk size larger than origin
Mikulas Patocka [Thu, 10 Dec 2009 23:51:54 +0000 (23:51 +0000)]
dm snapshot: cope with chunk size larger than origin

commit 8e87b9b81b3c370f7e53c1ab6e1c3519ef37a644 upstream.

Under some special conditions the snapshot hash_size is calculated as zero.
This patch instead sets a minimum value of 64, the same as for the
pending exception table.

rounddown_pow_of_two(0) is an undefined operation (it expands to shift
by -1).  init_exception_table with an argument of 0 would fail with -ENOMEM.

The way to trigger the problem is to create a snapshot with a chunk size
that is larger than the origin device.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm: avoid _hash_lock deadlock
Mikulas Patocka [Thu, 10 Dec 2009 23:51:52 +0000 (23:51 +0000)]
dm: avoid _hash_lock deadlock

commit 6076905b5ef39e0ea58db32583c9e0036c05e47b upstream.

Fix a reported deadlock if there are still unprocessed multipath events
on a device that is being removed.

_hash_lock is held during dev_remove while trying to send the
outstanding events.  Sending the events requests the _hash_lock
again in dm_copy_name_and_uuid.

This patch introduces a separate lock around regions that modify the
link to the hash table (dm_set_mdptr) or the name or uuid so that
dm_copy_name_and_uuid no longer needs _hash_lock.

Additionally, dm_copy_name_and_uuid can only be called if md exists
so we can drop the dm_get() and dm_put() which can lead to a BUG()
while md is being freed.

The deadlock:
 #0 [ffff8106298dfb48] schedule at ffffffff80063035
 #1 [ffff8106298dfc20] __down_read at ffffffff8006475d
 #2 [ffff8106298dfc60] dm_copy_name_and_uuid at ffffffff8824f740
 #3 [ffff8106298dfc90] dm_send_uevents at ffffffff88252685
 #4 [ffff8106298dfcd0] event_callback at ffffffff8824c678
 #5 [ffff8106298dfd00] dm_table_event at ffffffff8824dd01
 #6 [ffff8106298dfd10] __hash_remove at ffffffff882507ad
 #7 [ffff8106298dfd30] dev_remove at ffffffff88250865
 #8 [ffff8106298dfd60] ctl_ioctl at ffffffff88250d80
 #9 [ffff8106298dfee0] do_ioctl at ffffffff800418c4
#10 [ffff8106298dff00] vfs_ioctl at ffffffff8002fab9
#11 [ffff8106298dff40] sys_ioctl at ffffffff8004bdaf
#12 [ffff8106298dff80] tracesys at ffffffff8005d28d (via system_call)

Reported-by: guy keren <choo@actcom.co.il>
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm crypt: restructure essiv error path
Milan Broz [Thu, 10 Dec 2009 23:51:56 +0000 (23:51 +0000)]
dm crypt: restructure essiv error path

commit 5861f1be00b3b70f8ab5e5a81392a6cf69666cd2 upstream.

Use kzfree for salt deallocation because it is derived from the volume
key.  Use a common error path in ESSIV constructor.

Required by a later patch which fixes the way key material is wiped
from memory.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm crypt: move private iv fields to structs
Milan Broz [Thu, 10 Dec 2009 23:51:55 +0000 (23:51 +0000)]
dm crypt: move private iv fields to structs

commit 6047359277517c4e56d8bfd6ea4966d7a3924151 upstream.

Define private structures for IV so it's easy to add further attributes
in a following patch which fixes the way key material is wiped from
memory.  Also move ESSIV destructor and remove unnecessary 'status'
operation.

There are no functional changes in this patch.

Signed-off-by: Milan Broz <mbroz@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm snapshot: only take lock for statustype info not table
Mikulas Patocka [Thu, 10 Dec 2009 23:51:53 +0000 (23:51 +0000)]
dm snapshot: only take lock for statustype info not table

commit 94e76572b5dd37b1f0f4b3742ee8a565daead932 upstream.

Take snapshot lock only for STATUSTYPE_INFO, not STATUSTYPE_TABLE.

Commit 4c6fff445d7aa753957856278d4d93bcad6e2c14
(dm-snapshot-lock-snapshot-while-supplying-status.patch)
introduced this use of the lock, but userspace applications using
libdevmapper have been found to request STATUSTYPE_TABLE while the device
is suspended and the lock is already held, leading to deadlock.  Since
the lock is not necessary in this case, don't try to take it.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodm exception store: free tmp_store on persistent flag error
Julia Lawall [Thu, 10 Dec 2009 23:51:52 +0000 (23:51 +0000)]
dm exception store: free tmp_store on persistent flag error

commit 613978f8711c7fd4d0aa856872375d2abd7c92ff upstream.

Error handling code following a kmalloc should free the allocated data.

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: try harder to balloon up under memory pressure.
Ian Campbell [Fri, 5 Jun 2009 10:58:37 +0000 (11:58 +0100)]
xen: try harder to balloon up under memory pressure.

commit bc2c0303226ec716854d3c208c7f84fe7aa35cd7 upstream.

Currently if the balloon driver is unable to increase the guest's
reservation it assumes the failure was due to reaching its full
allocation, gives up on the ballooning operation and records the limit
it reached as the "hard limit". The driver will not try again until
the target is set again (even to the same value).

However it is possible that ballooning has in fact failed due to
memory pressure in the host and therefore it is desirable to keep
attempting to reach the target in case memory becomes available. The
most likely scenario is that some guests are ballooning down while
others are ballooning up and therefore there is temporary memory
pressure while things stabilise. You would not expect a well behaved
toolstack to ask a domain to balloon to more than its allocation nor
would you expect it to deliberately over-commit memory by setting
balloon targets which exceed the total host memory.

This patch drops the concept of a hard limit and causes the balloon
driver to retry increasing the reservation on a timer in the same
manner as when decreasing the reservation.

Also if we partially succeed in increasing the reservation
(i.e. receive less pages than we asked for) then we may as well keep
those pages rather than returning them to Xen.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoXen balloon: fix totalram_pages counting.
Gianluca Guida [Thu, 30 Jul 2009 21:54:36 +0000 (22:54 +0100)]
Xen balloon: fix totalram_pages counting.

commit 3d65c9488cadd2f11bd4d60c7266e639ece5d0d6 upstream.

Change totalram_pages when a single page is added/removed to the
ballooned list. This avoid totalram_pages to be set erroneously to
max_pfn at boot.

Signed-off-by: Gianluca Guida <gianluca.guida@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: explicitly create/destroy stop_machine workqueues outside suspend/resume region.
Ian Campbell [Tue, 1 Dec 2009 11:47:15 +0000 (11:47 +0000)]
xen: explicitly create/destroy stop_machine workqueues outside suspend/resume region.

commit b4606f2165153833247823e8c04c5e88cb3d298b upstream.

I have observed cases where the implicit stop_machine_destroy() done by
stop_machine() hangs while destroying the workqueues, specifically in
kthread_stop(). This seems to be because timer ticks are not restarted
until after stop_machine() returns.

Fortunately stop_machine provides a facility to pre-create/post-destroy
the workqueues so use this to ensure that workqueues are only destroyed
after everything is really up and running again.

I only actually observed this failure with 2.6.30. It seems that newer
kernels are somehow more robust against doing kthread_stop() without timer
interrupts (I tried some backports of some likely looking candidates but
did not track down the commit which added this robustness). However this
change seems like a reasonable belt&braces thing to do.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: use iret for return from 64b kernel to 32b usermode
Jeremy Fitzhardinge [Wed, 25 Nov 2009 21:15:38 +0000 (13:15 -0800)]
xen: use iret for return from 64b kernel to 32b usermode

commit 6aaf5d633bb6cead81b396d861d7bae4b9a0ba7e upstream.

If Xen wants to return to a 32b usermode with sysret it must use the
right form.  When using VCGF_in_syscall to trigger this, it looks at
the code segment and does a 32b sysret if it is FLAT_USER_CS32.
However, this is different from __USER32_CS, so it fails to return
properly if we use the normal Linux segment.

