safe/jmp/linux-2.6
17 years ago[PATCH] uml: fix device unplug crash
Jeff Dike [Thu, 29 Mar 2007 08:20:28 +0000 (01:20 -0700)]
[PATCH] uml: fix device unplug crash

Fix a NULL dereference when unplugging a device.  The default value of
err_msg wants to be "" in case the driver doesn't modify it.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: fix I/O hang when multiple devices are in use
Jeff Dike [Thu, 29 Mar 2007 08:20:27 +0000 (01:20 -0700)]
[PATCH] uml: fix I/O hang when multiple devices are in use

Commit 62f96cb01e8de7a5daee472e540f726db2801499 introduced per-devices
queues and locks, which was fine as far as it went, but left in place a
global which controlled access to submitting requests to the host.  This
should have been made per-device as well, since it causes I/O hangs when
multiple block devices are in use.

This patch fixes that by replacing the global with an activity flag in the
device structure in order to tell whether the queue is currently being run.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] bluetooth hid quirks: mightymouse quirk
Jiri Kosina [Thu, 29 Mar 2007 08:20:25 +0000 (01:20 -0700)]
[PATCH] bluetooth hid quirks: mightymouse quirk

I have a bugreport that scrollwheel of bluetooth version of apple
mightymouse doesn't work.  The USB version of mightymouse works, as there
is a quirk for handling scrollwheel in hid/usbhid for it.

Now that bluetooth git tree is hooked to generic hid layer, it could easily
use the quirks which are already present in generic hid parser, hid-input,
etc.

Below is a simple patch against bluetooth git tree, which adds quirk
handling to current bluetooth hidp code, and sets quirk flags for device
0x05ac/0x030c, which is the bluetooth version of the apple mightymouse.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] adaptec MAINTAINERS updates
Salyzyn, Mark [Thu, 29 Mar 2007 08:20:24 +0000 (01:20 -0700)]
[PATCH] adaptec MAINTAINERS updates

The IPS and DPT_I2O drivers are missing, so here is my 'hand coded'
addition to deal with overlap to the patch below (apply both). I
selected Maintained rather than supported for the ips and dpt_i2o driver
due to their legacy nature.

Cc: James Bottomley <James.Bottomley@steeleye.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] pid: Properly detect orphaned process groups in exit_notify
Eric W. Biederman [Thu, 29 Mar 2007 05:38:16 +0000 (23:38 -0600)]
[PATCH] pid: Properly detect orphaned process groups in exit_notify

In commit 0475ac0845f9295bc5f69af45f58dff2c104c8d1 when converting the
orphaned process group handling to use struct pid I made a small
mistake.  I accidentally replaced an == with a !=.

Besides just being a dumb thing to do apparently this has a bad side
effect.  The improper orphaned process group detection causes kwin to
die after a suspend/resume cycle.

I'm amazed this patch has been around as long as it has without anyone
else noticing something funny going on.

And the following people deserve credit for spotting and helping
to reproduce this.

Thanks to: Sid Boyce <g3vbv@blueyonder.co.uk>
Thanks to: "Michael Wu"

Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] x86_64 irq: Fix comments after changing IRQ0_VECTOR from 0x20 to 0x30
Yinghai Lu [Thu, 29 Mar 2007 05:10:29 +0000 (23:10 -0600)]
[PATCH] x86_64 irq: Fix comments after changing IRQ0_VECTOR from 0x20 to 0x30

Signed-off-by: Yinghai Lu <yinghai.lu@amd.com>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[POWERPC] Fix dcr_unmap()
Jean-Christophe Dubois [Wed, 28 Mar 2007 10:07:41 +0000 (11:07 +0100)]
[POWERPC] Fix dcr_unmap()

Fix a bug in dcr_unmap().

At unmap time the DCR offset need to be added instead of substracted.

Signed-off-by: Jean-Christophe Dubois <jdubois@mc.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
17 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfashe...
Linus Torvalds [Wed, 28 Mar 2007 21:02:03 +0000 (14:02 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/mfasheh/ocfs2

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfasheh/ocfs2:
  ocfs2_dlm: Check for migrateable lockres in dlm_empty_lockres()
  ocfs2_dlm: Fix lockres ref counting bug

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Wed, 28 Mar 2007 21:01:21 +0000 (14:01 -0700)]
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [VIDEO]: Fix section mismatch in cg3.c
  [SPARC]: sparc64 gcc-4.2.0 20070317 -Werror failure
  [VIDEO] ffb: Fix two DAC handling bugs.
  [SPARC32]: Fix SMP build regression
  [DRM]: Delete sparc64 FFB driver code that never gets built.