So avoid the whole mess by dropping VCGF_in_syscall and simply use
plain iret to return to usermode.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Acked-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: don't leak IRQs over suspend/resume.
Ian Campbell [Tue, 1 Dec 2009 16:15:30 +0000 (16:15 +0000)]
xen: don't leak IRQs over suspend/resume.

commit fed5ea87e02aaf902ff38c65b4514233db03dc09 upstream.

On resume irq_info[*].evtchn is reset to 0 since event channel mappings
are not preserved over suspend/resume. The other contents of irq_info
is preserved to allow rebind_evtchn_irq() to function.

However when a device resumes it will try to unbind from the
previous IRQ (e.g.  blkfront goes blkfront_resume() -> blkif_free() ->
unbind_from_irqhandler() -> unbind_from_irq()). This will fail due to the
check for VALID_EVTCHN in unbind_from_irq() and the IRQ is leaked. The
device will then continue to resume and allocate a new IRQ, eventually
leading to find_unbound_irq() panic()ing.

Fix this by changing unbind_from_irq() to handle teardown of interrupts
which have type!=IRQT_UNBOUND but are not currently bound to a specific
event channel.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: improve error handling in do_suspend.
Ian Campbell [Tue, 1 Dec 2009 11:47:14 +0000 (11:47 +0000)]
xen: improve error handling in do_suspend.

commit 65f63384b391bf4d384327d8a7c6de9860290b5c upstream.

The existing error handling has a few issues:
- If freeze_processes() fails it exits with shutting_down = SHUTDOWN_SUSPEND.
- If dpm_suspend_noirq() fails it exits without resuming xenbus.
- If stop_machine() fails it exits without resuming xenbus or calling
  dpm_resume_end().
- xs_suspend()/xs_resume() and dpm_suspend_noirq()/dpm_resume_noirq() were not
  nested in the obvious way.

Fix by ensuring each failure case goto's the correct label. Treat a failure of
stop_machine() as a cancelled suspend in order to follow the correct resume
path.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: call clock resume notifier on all CPUs
Ian Campbell [Wed, 25 Nov 2009 14:12:08 +0000 (14:12 +0000)]
xen: call clock resume notifier on all CPUs

commit f6eafe3665bcc374c66775d58312d1c06c55303f upstream.

tick_resume() is never called on secondary processors. Presumably this
is because they are offlined for suspend on native and so this is
normally taken care of in the CPU onlining path. Under Xen we keep all
CPUs online over a suspend.

This patch papers over the issue for me but I will investigate a more
generic, less hacky, way of doing to the same.

tick_suspend is also only called on the boot CPU which I presume should
be fixed too.

Signed-off-by: Ian Campbell <Ian.Campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: register runstate info for boot CPU early
Jeremy Fitzhardinge [Tue, 24 Nov 2009 17:38:25 +0000 (09:38 -0800)]
xen: register runstate info for boot CPU early

commit 499d19b82b586aef18727b9ae1437f8f37b66e91 upstream.

printk timestamping uses sched_clock, which in turn relies on runstate
info under Xen.  So make sure we set it up before any printks can
be called.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: don't call dpm_resume_noirq() with interrupts disabled.
Jeremy Fitzhardinge [Tue, 24 Nov 2009 17:58:49 +0000 (09:58 -0800)]
xen: don't call dpm_resume_noirq() with interrupts disabled.

commit 922cc38ab71d1360978e65207e4a4f4988987127 upstream.

dpm_resume_noirq() takes a mutex, so it can't be called from a no-interrupt
context.  Don't call it from within the stop-machine function, but just
afterwards, since we're resuming anyway, regardless of what happened.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: register runstate on secondary CPUs
Ian Campbell [Tue, 24 Nov 2009 17:32:48 +0000 (09:32 -0800)]
xen: register runstate on secondary CPUs

commit 028896721ac04f6fa0697f3ecac3f98761746363 upstream.

The commit "xen: re-register runstate area earlier on resume" caused us
to never try and setup the runstate area for secondary CPUs. Ensure that
we do this...

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: register timer interrupt with IRQF_TIMER
Ian Campbell [Tue, 24 Nov 2009 10:16:23 +0000 (10:16 +0000)]
xen: register timer interrupt with IRQF_TIMER

commit f350c7922faad3397c98c81a9e5658f5a1ef0214 upstream.

Otherwise the timer is disabled by dpm_suspend_noirq() which in turn prevents
correct operation of stop_machine on multi-processor systems and breaks
suspend.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: correctly restore pfn_to_mfn_list_list after resume
Ian Campbell [Sat, 21 Nov 2009 11:32:49 +0000 (11:32 +0000)]
xen: correctly restore pfn_to_mfn_list_list after resume

commit fa24ba62ea2869308ffc9f0b286ac9650b4ca6cb upstream.

pvops kernels >= 2.6.30 can currently only be saved and restored once. The
second attempt to save results in:

    ERROR Internal error: Frame# in pfn-to-mfn frame list is not in pseudophys
    ERROR Internal error: entry 0: p2m_frame_list[0] is 0xf2c2c2c2, max 0x120000
    ERROR Internal error: Failed to map/save the p2m frame list

I finally narrowed it down to:

    commit cdaead6b4e657f960d6d6f9f380e7dfeedc6a09b
        Author: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
        Date:   Fri Feb 27 15:34:59 2009 -0800

            xen: split construction of p2m mfn tables from registration

            Build the p2m_mfn_list_list early with the rest of the p2m table, but
            register it later when the real shared_info structure is in place.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
The unforeseen side-effect of this change was to cause the mfn list list to not
be rebuilt on resume. Prior to this change it would have been rebuilt via
xen_post_suspend() -> xen_setup_shared_info() -> xen_setup_mfn_list_list().

Fix by explicitly calling xen_build_mfn_list_list() from xen_post_suspend().

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: restore runstate_info even if !have_vcpu_info_placement
Jeremy Fitzhardinge [Sat, 21 Nov 2009 00:46:29 +0000 (08:46 +0800)]
xen: restore runstate_info even if !have_vcpu_info_placement

commit 3905bb2aa7bb801b31946b37a4635ebac4009051 upstream.

Even if have_vcpu_info_placement is not set, we still need to set up
the runstate area on each resumed vcpu.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen: re-register runstate area earlier on resume.
Ian Campbell [Sat, 21 Nov 2009 00:35:55 +0000 (08:35 +0800)]
xen: re-register runstate area earlier on resume.

commit be012920ecba161ad20303a3f6d9e96c58cf97c7 upstream.

This is necessary to ensure the runstate area is available to
xen_sched_clock before any calls to printk which will require it in
order to provide a timestamp.

I chose to pull the xen_setup_runstate_info out of xen_time_init into
the caller in order to maintain parity with calling
xen_setup_runstate_info separately from calling xen_time_resume.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoxen/xenbus: make DEVICE_ATTR()s static
Jeremy Fitzhardinge [Wed, 25 Nov 2009 00:41:47 +0000 (16:41 -0800)]
xen/xenbus: make DEVICE_ATTR()s static

commit db05fed0ad72f264e39bcb366795f7367384ec92 upstream.

They don't need to be global, and may cause linker clashes.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/i915: Add the missing clonemask for display port on Ironlake
Zhao Yakui [Wed, 2 Dec 2009 02:03:33 +0000 (10:03 +0800)]
drm/i915: Add the missing clonemask for display port on Ironlake

commit 652af9d74e1a3a10bb10f0d8e8f42ddac26bbc1a upstream.

Add the missing clonemask for display port on Ironlake.

Signed-off-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/i915: Set the error code after failing to insert new offset into mm ht.
Chris Wilson [Wed, 2 Dec 2009 15:15:30 +0000 (15:15 +0000)]
drm/i915: Set the error code after failing to insert new offset into mm ht.

commit 5618ca6abc2d6f475b258badc017a5254cf43d1b upstream.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/ttm: Fix build failure due to missing struct page
Martin Michlmayr [Thu, 19 Nov 2009 16:29:45 +0000 (16:29 +0000)]
drm/ttm: Fix build failure due to missing struct page

commit c3a73ba13bac7fd96030f39202b2d37fb19c46a6 upstream.

drm/ttm fails to build on MIPS because "struct page" is not known:
| In file included from drivers/gpu/drm/ttm/ttm_memory.c:28:
| include/drm/ttm/ttm_memory.h:154: warning: 'struct page' declared inside parameter list
| include/drm/ttm/ttm_memory.h:154: warning: its scope is only this definition or declaration, which is probably not what you want
| include/drm/ttm/ttm_memory.h:156: warning: 'struct page' declared inside parameter list
| drivers/gpu/drm/ttm/ttm_memory.c:540: error: conflicting types for 'ttm_mem_global_alloc_page'
| include/drm/ttm/ttm_memory.h:154: error: previous declaration of 'ttm_mem_global_alloc_page' was here
| drivers/gpu/drm/ttm/ttm_memory.c:561: error: conflicting types for 'ttm_mem_global_free_page'
| include/drm/ttm/ttm_memory.h:156: error: previous declaration of 'ttm_mem_global_free_page' was here

Signed-off-by: Martin Michlmayr <tbm@cyrius.com>
Acked-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/radeon/kms: rs6xx/rs740: clamp vram to aperture size
Alex Deucher [Thu, 3 Dec 2009 21:28:02 +0000 (16:28 -0500)]
drm/radeon/kms: rs6xx/rs740: clamp vram to aperture size

commit 0088dbdb809e8799cb8f26da5ac64b15201fa99d upstream.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/radeon/kms: fix vram setup on rs600
Alex Deucher [Thu, 3 Dec 2009 21:18:19 +0000 (16:18 -0500)]
drm/radeon/kms: fix vram setup on rs600

commit 722f29434e72188b2d20f9b41f4b5952073ed568 upstream.

also fix up rs690 mem width.

should fix fdo bug 25408

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/radeon/kms: fix legacy crtc2 dpms
Alex Deucher [Thu, 3 Dec 2009 17:15:54 +0000 (12:15 -0500)]
drm/radeon/kms: fix legacy crtc2 dpms

commit 8de21525439e6b5bb8d8c81e49094d867bf82f6d upstream.

noticed by Matthijs Kooijman on fdo bug 22140

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/radeon/kms: handle vblanks properly with dpms on
Alex Deucher [Wed, 2 Dec 2009 16:46:52 +0000 (11:46 -0500)]
drm/radeon/kms: handle vblanks properly with dpms on

commit 500b758725314ab1b5316eb0caa5b0fa26740e6b upstream.

avivo chips

Copied from pre-avivo code.

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodrm/radeon/kms: Add quirk for HIS X1300 board
Alex Deucher [Tue, 1 Dec 2009 19:49:50 +0000 (14:49 -0500)]
drm/radeon/kms: Add quirk for HIS X1300 board

commit 4e3f9b78ff917cc5c833858fdb5d96bc262e0bf3 upstream.

Board is DVI+VGA, not DVI+DVI

Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agopowerpc: Fix usage of 64-bit instruction in 32-bit altivec code
Benjamin Herrenschmidt [Tue, 8 Dec 2009 18:45:45 +0000 (18:45 +0000)]
powerpc: Fix usage of 64-bit instruction in 32-bit altivec code

commit e090aa80321b64c3b793f3b047e31ecf1af9538d upstream.

e821ea70f3b4873b50056a1e0f74befed1014c09 introduced a bug by copying
some 64-bit originated code as-is to be used by both 32 and 64-bit
but this code contains a 64-bit ony "cmpdi" instruction.

This changes it to cmpwi, which is fine since VRSAVE can only contains
a 32-bit value anyway.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agopowerpc/therm_adt746x: Record pwm invert bit at module load time]
Darrick J. Wong [Thu, 3 Dec 2009 16:19:59 +0000 (16:19 +0000)]
powerpc/therm_adt746x: Record pwm invert bit at module load time]

commit 1496e89ae2a0962748e55165a590fa3209c6f158 upstream.

In commit 0512a9a8e277a9de2820211eef964473b714ae65, we unilaterally zero the
"pwm invert" bit in the fan behavior configuration register.  On my PowerBook
G4, this results in the fans going to full speed at low temperature and
shutting off at high temperature because the pwm invert bit is supposed to be
set.

Therefore, record the pwm invert bit at driver load time, and write the bit
into the fan behavior control register.  This restores correct behavior on my
PBG4 and should work around the bit being set to the wrong value after
suspend/resume (which is what the original patch was trying to fix).  It also
fixes a minor omission where the pwm invert bit correction is NOT performed
when switching into automatic mode.

Signed-off-by: Darrick J. Wong <djwong@us.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agopowerpc/windfarm: Add detection for second cpu pump
Bolko Maass [Fri, 27 Nov 2009 05:44:33 +0000 (05:44 +0000)]
powerpc/windfarm: Add detection for second cpu pump

commit 529586dc39b0ec47c6290c4e7bed6ea3ffd1d8fb upstream.

Windfarm SMU control is explicitly missing support for a second CPU pump in G5 PowerMacs. Such machines actually exist  (specifically Quads with a second pump), so this patch adds detection for it.

Signed-off by: Bolko Maass <bmaass@math.uni-bremen.de>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomm: hugetlb: fix hugepage memory leak in walk_page_range()
Naoya Horiguchi [Tue, 15 Dec 2009 01:59:59 +0000 (17:59 -0800)]
mm: hugetlb: fix hugepage memory leak in walk_page_range()

commit d33b9f45bd24a6391bc05e2b5a13c1b5787ca9c2 upstream.

Most callers of pmd_none_or_clear_bad() check whether the target page is
in a hugepage or not, but walk_page_range() do not check it.  So if we
read /proc/pid/pagemap for the hugepage on x86 machine, the hugepage
memory is leaked as shown below.  This patch fixes it.

Details
=======
My test program (leak_pagemap) works as follows:
 - creat() and mmap() a file on hugetlbfs (file size is 200MB == 100 hugepages,)
 - read()/write() something on it,
 - call page-types with option -p (walk around the page tables),
 - munmap() and unlink() the file on hugetlbfs

Without my patches
------------------
$ cat /proc/meminfo |grep "HugePage"
HugePages_Total:    1000
HugePages_Free:     1000
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ./leak_pagemap
[snip output]
$ cat /proc/meminfo |grep "HugePage"
HugePages_Total:    1000
HugePages_Free:      900
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ls /hugetlbfs/
$

100 hugepages are accounted as used while there is no file on hugetlbfs.

With my patches
---------------
$ cat /proc/meminfo |grep "HugePage"
HugePages_Total:    1000
HugePages_Free:     1000
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ./leak_pagemap
[snip output]
$ cat /proc/meminfo |grep "HugePage"
HugePages_Total:    1000
HugePages_Free:     1000
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ls /hugetlbfs
$

No memory leaks.

Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Andy Whitcroft <apw@canonical.com>
Cc: David Rientjes <rientjes@google.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>
14 years agomm: hugetlb: fix hugepage memory leak in mincore()
Naoya Horiguchi [Tue, 15 Dec 2009 01:59:58 +0000 (17:59 -0800)]
mm: hugetlb: fix hugepage memory leak in mincore()

commit 4f16fc107d9c9b8a72aa19b189a9216e90a7aaef upstream.

Most callers of pmd_none_or_clear_bad() check whether the target page is
in a hugepage or not, but mincore() and walk_page_range() do not check it.
 So if we use mincore() on a hugepage on x86 machine, the hugepage memory
is leaked as shown below.  This patch fixes it by extending mincore()
system call to support hugepages.