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 28 Mar 2007 21:00:27 +0000 (14:00 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [DCCP] getsockopt: Fix DCCP_SOCKOPT_[SEND,RECV]_CSCOV

17 years agoMerge branch 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Wed, 28 Mar 2007 21:00:01 +0000 (14:00 -0700)]
Merge branch 'for-linus' of /linux/kernel/git/roland/infiniband

* 'for-linus' of master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband:
  IB/iser: Handle aborting a command after it is sent
  IB/mthca: Fix thinko in init_mr_table()
  RDMA/cxgb3: Fix resource leak in cxio_hal_init_ctrl_qp()

17 years ago[PATCH] MSI-X: fix resume crash
Eric W. Biederman [Wed, 28 Mar 2007 13:36:09 +0000 (15:36 +0200)]
[PATCH] MSI-X: fix resume crash

So I think the right solution is to simply make pci_enable_device just
flip enable bits and move the rest of the work someplace else.

However a thorough cleanup is a little extreme for this point in the
release cycle, so I think a quick hack that makes the code not stomp the
irq when msi irq's are enabled should be the first fix.  Then we can
later make the code not change the irqs at all.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] oprofile: fix potential deadlock on oprofilefs_lock
Jiri Kosina [Wed, 28 Mar 2007 16:12:34 +0000 (18:12 +0200)]
[PATCH] oprofile: fix potential deadlock on oprofilefs_lock

nmi_cpu_setup() is called from hardirq context and acquires oprofilefs_lock.
alloc_event_buffer() and oprofilefs_ulong_from_user() acquire this lock
without disabling irqs, which could deadlock.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6
Linus Torvalds [Wed, 28 Mar 2007 20:53:40 +0000 (13:53 -0700)]
Merge /pub/scm/linux/kernel/git/lethal/sh-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/lethal/sh-2.6:
  sh: Trivial fix for hp6xx build.
  sh: Fixup __cmpxchg() compile breakage with gcc4.
  sh: Kill bogus GCC4 symbol exports.

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Wed, 28 Mar 2007 20:46:00 +0000 (13:46 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  ata: NCQ is broken on Maxtor 6L250S0
  pata_pdc202xx_old: LBA48 bug
  libata: IDENTIFY backwards for drive side cable detection
  ahci.c: walkaround for SB600 SATA internal error issue
  [libata] Disable ACPI by default; fix namespace problems

17 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Wed, 28 Mar 2007 20:45:13 +0000 (13:45 -0700)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/netdev-2.6

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/netdev-2.6:
  SUN3/3X Lance trivial fix improved
  mv643xx_eth: Fix use of uninitialized port_num field
  forcedeth: fix tx timeout
  forcedeth: fix nic poll
  qla3xxx: bugfix: Jumbo frame handling.
  qla3xxx: bugfix: Dropping interrupt under heavy network load.
  qla3xxx: bugfix: Multi segment sends were getting whacked.
  qla3xxx: bugfix: Add tx control block memset.
  atl1: remove unnecessary crc inversion
  myri10ge: correctly detect when TSO should be used
  [PATCH] WE-22 : prevent information leak on 64 bit
  [PATCH] wext: Add missing ioctls to 64<->32 conversion
  [PATCH] bcm43xx: Fix machine check on PPC for version 1 PHY
  [PATCH] bcm43xx: fix radio_set_tx_iq
  [PATCH] bcm43xx: Fix code for confusion between PHY revision and PHY version

17 years ago[PATCH] hrtimers: fix reprogramming SMP race
Ingo Molnar [Wed, 28 Mar 2007 11:17:18 +0000 (13:17 +0200)]
[PATCH] hrtimers: fix reprogramming SMP race

hrtimer_start() incorrectly set the 'reprogram' flag to enqueue_hrtimer(),
which should only be 1 if the hrtimer is queued to the current CPU.