Details
=======
My test program (leak_mincore) works as follows:
 - creat() and mmap() a file on hugetlbfs (file size is 200MB == 100 hugepages,)
 - read()/write() something on it,
 - call mincore() for first ten pages and printf() the values of *vec
 - munmap() and unlink() the file on hugetlbfs

Without my patch
----------------
$ cat /proc/meminfo| grep "HugePage"
HugePages_Total:    1000
HugePages_Free:     1000
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ./leak_mincore
vec[0] 0
vec[1] 0
vec[2] 0
vec[3] 0
vec[4] 0
vec[5] 0
vec[6] 0
vec[7] 0
vec[8] 0
vec[9] 0
$ cat /proc/meminfo |grep "HugePage"
HugePages_Total:    1000
HugePages_Free:      999
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ls /hugetlbfs/
$

Return values in *vec from mincore() are set to 0, while the hugepage
should be in memory, and 1 hugepage is still accounted as used while
there is no file on hugetlbfs.

With my patch
-------------
$ cat /proc/meminfo| grep "HugePage"
HugePages_Total:    1000
HugePages_Free:     1000
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ./leak_mincore
vec[0] 1
vec[1] 1
vec[2] 1
vec[3] 1
vec[4] 1
vec[5] 1
vec[6] 1
vec[7] 1
vec[8] 1
vec[9] 1
$ cat /proc/meminfo |grep "HugePage"
HugePages_Total:    1000
HugePages_Free:     1000
HugePages_Rsvd:        0
HugePages_Surp:        0
$ ls /hugetlbfs/
$

Return value in *vec set to 1 and no memory leaks.

[akpm@linux-foundation.org: cleanup]
[akpm@linux-foundation.org: build fix]
Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Cc: Hugh Dickins <hugh.dickins@tiscali.co.uk>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Cc: Andy Whitcroft <apw@canonical.com>
Cc: David Rientjes <rientjes@google.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>
14 years agox86: Fix bogus warning in apic_noop.apic_write()
Thomas Gleixner [Mon, 7 Dec 2009 11:59:46 +0000 (12:59 +0100)]
x86: Fix bogus warning in apic_noop.apic_write()

commit a946d8f11f0da9cfc714248036fcfd3a794d1e27 upstream.

apic_noop is used to provide dummy apic functions. It's installed
when the CPU has no APIC or when the APIC is disabled on the kernel
command line.

The apic_noop implementation of apic_write() warns when the CPU has
an APIC or when the APIC is not disabled.

That's bogus. The warning should only happen when the CPU has an
APIC _AND_ the APIC is not disabled. apic_noop.apic_read() has the
correct check.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
LKML-Reference: <alpine.LFD.2.00.0912071255420.3089@localhost.localdomain>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agortl8187: Fix wrong rfkill switch mask for some models
Larry Finger [Sun, 6 Dec 2009 01:25:22 +0000 (19:25 -0600)]
rtl8187: Fix wrong rfkill switch mask for some models

commit 70d57139f932b9ca21026253d02af71cf53d764a upstream.

There are different bits used to convey the setting of the rfkill
switch to the driver. The current driver only supports one of these
possibilities. These changes were derived from the latest version
of the vendor driver.

This patch fixes the regression noted in kernel Bugzilla #14743.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Reported-and-tested-by: Antti Kaijanmäki <antti@kaijanmaki.net>
Tested-by: Hin-Tak Leung <hintak.leung@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agowireless: correctly report signal value for IEEE80211_HW_SIGNAL_UNSPEC
John W. Linville [Tue, 8 Dec 2009 22:10:13 +0000 (17:10 -0500)]
wireless: correctly report signal value for IEEE80211_HW_SIGNAL_UNSPEC

commit 19deffbeba930030cfaf000b920333c6ba99ad52 upstream.

This part was missed in "cfg80211: implement get_wireless_stats",
probably because sta_set_sinfo already existed and was only handling
dBm signals.

Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomac80211: fix scan abort sanity checks
Johannes Berg [Sat, 31 Oct 2009 06:44:08 +0000 (07:44 +0100)]
mac80211: fix scan abort sanity checks

commit 6d3560d4fc9c5b9fe1a07a63926ea70512c69c32 upstream.

Since sometimes mac80211 queues up a scan request
to only act on it later, it must be allowed to
(internally) cancel a not-yet-running scan, e.g.
when the interface is taken down. This condition
was missing since we always checked only the
local->scanning variable which isn't yet set in
that situation.

Reported-by: Luis R. Rodriguez <mcgrof@gmail.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomac80211: Revert 'Use correct sign for mesh active path refresh'
Javier Cardona [Thu, 10 Dec 2009 02:43:01 +0000 (18:43 -0800)]
mac80211: Revert 'Use correct sign for mesh active path refresh'

commit 7b324d28a94dac5a451e8cba66e8d324601e5b9a upstream.

The patch ("mac80211: Use correct sign for mesh active path
refresh.") was actually a bug.  Reverted it and improved the
explanation of how mesh path refresh works.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: Andrey Yurovsky <andrey@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomac80211: Fixed bug in mesh portal paths
Javier Cardona [Thu, 10 Dec 2009 02:43:00 +0000 (18:43 -0800)]
mac80211: Fixed bug in mesh portal paths

commit 5d618cb81aeea19879975cd1f9a1e707694dfd7c upstream.

Paths to mesh portals were being timed out immediately after each use in
intermediate forwarding nodes.  mppath->exp_time is set to the expiration time
so assigning it to jiffies was marking the path as expired.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: Andrey Yurovsky <andrey@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomac80211: Fix bug in computing crc over dynamic IEs in beacon
Vasanthakumar Thiagarajan [Fri, 4 Dec 2009 12:11:34 +0000 (17:41 +0530)]
mac80211: Fix bug in computing crc over dynamic IEs in beacon

commit 1814077fd12a9cdf478c10076e9c42094e9d9250 upstream.

On a 32-bit machine, BIT() macro does not give the required
bit value if the bit is mroe than 31. In ieee802_11_parse_elems_crc(),
BIT() is suppossed to get the bit value more than 31 (42 (id of ERP_INFO_IE),
37 (CHANNEL_SWITCH_IE), (42), 32 (POWER_CONSTRAINT_IE), 45 (HT_CAP_IE),
61 (HT_INFO_IE)). As we do not get the required bit value for the above
IEs, crc over these IEs are never calculated, so any dynamic change in these
IEs after the association is not really handled on 32-bit platforms.
This patch fixes this issue.

Signed-off-by: Vasanthakumar Thiagarajan <vasanth@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoSerial: Do not read IIR in serial8250_start_tx when UART_BUG_TXEN
Ian Jackson [Wed, 18 Nov 2009 10:08:11 +0000 (11:08 +0100)]
Serial: Do not read IIR in serial8250_start_tx when UART_BUG_TXEN

commit 68cb4f8e246bbbc649980be0628cae9265870a91 upstream.

Do not read IIR in serial8250_start_tx when UART_BUG_TXEN

Reading the IIR clears some oustanding interrupts so it is not safe.
Instead, simply transmit immediately if the buffer is empty without
regard to IIR.

Signed-off-by: Ian Jackson <ian.jackson@eu.citrix.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Jiri Kosina <jkosina@suse.cz>
Cc: Alan Cox <alan@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoDriver core: fix race in dev_driver_string
Alan Stern [Fri, 4 Dec 2009 16:06:57 +0000 (11:06 -0500)]
Driver core: fix race in dev_driver_string

commit 3589972e51fac1e02d0aaa576fa47f568cb94d40 upstream.

This patch (as1310) works around a race in dev_driver_string().  If
the device is unbound while the function is running, dev->driver might
become NULL after we test it and before we dereference it.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Cc: Oliver Neukum <oliver@neukum.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodebugfs: fix create mutex racy fops and private data
Mathieu Desnoyers [Tue, 17 Nov 2009 22:40:26 +0000 (14:40 -0800)]
debugfs: fix create mutex racy fops and private data

commit d3a3b0adad0865c12e39b712ca89efbd0a3a0dbc upstream.

Setting fops and private data outside of the mutex at debugfs file
creation introduces a race where the files can be opened with the wrong
file operations and private data.  It is easy to trigger with a process
waiting on file creation notification.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agodevpts_get_tty() should validate inode
Sukadev Bhattiprolu [Wed, 18 Nov 2009 02:35:43 +0000 (18:35 -0800)]
devpts_get_tty() should validate inode

commit edfacdd6f81119b9005615593f2cbd94b8c7e2d8 upstream.

devpts_get_tty() assumes that the inode passed in is associated with a valid
pty.  But if the only reference to the pty is via a bind-mount, the inode
passed to devpts_get_tty() while valid, would refer to a pty that no longer
exists.

With a lot of debug effort, Grzegorz Nosek developed a small program (see
below) to reproduce a crash on recent kernels. This crash is a regression
introduced by the commit:

commit 527b3e4773628b30d03323a2cb5fb0d84441990f
Author: Sukadev Bhattiprolu <sukadev@us.ibm.com>
Date:   Mon Oct 13 10:43:08 2008 +0100

To fix, ensure that the dentry associated with the inode has not yet been
deleted/unhashed by devpts_pty_kill().

See also:
https://lists.linux-foundation.org/pipermail/containers/2009-July/019273.html

tty-bug.c:

#define _GNU_SOURCE
#include <fcntl.h>
#include <sched.h>
#include <stdlib.h>
#include <sys/mount.h>
#include <sys/signal.h>
#include <unistd.h>
#include <stdio.h>

#include <linux/fs.h>

void dummy(int sig)
{
}

static int child(void *unused)
{
int fd;

signal(SIGINT, dummy); signal(SIGHUP, dummy);
pause(); /* cheesy synchronisation to wait for /dev/pts/0 to appear */

mount("/dev/pts/0", "/dev/console", NULL, MS_BIND, NULL);
sleep(2);

fd = open("/dev/console", O_RDWR);
dup(0); dup(0);
write(1, "Hello world!\n", sizeof("Hello world!\n")-1);
return 0;
}

int main(void)
{
pid_t pid;
char *stack;

stack = malloc(16384);
pid = clone(child, stack+16384, CLONE_NEWNS|SIGCHLD, NULL);

open("/dev/ptmx", O_RDWR|O_NOCTTY|O_NONBLOCK);

unlockpt(fd); grantpt(fd);

sleep(2);
kill(pid, SIGHUP);
sleep(1);
return 0; /* exit before child opens /dev/console */
}

Reported-by: Grzegorz Nosek <root@localdomain.pl>
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Tested-by: Serge Hallyn <serue@us.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agofutex: Take mmap_sem for get_user_pages in fault_in_user_writeable
Andi Kleen [Tue, 8 Dec 2009 12:19:42 +0000 (13:19 +0100)]
futex: Take mmap_sem for get_user_pages in fault_in_user_writeable

commit 722d0172377a5697919b9f7e5beb95165b1dec4e upstream.

get_user_pages() must be called with mmap_sem held.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Andrew Morton <akpm@linuxfoundation.org>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Darren Hart <dvhltc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <20091208121942.GA21298@basil.fritz.box>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agomd/bitmap: protect against bitmap removal while being updated.
NeilBrown [Mon, 14 Dec 2009 01:49:46 +0000 (12:49 +1100)]
md/bitmap: protect against bitmap removal while being updated.

commit aa5cbd103887011b4830355f88fb055f9ad2d556 upstream.

A write intent bitmap can be removed from an array while the
array is active.
When this happens, all IO is suspended and flushed before the
bitmap is removed.
However it is possible that bitmap_daemon_work is still running to
clear old bits from the bitmap.  If it is, it can dereference the
bitmap after it has been freed.

So introduce a new mutex to protect bitmap_daemon_work and get it
before destroying a bitmap.

This is suitable for any current -stable kernel.

Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoNFS: Fix nfs_migrate_page()
Trond Myklebust [Thu, 10 Dec 2009 14:05:55 +0000 (09:05 -0500)]
NFS: Fix nfs_migrate_page()

commit 190f38e5cedc910940b1da9015f00458c18f97b4 upstream.

The call to migrate_page() will cause the page->private field to be
cleared.
Also fix up the locking around the page->private transfer, so that we ensure
that calls to nfs_page_find_request() don't end up racing.

Finally, fix up a double free bug: nfs_unlock_request() already calls
nfs_release_request() for us...

Reported-by: Wu Fengguang <fengguang.wu@intel.com>
Tested-by: Andi Kleen <andi@firstfloor.org>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoSUNRPC: IS_ERR/PTR_ERR confusion
Roel Kluin [Tue, 8 Dec 2009 18:13:03 +0000 (13:13 -0500)]
SUNRPC: IS_ERR/PTR_ERR confusion

commit 480e3243df156e39eea6c91057e2ae612a6bbe19 upstream.

IS_ERR returns 1 or 0, PTR_ERR returns the error value.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agohfs: fix a potential buffer overflow
Amerigo Wang [Tue, 15 Dec 2009 01:57:37 +0000 (17:57 -0800)]
hfs: fix a potential buffer overflow

commit ec81aecb29668ad71f699f4e7b96ec46691895b6 upstream.

A specially-crafted Hierarchical File System (HFS) filesystem could cause
a buffer overflow to occur in a process's kernel stack during a memcpy()
call within the hfs_bnode_read() function (at fs/hfs/bnode.c:24).  The
attacker can provide the source buffer and length, and the destination
buffer is a local variable of a fixed length.  This local variable (passed
as "&entry" from fs/hfs/dir.c:112 and allocated on line 60) is stored in
the stack frame of hfs_bnode_read()'s caller, which is hfs_readdir().
Because the hfs_readdir() function executes upon any attempt to read a
directory on the filesystem, it gets called whenever a user attempts to
inspect any filesystem contents.

[amwang@redhat.com: modify this patch and fix coding style problems]
Signed-off-by: WANG Cong <amwang@redhat.com>
Cc: Eugene Teo <eteo@redhat.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Dave Anderson <anderson@redhat.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>
14 years agopxa/em-x270: fix usb hub power up/reset sequence
Igor Grinberg [Sun, 6 Dec 2009 13:45:43 +0000 (15:45 +0200)]
pxa/em-x270: fix usb hub power up/reset sequence

commit 1b82e4c32fba96d8805b1e2126ba5382e56fac32 upstream.

Signed-off-by: Igor Grinberg <grinberg@compulab.co.il>
Signed-off-by: Mike Rapoport <mike@compulab.co.il>
Signed-off-by: Eric Miao <eric.y.miao@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: Close usb_find_interface race v3
Russ Dill [Tue, 15 Dec 2009 04:45:35 +0000 (21:45 -0700)]
USB: Close usb_find_interface race v3

commit c2d284ee04ab6f6718de2ddcf1b43160e046c41d upstream.

USB drivers that create character devices call usb_register_dev in their
probe function. This associates the usb_interface device with that minor
number and creates the character device and announces it to the world.
However, the driver's probe function is called before the new
usb_interface is added to the driver's klist_devices.

This is a problem because userspace will respond to the character device
creation announcement by opening the character device. The driver's open
function will the call usb_find_interface to find the usb_interface
associated with that minor number. usb_find_interface will walk the
driver's list of devices and find the usb_interface with the matching
minor number.

Because the announcement happens before the usb_interface is added to the
driver's klist_devices, a race condition exists. A straightforward fix
is to walk the list of devices on usb_bus_type instead since the device
is added to that list before the announcement occurs.

bus_find_device calls get_device to bump the reference count on the found
device. It is arguable that the reference count should be dropped by the
caller of usb_find_interface instead of usb_find_interface, however,
the current users of usb_find_interface do not expect this.

The original version of this patch only matched against minor number
instead of driver and minor number. This version matches against both.