Doing otherwise could result in a reprogramming of the current CPU's
clockevents device, with a timer that is not queued to it - resulting in a
bogus next expiry value.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Michal Piotrowski <michal.k.k.piotrowski@gmail.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[VIDEO]: Fix section mismatch in cg3.c
Robert Reif [Wed, 28 Mar 2007 19:50:56 +0000 (12:50 -0700)]
[VIDEO]: Fix section mismatch in cg3.c

Fix section mismatch warning by moving data into __devinitdata section.
Add __devinit to initialization functions.

Signed-off-by: Robert Reif <reif@earthlink.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[DCCP] getsockopt: Fix DCCP_SOCKOPT_[SEND,RECV]_CSCOV
Arnaldo Carvalho de Melo [Wed, 28 Mar 2007 18:54:32 +0000 (11:54 -0700)]
[DCCP] getsockopt: Fix DCCP_SOCKOPT_[SEND,RECV]_CSCOV

We were only checking if there was enough space to put the int, but
left len as specified by the (malicious) user, sigh, fix it by setting
len to sizeof(val) and transfering just one int worth of data, the one
asked for.

Also check for negative len values.

Signed-off-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agosh: Trivial fix for hp6xx build.
Kristoffer Ericson [Wed, 28 Mar 2007 10:45:59 +0000 (19:45 +0900)]
sh: Trivial fix for hp6xx build.

The IRQ3 define was removed when asm-sh/irq.h was cleaned up,
this updates the hp6xx header to use the IRQ number directly.

Signed-off-by: Kristoffer Ericson <kristoffer_e1@hotmail.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
17 years agosh: Fixup __cmpxchg() compile breakage with gcc4.
Paul Mundt [Wed, 28 Mar 2007 08:26:19 +0000 (17:26 +0900)]
sh: Fixup __cmpxchg() compile breakage with gcc4.

As reported by Manuel:

When I build linux with GCC-4.x and enable
CONFIG_CC_OPTIMIZE_FOR_SIZE linking fails with this error:

  LD      .tmp_vmlinux1
  kernel/built-in.o: In function '__cmpxchg_called_with_bad_pointer'
  make[1]: *** [.tmp_vmlinux1] Error 1
  make: *** [_all] Error 2

This ended up being an inlining problem, fixed by explicitly
including linux/compiler.h and grabbing the definitions from there.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
17 years agosh: Kill bogus GCC4 symbol exports.
Paul Mundt [Wed, 28 Mar 2007 08:24:47 +0000 (17:24 +0900)]
sh: Kill bogus GCC4 symbol exports.

__sdivsi3_i4i, __udiv_qrnnd_16, and __udivsi3_i4i don't exist
outside of the ST compiler, so kill them off.

This causes compile failures with other GCC4 compilers.

Signed-off-by: Paul Mundt <lethal@linux-sh.org>
17 years agoMerge branch 'upstream-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/linvil...
Jeff Garzik [Wed, 28 Mar 2007 06:21:18 +0000 (02:21 -0400)]
Merge branch 'upstream-fixes' of git://git./linux/kernel/git/linville/wireless-2.6 into upstream-fixes

17 years agoSUN3/3X Lance trivial fix improved
Cyrill V. Gorcunov [Tue, 27 Mar 2007 05:47:26 +0000 (21:47 -0800)]
SUN3/3X Lance trivial fix improved

This patch adds checking for allocated DVMA memory and granted IRQ line.

Signed-off-by: Cyrill V. Gorcunov <gorcunov@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agomv643xx_eth: Fix use of uninitialized port_num field
Gabriel Paubert [Fri, 23 Mar 2007 19:03:52 +0000 (12:03 -0700)]
mv643xx_eth: Fix use of uninitialized port_num field

In this driver, the default ethernet address is first set by by calling
eth_port_uc_addr_get() which reads the relevant registers of the
corresponding port as initially set by firmware. However that function
used the port_num field accessed through the private area of net_dev
before it was set.

The result was that one board I have ended up with the unicast address
set to 00:00:00:00:00:00 (only port 1 is connected on this board). The
problem appeared after commit 84dd619e4dc3b0b1c40dafd98c90fd950bce7bc5.

This patch fixes the bug by setting mp->port_num prior to calling
eth_port_uc_get_addr().

Signed-off-by: Gabriel Paubert <paubert@iram.es>
Signed-off-by: Dale Farnsworth <dale@farnsworth.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoforcedeth: fix tx timeout
Ayaz Abdulla [Fri, 23 Mar 2007 10:50:02 +0000 (05:50 -0500)]
forcedeth: fix tx timeout

The tx timeout routine was waking the tx queue conditionally. However,
it must call it unconditionally since the dev_watchdog has halted the tx
queue before calling the timeout function.