Signed-off-by: Russ Dill <Russ.Dill@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: usb-storage: add BAD_SENSE flag
Alan Stern [Mon, 7 Dec 2009 21:39:16 +0000 (16:39 -0500)]
USB: usb-storage: add BAD_SENSE flag

commit a0bb108112a872c0b0c4b3ef4974f95fb75b155d upstream.

This patch (as1311) fixes a problem in usb-storage: Some devices are
pretty broken when it comes to reporting sense data.  The information
they send back indicates that they have more than 18 bytes of sense
data available, but when the system asks for more than 18 they fail or
hang.  The symptom is that probing fails with multiple resets.

The patch adds a new BAD_SENSE flag to indicate that usb-storage
should never ask for more than 18 bytes of sense data.  The flag can
be set in an unusual_devs entry or via the "quirks=" module parameter,
and it is set automatically whenever a REQUEST SENSE command for more
than 18 bytes fails or times out.

An unusual_devs entry is added for the Agfa photo frame, which uses a
Prolific chip having this bug.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Daniel Kukula <daniel.kuku@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: usbtmc: repeat usb_bulk_msg until whole message is transfered
Andre Herms [Thu, 19 Nov 2009 17:14:49 +0000 (18:14 +0100)]
USB: usbtmc: repeat usb_bulk_msg until whole message is transfered

commit ec412b92dbe3ea839716853eea058d1bcc5e6ca4 upstream.

usb_bulk_msg() transfers only bytes up to the maximum packet size.
It must be repeated by the usbtmc driver until all bytes of a TMC message
are transfered.

Without this patch, ETIMEDOUT is reported when writing TMC messages
larger than the maximum USB bulk size and the transfer remains incomplete.
The user will notice that the device hangs and must be reset by either closing
the application or pulling the plug.

Signed-off-by: Andre Herms <andre.herms@tec-venture.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: option.c: add support for D-Link DWM-162-U5
Zhang Le [Tue, 17 Nov 2009 22:53:42 +0000 (14:53 -0800)]
USB: option.c: add support for D-Link DWM-162-U5

commit 54a8e144acad6506920f385f4ef2779664f05b21 upstream.

Add D-Link DWM-162-U5 device id 1e0e:ce16 into option driver.  The device
has 4 interfaces, of which 1 is handled by storage and the other 3 by
option driver.

The device appears first as CD-only 05c6:2100 device and must be switched
to 1e0e:ce16 mode either by using "eject CD" or usb_modeswitch.

The MessageContent for usb_modeswitch.conf is:
"55534243e0c26a85000000000000061b000000020000000000000000000000"

Signed-off-by: Zhang Le <r0bertz@gentoo.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: musb_gadget_ep0: fix unhandled endpoint 0 IRQs, again
Sergei Shtylyov [Mon, 16 Nov 2009 10:54:05 +0000 (16:24 +0530)]
USB: musb_gadget_ep0: fix unhandled endpoint 0 IRQs, again

commit 196f1b7a387546f425df2f1fad26772e3d513aea upstream.

Commit a5073b52833e4df8e16c93dc4cbb7e0c558c74a2 (musb_gadget: fix
unhandled endpoint 0 IRQs) somehow missed its key change:

"The gadget EP0 code routinely ignores an interrupt at end of
the data phase because of musb_g_ep0_giveback() resetting the
state machine to "idle, waiting for SETUP" phase prematurely."

So, the majority of the cases of unhandled IRQs is still unfixed...

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUSB: xhci: Add correct email and files to MAINTAINERS entry.
Sarah Sharp [Tue, 1 Dec 2009 18:37:07 +0000 (10:37 -0800)]
USB: xhci: Add correct email and files to MAINTAINERS entry.

commit 36d0344c254a7b333272757f858c403ea3a2d29f upstream.

Add the xHCI driver files to its MAINTAINERS entry so that I'm Cc'd on
cleanup patches.  Update the email address to one I actually use for
sending patches and responding to Linux mailing list emails.

Signed-off-by: Sarah Sharp <sarah.a.sharp@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agojbd2: don't wipe the journal on a failed journal checksum
Theodore Ts'o [Sun, 15 Nov 2009 20:31:37 +0000 (15:31 -0500)]
jbd2: don't wipe the journal on a failed journal checksum

commit e6a47428de84e19fda52f21ab73fde2906c40d09 upstream.

If there is a failed journal checksum, don't reset the journal.  This
allows for userspace programs to decide how to recover from this
situation.  It may be that ignoring the journal checksum failure might
be a better way of recovering the file system.  Once we add per-block
checksums, we can definitely do better.  Until then, a system
administrator can try backing up the file system image (or taking a
snapshot) and and trying to determine experimentally whether ignoring
the checksum failure or aborting the journal replay results in less
data loss.

Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoUBI: flush wl before clearing update marker
Sebastian Andrzej Siewior [Sun, 29 Nov 2009 18:46:02 +0000 (19:46 +0100)]
UBI: flush wl before clearing update marker

commit 6afaf8a484cbbfd2ccf58a4e5396d1f280469789 upstream.

ubiupdatevol -t does the following:
- ubi_start_update()
  - set_update_marker()
  - for all LEBs ubi_eba_unmap_leb()
  - clear_update_marker()
  - ubi_wl_flush()

ubi_wl_flush() physically erases all PEB, once it returns all PEBs are
empty. clear_update_marker() has the update marker written after return.
If there is a power cut between the last two functions then the UBI
volume has no longer the "update" marker set and may have some valid
LEBs while some of them may be gone.
If that volume in question happens to be a UBIFS volume, then mount
will fail with

|UBIFS error (pid 1361): ubifs_read_node: bad node type (255 but expected 6)
|UBIFS error (pid 1361): ubifs_read_node: bad node at LEB 0:0
|Not a node, first 24 bytes:
|00000000: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

if there is at least one valid LEB and the wear-leveling worker managed
to clear LEB 0.

The patch waits for the wl worker to finish prior clearing the "update"
marker on flash. The two new LEB which are scheduled for erasing after
clear_update_marker() should not matter because they are only visible to
UBI.

Signed-off-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agobsdacct: fix uid/gid misreporting
Alexey Dobriyan [Tue, 15 Dec 2009 01:57:34 +0000 (17:57 -0800)]
bsdacct: fix uid/gid misreporting

commit 4b731d50ff3df6b9141a6c12b088e8eb0109e83c upstream.

commit d8e180dcd5bbbab9cd3ff2e779efcf70692ef541 "bsdacct: switch
credentials for writing to the accounting file" introduced credential
switching during final acct data collecting.  However, uid/gid pair
continued to be collected from current which became credentials of who
created acct file, not who exits.

Addresses http://bugzilla.kernel.org/show_bug.cgi?id=14676

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Reported-by: Juho K. Juopperi <jkj@kapsi.fi>
Acked-by: Serge Hallyn <serue@us.ibm.com>
Acked-by: David Howells <dhowells@redhat.com>
Reviewed-by: Michal Schmidt <mschmidt@redhat.com>
Cc: James Morris <jmorris@namei.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>
14 years agoV4L/DVB: Fix test in copy_reg_bits()
Roel Kluin [Fri, 20 Nov 2009 18:34:13 +0000 (15:34 -0300)]
V4L/DVB: Fix test in copy_reg_bits()

commit c95a419a5604ec8a23cd73f61e9bb151e8cbe89b upstream.

The reg_pair2[j].reg was tested twice.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Acked-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agos390: clear high-order bits of registers after sam64
Hendrik Brueckner [Mon, 7 Dec 2009 11:44:42 +0000 (12:44 +0100)]
s390: clear high-order bits of registers after sam64

commit cf87b7439ec81b9374e7772e44e9cb2eb9e57160 upstream.

When the kernel is IPLed without the CLEAR option and switches
to 64-bit, the high-order half of the registers might contain
random values.  This can cause addressing exceptions and the
kernel enters an interrupt loop.

Initialize the high-order half of the general purpose registers
with zeros after switching to 64-bit mode.

Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agopata_hpt{37x|3x2n}: fix timing register masks (take 2)
Sergei Shtylyov [Fri, 27 Nov 2009 18:29:02 +0000 (22:29 +0400)]
pata_hpt{37x|3x2n}: fix timing register masks (take 2)

commit 5600c70e576199a7552e1c0fff43f3fe16f5566e upstream.

These drivers inherited from the older 'hpt366' IDE driver the buggy timing
register masks in their set_piomode() metods. As a result, too low command
cycle active time is programmed for slow PIO modes.  Quite fortunately, it's
later "fixed up" by the set_dmamode() methods which also "helpfully" reprogram
the command timings, usually to PIO mode 4; unfortunately, setting an UltraDMA
mode #N also reprograms already set PIO data timings, usually to MWDMA mode #
max(N, 2) timings...

However, the drivers added some breakage of their own too:  the bit that they
set/clear to control the FIFO is sometimes wrong -- it's actually the MSB of
the command cycle setup time; also, setting it in DMA mode is wrong as this
bit is only for PIO actually and clearing it for PIO modes is not needed as
no mode in any timing table has it set...

Fix all this, inverting the masks while at it, like in the 'hpt366' and
'pata_hpt366' drivers; bump the drivers' versions, accounting for recent
patches that forgot to do it...

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86: Fix typo in Intel CPU cache size descriptor
Dave Jones [Tue, 10 Nov 2009 20:01:20 +0000 (15:01 -0500)]
x86: Fix typo in Intel CPU cache size descriptor

commit e02e0e1a130b9ca37c5186d38ad4b3aaf58bb149 upstream.

I double-checked the datasheet. One of the existing
descriptors has a typo: it should be 2MB not 2038 KB.

Signed-off-by: Dave Jones <davej@redhat.com>
LKML-Reference: <20091110200120.GA27090@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86: Add new Intel CPU cache size descriptors
Dave Jones [Tue, 10 Nov 2009 18:49:24 +0000 (13:49 -0500)]
x86: Add new Intel CPU cache size descriptors

commit 85160b92fbd35321104819283c91bfed2b553e3c upstream.

The latest rev of Intel doc AP-485 details new cache descriptors
that we don't yet support. 12MB, 18MB and 24MB 24-way assoc L3
caches.

Signed-off-by: Dave Jones <davej@redhat.com>
LKML-Reference: <20091110184924.GA20337@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86: Fix duplicated UV BAU interrupt vector
Cliff Wickman [Fri, 11 Dec 2009 17:36:18 +0000 (11:36 -0600)]
x86: Fix duplicated UV BAU interrupt vector

commit 1d865fb728bd6bbcdfbd6ec1e2b8ade3b4805641 upstream.

Interrupt vector 0xec has been doubly defined in irq_vectors.h

It seems arbitrary whether LOCAL_PENDING_VECTOR or
UV_BAU_MESSAGE is the higher number.  As long as they are
unique. If they are not unique we'll hit a BUG in
alloc_system_vector().

Signed-off-by: Cliff Wickman <cpw@sgi.com>
LKML-Reference: <E1NJ9Pe-0004P7-0Q@eag09.americas.sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86: SGI UV: Fix BAU initialization
Cliff Wickman [Thu, 19 Nov 2009 23:12:43 +0000 (17:12 -0600)]
x86: SGI UV: Fix BAU initialization

commit e38e2af1c57c3eb5211331a5b4fcaae0c4a2a918 upstream.

A memory mapped register that affects the SGI UV Broadcast
Assist Unit's interrupt handling may sometimes be unintialized.

Remove the condition on its initialization, as that condition
can be randomly satisfied by a hardware reset.

Signed-off-by: Cliff Wickman <cpw@sgi.com>
LKML-Reference: <E1NBGB9-0005nU-Dp@eag09.americas.sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86/mce: Set up timer unconditionally
Jan Beulich [Tue, 8 Dec 2009 02:21:37 +0000 (11:21 +0900)]
x86/mce: Set up timer unconditionally

commit bc09effabf0c5c6c7021e5ef9af15a23579b32a8 upstream.

mce_timer must be passed to setup_timer() in all cases, no
matter whether it is going to be actually used. Otherwise, when
the CPU gets brought down, its call to del_timer_sync() will
never return, as the timer won't have a base associated, and
hence lock_timer_base() will loop infinitely.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
LKML-Reference: <4B1DB831.2030801@jp.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86, mce: don't restart timer if disabled
Hidetoshi Seto [Thu, 3 Dec 2009 02:33:08 +0000 (11:33 +0900)]
x86, mce: don't restart timer if disabled

commit fe5ed91ddce85a0ed0e4f92c10b099873ef62167 upstream.

Even it is in error path unlikely taken, add_timer_on() at
CPU_DOWN_FAILED* needs to be skipped if mce_timer is disabled.

Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Jan Beulich <jbeulich@novell.com>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86: Use -maccumulate-outgoing-args for sane mcount prologues
Thomas Gleixner [Sat, 28 Nov 2009 14:03:03 +0000 (15:03 +0100)]
x86: Use -maccumulate-outgoing-args for sane mcount prologues

commit b8b7d791a8ff01d2380089279a69afa99115fb23 upstream.

commit 746357d (x86: Prevent GCC 4.4.x (pentium-mmx et al) function
prologue wreckage) uses -mtune=generic to work around the function
prologue problem with mcount on -march=pentium-mmx and others.

Jakub pointed out that we can use -maccumulate-outgoing-args instead
which is selected by -mtune=generic and prevents the problem without
losing the -march specific optimizations.

Pointed-out-by: Jakub Jelinek <jakub@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agox86: Prevent GCC 4.4.x (pentium-mmx et al) function prologue wreckage
Thomas Gleixner [Fri, 20 Nov 2009 11:01:43 +0000 (12:01 +0100)]
x86: Prevent GCC 4.4.x (pentium-mmx et al) function prologue wreckage

commit 746357d6a526d6da9d89a2ec645b28406e959c2e upstream.

When the kernel is compiled with -pg for tracing GCC 4.4.x inserts
stack alignment of a function _before_ the mcount prologue if the
-march=pentium-mmx is set and -mtune=generic is not set. This breaks
the assumption of the function graph tracer which expects that the
mcount prologue

       push %ebp
       mov  %esp, %ebp

is the first stack operation in a function because it needs to modify
the function return address on the stack to trap into the tracer
before returning to the real caller.

The generated code is:

        push   %edi
        lea    0x8(%esp),%edi
        and    $0xfffffff0,%esp
        pushl  -0x4(%edi)
        push   %ebp
        mov    %esp,%ebp

so the tracer modifies the copy of the return address which is stored
after the stack alignment and therefor does not trap the return which
in turn breaks the call chain logic of the tracer and leads to a
kernel panic.

Aside of the fact that the generated code is horrible for no good
reason other -march -mtune options generate the expected:

        push   %ebp
        mov    %esp,%ebp
        and    $0xfffffff0,%esp

which does the same and keeps everything intact.

After some experimenting we found out that this problem is restricted
to gcc4.4.x and to the following -march settings:

i586, pentium, pentium-mmx, k6, k6-2, k6-3, winchip-c6, winchip2, c3,
geode

By adding -mtune=generic the code generator produces always the
expected code.

So forcing -mtune=generic when CONFIG_FUNCTION_GRAPH_TRACER=y is not
pretty, but at the moment the only way to prevent that the kernel
trips over gcc-shrooms induced code madness.

Most distro kernels have CONFIG_X86_GENERIC=y anyway which forces
-mtune=generic as well so it will not impact those.