Signed-Off-By: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoforcedeth: fix nic poll
Ayaz Abdulla [Fri, 23 Mar 2007 10:49:37 +0000 (05:49 -0500)]
forcedeth: fix nic poll

The nic poll routine was missing the call to the optimized irq routine.
This patch adds the missing call for the optimized path.

See http://bugzilla.kernel.org/show_bug.cgi?id=7950 for more information.

Signed-Off-By: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoqla3xxx: bugfix: Jumbo frame handling.
Ron Mercer [Mon, 26 Mar 2007 20:43:00 +0000 (13:43 -0700)]
qla3xxx: bugfix: Jumbo frame handling.

Fixed rx checksum bits. Turn on TCP processing for rx checksum.
Fixed max frame length register write.  It wasn't getting set
in multi-port system. Set rx buffer queue length properly
for jumbo frames.

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoqla3xxx: bugfix: Dropping interrupt under heavy network load.
Ron Mercer [Mon, 26 Mar 2007 20:42:59 +0000 (13:42 -0700)]
qla3xxx: bugfix: Dropping interrupt under heavy network load.

Update the rx queue pointer when exiting NAPI poll rather than
at the end of each iteration.  Remove unnecessary PCI flushes
that occurred after every write.  Now write all regs and
flush once.

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoqla3xxx: bugfix: Multi segment sends were getting whacked.
Ron Mercer [Mon, 26 Mar 2007 20:42:58 +0000 (13:42 -0700)]
qla3xxx: bugfix: Multi segment sends were getting whacked.

The proper header length was not being used.

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoqla3xxx: bugfix: Add tx control block memset.
Ron Mercer [Mon, 26 Mar 2007 20:42:57 +0000 (13:42 -0700)]
qla3xxx: bugfix: Add tx control block memset.

This was removed in a previous patch to increase performance, but
caused a transmit error for the 4032 chip.

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoatl1: remove unnecessary crc inversion
Jay Cliburn [Wed, 28 Mar 2007 00:43:49 +0000 (19:43 -0500)]
atl1: remove unnecessary crc inversion

The original vendor driver contained a private ether_crc_le() function
that produced an inverted crc.  When we changed to the kernel version of
ether_crc_le(), we neglected to undo the inversion.  Let's do it now.
Discovered by and patch proffered by Jose Alberto Reguero.

Signed-off-by: Jose Alberto Reguero <jareguero@telefonica.net>
Signed-off-by: Jay Cliburn <jacliburn@bellsouth.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agomyri10ge: correctly detect when TSO should be used
Brice Goglin [Tue, 27 Mar 2007 19:54:53 +0000 (21:54 +0200)]
myri10ge: correctly detect when TSO should be used

Correctly detect when TSO should be used on transmit by looking at the
skb->gso_size rather than seeing if the frame was larger than our MTU.
The old method causes problems when a host with a large (jumbo) MTU is
sending to a host with a small (standard) MTU.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoata: NCQ is broken on Maxtor 6L250S0
Paul Rolland [Tue, 27 Mar 2007 05:43:44 +0000 (21:43 -0800)]
ata: NCQ is broken on Maxtor 6L250S0

With this applied, my machine has stopped all those painful messages.
dmesg now says :

root@riri:/Kernels# dmesg | grep LBA
ata1.00: 490234752 sectors, multi 0: LBA48 NCQ (not used)
ata2.00: 640 sectors, multi 1: LBA
ata3.00: 490234752 sectors, multi 0: LBA48 NCQ (not used)

Signed-off-by: Paul Rolland <rol@as2917.net>
Acked-by: Alan Cox <alan@redhat.com>
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Tejun Heo <htejun@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agopata_pdc202xx_old: LBA48 bug
Alan Cox [Fri, 23 Mar 2007 18:57:23 +0000 (18:57 +0000)]
pata_pdc202xx_old: LBA48 bug

In LBA48 mode we have to help the controller to get anything to work. The
chip provides a register giving word counts meant for ATAPI use which we
can use. However we need to load the count in words not bytes..

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agolibata: IDENTIFY backwards for drive side cable detection
Tejun Heo [Thu, 22 Mar 2007 13:24:19 +0000 (22:24 +0900)]
libata: IDENTIFY backwards for drive side cable detection

For drive side cable detection to work correctly, drives need to be
identified backwards such that the slave device releases PDIAG- before
the mater drive tries to detect cable type.  ata_bus_probe() was fixed
by commit f31f0cc2f0b7527072d94d02da332d9bb8d7d94c but the new EH path
wasn't fixed.  This patch makes new EH path do IDENTIFY backwards.

ata_dev_configure() for new devices are still performed master first.
This is to keep the detection messages in forward order.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoahci.c: walkaround for SB600 SATA internal error issue
Conke Hu [Tue, 27 Mar 2007 10:33:05 +0000 (18:33 +0800)]
ahci.c: walkaround for SB600 SATA internal error issue

   There is a HW issue in ATI SB600 SATA that PxSERR.E should not be
set on some conditions, for example, when there is no media in SATA
CD/DVD drive or media is not ready, AHCI controller fails to execute
ATAPI commands and reports PORT_IRQ_TF_ERR, but ATI SB600 SATA
controller sets PxSERR.E at the
same time, which is not necessary.
    This patch is just to ignore the INTERNAL ERROR in such case.
Without this patch, ahci error handler will report many errors as
below:
    ----------- cut from dmesg -----------
ata9: soft resetting port
ata9: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata9.00: configured for UDMA/33
ata9: EH complete
ata9.00: exception Emask 0x40 SAct 0x0 SErr 0x800 action 0x2
ata9.00: (irq_stat 0x40000001)
ata9.00: cmd a0/00:00:00:00:20/00:00:00:00:00/a0 tag 0 cdb 0x0 data 0
        res 51/24:03:00:00:20/00:00:00:00:00/a0 Emask 0x40 (internal error)
ata9: soft resetting port
ata9: SATA link up 1.5 Gbps (SStatus 113 SControl 300)
ata9.00: configured for UDMA/33
ata9: EH complete
ata9.00: exception Emask 0x40 SAct 0x0 SErr 0x800 action 0x2
ata9.00: (irq_stat 0x40000001)
ata9.00: cmd a0/01:00:00:00:00/00:00:00:00:00/a0 tag 0 cdb 0x43 data 12 in
        res 51/24:03:00:00:00/00:00:00:00:00/a0 Emask 0x40 (internal error)
    -------- end cut ---------

Signed-off-by: Conke Hu <conke.hu@amd.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years ago[libata] Disable ACPI by default; fix namespace problems
Jeff Garzik [Wed, 28 Mar 2007 05:57:37 +0000 (01:57 -0400)]
[libata] Disable ACPI by default; fix namespace problems

Not yet ready to turn on ATA ACPI by default, for either PATA or SATA.

Also, rename the global-scope module parameter variable 'noacpi' to
something more libata-specific, reducing the potential for namespace
collision.

Signed-off-by: Jeff Garzik <jeff@garzik.org>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 28 Mar 2007 00:56:42 +0000 (17:56 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [IPV6]: Set IF_READY if the device is up and has carrier
  [NET_SCHED]: sch_htb/sch_hfsc: fix oops in qlen_notify
  [NET]: Remove dead net/sched/Makefile entry for sch_hpfq.o.

17 years ago[PATCH] aio: remove bare user-triggerable error printk
Zach Brown [Tue, 27 Mar 2007 22:44:01 +0000 (15:44 -0700)]
[PATCH] aio: remove bare user-triggerable error printk

The user can generate console output if they cause do_mmap() to fail
during sys_io_setup().  This was seen in a regression test that does
exactly that by spinning calling mmap() until it gets -ENOMEM before
calling io_setup().

We don't need this printk at all, just remove it.

Signed-off-by: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[IPV6]: Set IF_READY if the device is up and has carrier
Herbert Xu [Tue, 27 Mar 2007 21:31:52 +0000 (14:31 -0700)]
[IPV6]: Set IF_READY if the device is up and has carrier

We still need to set the IF_READY flag in ipv6_add_dev for the case
where all addresses (including the link-local) are deleted and then
recreated.  In that case the IPv6 device too will be destroyed and
then recreated.