References: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=42109
    http://lkml.org/lkml/2009/11/19/17

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
LKML-Reference: <alpine.LFD.2.00.0911200206570.24119@localhost.localdomain>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>,
Cc: Jeff Law <law@redhat.com>
Cc: gcc@gcc.gnu.org
Cc: David Daney <ddaney@caviumnetworks.com>
Cc: Andrew Haley <aph@redhat.com>
Cc: Richard Guenther <richard.guenther@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoKVM: x86: include pvclock MSRs in msrs_to_save
Glauber Costa [Tue, 6 Oct 2009 17:24:50 +0000 (13:24 -0400)]
KVM: x86: include pvclock MSRs in msrs_to_save

commit e3267cbbbfbcbe9c18833e89b10beabb1117cb55 upstream.

For a while now, we are issuing a rdmsr instruction to find out which
msrs in our save list are really supported by the underlying machine.
However, it fails to account for kvm-specific msrs, such as the pvclock
ones.

This patch moves then to the beginning of the list, and skip testing them.

Signed-off-by: Glauber Costa <glommer@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoKVM: fix irq_source_id size verification
Marcelo Tosatti [Sun, 18 Oct 2009 01:47:23 +0000 (22:47 -0300)]
KVM: fix irq_source_id size verification

commit cd5a2685de4a642fd0bd763e8c19711ef08dbe27 upstream.

find_first_zero_bit works with bit numbers, not bytes.

Fixes

https://sourceforge.net/tracker/?func=detail&aid=2847560&group_id=180599&atid=893831

Reported-by: "Xu, Jiajun" <jiajun.xu@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoKVM: s390: Make psw available on all exits, not just a subset
Carsten Otte [Thu, 19 Nov 2009 13:21:16 +0000 (14:21 +0100)]
KVM: s390: Make psw available on all exits, not just a subset

commit d7b0b5eb3000c6fb902f08c619fcd673a23d8fab upstream.

This patch moves s390 processor status word into the base kvm_run
struct and keeps it up-to date on all userspace exits.

The userspace ABI is broken by this, however there are no applications
in the wild using this.  A capability check is provided so users can
verify the updated API exists.

Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoKVM: s390: Fix prefix register checking in arch/s390/kvm/sigp.c
Carsten Otte [Mon, 30 Nov 2009 16:14:41 +0000 (17:14 +0100)]
KVM: s390: Fix prefix register checking in arch/s390/kvm/sigp.c

commit f50146bd7bdb75435638e60d4960edd9bcdf88b8 upstream.

This patch corrects the checking of the new address for the prefix register.
On s390, the prefix register is used to address the cpu's lowcore (address
0...8k). This check is supposed to verify that the memory is readable and
present.
copy_from_guest is a helper function, that can be used to read from guest
memory. It applies prefixing, adds the start address of the guest memory in
user, and then calls copy_from_user. Previous code was obviously broken for
two reasons:
- prefixing should not be applied here. The current prefix register is
  going to be updated soon, and the address we're looking for will be
  0..8k after we've updated the register
- we're adding the guest origin (gmsor) twice: once in subject code
  and once in copy_from_guest

With kuli, we did not hit this problem because (a) we were lucky with
previous prefix register content, and (b) our guest memory was mmaped
very low into user address space.

Signed-off-by: Carsten Otte <cotte@de.ibm.com>
Reported-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoKVM: x86 emulator: limit instructions to 15 bytes
Avi Kivity [Tue, 24 Nov 2009 13:20:15 +0000 (15:20 +0200)]
KVM: x86 emulator: limit instructions to 15 bytes

commit eb3c79e64a70fb8f7473e30fa07e89c1ecc2c9bb upstream.

While we are never normally passed an instruction that exceeds 15 bytes,
smp games can cause us to attempt to interpret one, which will cause
large latencies in non-preempt hosts.

Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoALSA: hda - Terradici HDA controllers does not support 64-bit mode
Jaroslav Kysela [Wed, 9 Dec 2009 09:44:47 +0000 (10:44 +0100)]
ALSA: hda - Terradici HDA controllers does not support 64-bit mode

commit 396087eaead95fcb29eb36f1e59517aeb58c545e upstream.

Confirmed from vendor and tests in RedHat bugzilla #536782 .

Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoALSA: hrtimer - Fix lock-up
Takashi Iwai [Fri, 11 Dec 2009 11:51:05 +0000 (12:51 +0100)]
ALSA: hrtimer - Fix lock-up

commit fcfdebe70759c74e2e701f69aaa7f0e5e32cf5a6 upstream.

The timer stop callback can be called from snd_timer_interrupt(), which
is called from the hrtimer callback.  Since hrtimer_cancel() waits for
the callback completion, this eventually results in a lock-up.

This patch fixes the problem by just toggling a flag at stop callback
and call hrtimer_cancel() later.

Reported-and-tested-by: Wojtek Zabolotny <W.Zabolotny@elka.pw.edu.pl>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agohrtimer: Fix /proc/timer_list regression
Feng Tang [Thu, 3 Sep 2009 08:32:53 +0000 (16:32 +0800)]
hrtimer: Fix /proc/timer_list regression

commit 8629ea2eaba8ca0de2e38ce1b4a825e16255976e upstream.

commit 507e1231 (timer stats: Optimize by adding quick check to avoid
function calls) introduced a regression in /proc/timer_list.

/proc/timer_list shows now
 #0: <c27d46b0>, tick_sched_timer, S:01, <(null)>, /-1
instead of
 #0: <c27d46b0>, tick_sched_timer, S:01, hrtimer_start, swapper/0

Revert the hrtimer quick check for now. The optimization needs more
thought, but this is neither 2.6.32-rc7 nor stable material.

[ tglx: - Removed unrelated changes from the original patch
   - Prevent unneccesary call to timer_stats_update_stats
- massaged the changelog ]

Signed-off-by: Feng Tang <feng.tang@intel.com>
LKML-Reference: <alpine.LFD.2.00.0911181933540.24119@localhost.localdomain>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoath5k: enable EEPROM checksum check
Luis R. Rodriguez [Wed, 25 Nov 2009 22:23:26 +0000 (17:23 -0500)]
ath5k: enable EEPROM checksum check

commit 512414b0bed0d376ac4d5ec1dd6f0b1a3551febc upstream.

Without this we have no gaurantee of the integrity of the
EEPROM and are likely to encounter a lot of bogus bug reports
due to actual issues on the EEPROM. With the EEPROM checksum
check in place we can easily rule those issues out.

If you run patch during a revert *you* have a card with a busted
EEPROM and only older kernel will support that concoction. This
patch is a trade off between not accepitng bogus EEPROMs and
avoiding bogus bug reports allowing developers to focus instead
on real concrete issues.

If stable keeps bogus bug reports because of a possibly busted EEPROM
feel free to apply this there too.

Tested on an AR5414

Cc: jirislaby@gmail.com
Cc: akpm@linux-foundation.org
Cc: rjw@sisk.pl
Cc: me@bobcopeland.com
Cc: david.quan@atheros.com
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agoath5k: allow setting txpower to 0
Bob Copeland [Mon, 16 Nov 2009 13:30:29 +0000 (08:30 -0500)]
ath5k: allow setting txpower to 0

commit 2eb2fa67e5462a36e98172fb92c78bc405b3035f upstream.

As a holdover from earlier code when we used to set
the power limit to '0' after a reset to configure the
default transmit power, ath5k interprets txpower=0 as
12.5 dBm.  Fix that by just passing 0 through.

This fixes http://bugzilla.kernel.org/show_bug.cgi?id=14567

Reported-by: Daniel Folkers <daniel.folkers@task24.nl>
Tested-by: Daniel Folkers <daniel.folkers@task24.nl>
Signed-off-by: Bob Copeland <me@bobcopeland.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
14 years agossb: Fix range check in sprom write
Michael Buesch [Mon, 23 Nov 2009 19:58:06 +0000 (20:58 +0100)]
ssb: Fix range check in sprom write

commit e33761e6f23881de9f3ee77cc2204ab2e26f3d9a upstream.

The range check in the sprom image parser hex2sprom() is broken.
One sprom word is 4 hex characters.
This fixes the check and also adds much better sanity checks to the code.
We better make sure the image is OK by doing some sanity checks to avoid
bricking the device by accident.

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>