In order to prevent the original problem, we simply ensure that
the device is up before setting IF_READY.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[NET_SCHED]: sch_htb/sch_hfsc: fix oops in qlen_notify
Patrick McHardy [Tue, 27 Mar 2007 21:04:24 +0000 (14:04 -0700)]
[NET_SCHED]: sch_htb/sch_hfsc: fix oops in qlen_notify

During both HTB and HFSC class deletion the class is removed from the
class hash before calling qdisc_tree_decrease_qlen. This makes the
->get operation in qdisc_tree_decrease_qlen fail, so it passes a NULL
pointer to ->qlen_notify, causing an oops.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[PATCH] WE-22 : prevent information leak on 64 bit
Jean Tourrilhes [Fri, 23 Mar 2007 00:31:16 +0000 (00:31 +0000)]
[PATCH] WE-22 : prevent information leak on 64 bit

  Johannes Berg discovered that kernel space was leaking to
userspace on 64 bit platform. He made a first patch to fix that. This
is an improved version of his patch.

Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] wext: Add missing ioctls to 64<->32 conversion
Jean Tourrilhes [Fri, 23 Mar 2007 00:26:49 +0000 (00:26 +0000)]
[PATCH] wext: Add missing ioctls to 64<->32 conversion

  Johannes Berg and Michael Buesch noticed that the WPA ioctls
were missing from the 64<->32 bit conversion. This means that when
using a 32 bits userspace on a 64 bit kernel, those ioctls fail.

Signed-off-by: Jean Tourrilhes <jt@hpl.hp.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] bcm43xx: Fix machine check on PPC for version 1 PHY
David Woodhouse [Sun, 25 Mar 2007 13:45:54 +0000 (08:45 -0500)]
[PATCH] bcm43xx: Fix machine check on PPC for version 1 PHY

Recent changes in the specs that were introduced in commit
740ac4fb08866d702be90f167665d03759bd27d0 were incorrect and resulted in machine check
errors on the PPC architecture for G PHY's with a revision number equal to 1. The
two offending changes are reverted.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] bcm43xx: fix radio_set_tx_iq
Stefano Brivio [Fri, 23 Mar 2007 19:21:39 +0000 (20:21 +0100)]
[PATCH] bcm43xx: fix radio_set_tx_iq

Fix a duplicated leftshift in bcm43xx_radio_set_tx_iq. data_high values are
already leftshifted. Thanks to Michael Buesch for spotting this.

Signed-off-by: Stefano Brivio <stefano.brivio@polimi.it>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years ago[PATCH] bcm43xx: Fix code for confusion between PHY revision and PHY version
Larry Finger [Sat, 17 Mar 2007 16:28:21 +0000 (11:28 -0500)]
[PATCH] bcm43xx: Fix code for confusion between PHY revision and PHY version

There are several places where the PHY version and revision were interchanged.
These are changed in the specifications on 2/13/07 and now use "analog" instead
instead of "version" to help reduce confusion.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Tue, 27 Mar 2007 17:06:30 +0000 (10:06 -0700)]
Merge /linux/kernel/git/jejb/scsi-rc-fixes-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6:
  [SCSI] gdth: fix oops in gdth_copy_cmd()
  [SCSI] mptsas: Fix oops for insmod during kexec
  [SCSI] lpfc: avoid double-free during PCI error failure

17 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb
Linus Torvalds [Tue, 27 Mar 2007 16:23:59 +0000 (09:23 -0700)]
Merge git://git./linux/kernel/git/mchehab/v4l-dvb

* git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb:
  V4L/DVB (5472): Isl6421: don't reference freed memory
  V4L/DVB (5441): Saa7146: Fix allocation of clipping memory
  V4L/DVB (5421): Fix suspend/resume in msp3400 and tuner
  V4L/DVB (5415): Msp_attach must return 0 if no msp3400 was found.
  V4L/DVB (5408): Fix SECAM handling on saa7115
  V4L/DVB (5400): Core: fix several locking related problems
  V4L/DVB (5390): Radio: Fix error in Kbuild file
  V4L/DVB (5332): Ir_rc5_timer_end decoder lockup fix

17 years agoMerge git://kvm.qumranet.com/home/avi/kvm
Linus Torvalds [Tue, 27 Mar 2007 16:22:05 +0000 (09:22 -0700)]
Merge git://kvm.qumranet.com/home/avi/kvm

* git://kvm.qumranet.com/home/avi/kvm:
  KVM: always reload segment selectors
  KVM: Prevent system selectors leaking into guest on real->protected mode transition on vmx

17 years ago[PATCH] Revert "swsusp: disable nonboot CPUs before entering platform suspend"
Rafael J. Wysocki [Tue, 27 Mar 2007 10:09:13 +0000 (12:09 +0200)]
[PATCH] Revert "swsusp: disable nonboot CPUs before entering platform suspend"

This reverts commit 94985134b7b46848267ed6b734320db01c974e72 and
insteads removes the WARN_ON() that caused that commit in the first
place.

The problem is that we call disable_nonboot_cpus() in swsusp before
powering down the system in order to avoid triggering the WARN_ON()
in arch/x86_64/kernel/acpi/sleep.c:init_low_mapping() and this doesn't
work well on Thomas' system.

So instead, remove the WARN_ON() in arch/x86_64/kernel/acpi/sleep.c:
init_low_mapping(), which triggers every time during the suspend to disk
in the platform mode, as the potential problem it is related to doesn't
seem to occur in practice.

[ I think we might want to disallow the case of multiple users of that
  mm, or something.  Normally, playing with the current process page
  tables on the current CPU should be fine as long as we don't have
  other threads using those tables at the same time..

  Anyway, not pretty, but better than the warning or the lockup - Linus ]

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] i386: Fix bogus return value in hpet_next_event()
Thomas Gleixner [Tue, 27 Mar 2007 07:08:26 +0000 (09:08 +0200)]
[PATCH] i386: Fix bogus return value in hpet_next_event()

The clockevents / tick management code expects an error value, when the
event is already expired. hpet_next_event() returns 1 in that case.

Fix it to return the proper -ETIME error code.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.dk/data/git/linux-2.6-block
Linus Torvalds [Tue, 27 Mar 2007 16:05:49 +0000 (09:05 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/data/git/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/data/git/linux-2.6-block:
  Export __splice_from_pipe()
  2/2 splice: dont readpage
  1/2 splice: dont steal
  make elv_register() output atomic
  block: blk_max_pfn is somtimes wrong

17 years ago[PATCH] Fix kernel build with EMBEDDED & PROC_FS & !PROC_SYSCTL
Mika Kukkonen [Tue, 27 Mar 2007 05:32:33 +0000 (21:32 -0800)]
[PATCH] Fix kernel build with EMBEDDED & PROC_FS & !PROC_SYSCTL

Without attached patch against current -git I get following with
!PROC_SYSCTL (with EMBEDDED and PROC_FS set):

    CC      init/version.o
    LD      init/built-in.o
    LD      vmlinux
  fs/built-in.o: In function `do_proc_sys_lookup':
  proc_sysctl.c:(.text+0x26583): undefined reference to `sysctl_head_next'
  fs/built-in.o: In function `proc_sys_revalidate':
  proc_sysctl.c:(.text+0x265bb): undefined reference to `sysctl_head_finish'
  fs/built-in.o: In function `proc_sys_readdir':
  proc_sysctl.c:(.text+0x26720): undefined reference to `sysctl_head_next'
  proc_sysctl.c:(.text+0x267d8): undefined reference to `sysctl_head_finish'
  proc_sysctl.c:(.text+0x268e7): undefined reference to `sysctl_head_next'
  proc_sysctl.c:(.text+0x26910): undefined reference to `sysctl_head_finish'
  fs/built-in.o: In function `proc_sys_write':
  proc_sysctl.c:(.text+0x2695d): undefined reference to `sysctl_perm'
  proc_sysctl.c:(.text+0x2699c): undefined reference to `sysctl_head_finish'
  fs/built-in.o: In function `proc_sys_read':
  proc_sysctl.c:(.text+0x269e9): undefined reference to `sysctl_perm'
  proc_sysctl.c:(.text+0x26a25): undefined reference to `sysctl_head_finish'
  fs/built-in.o: In function `proc_sys_permission':
  proc_sysctl.c:(.text+0x26ad1): undefined reference to `sysctl_perm'
  proc_sysctl.c:(.text+0x26adb): undefined reference to `sysctl_head_finish'
  fs/built-in.o: In function `proc_sys_lookup':
  proc_sysctl.c:(.text+0x26b39): undefined reference to `sysctl_head_finish'
  make: *** [vmlinux] Virhe 1

All those functions are in fs/proc/proc_sysctl.c, which has no CONFIG_
#define's in it, so the patch makes the compilation of that file to depend
on CONFIG_PROC_SYSCTL (the simplest choice).

Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] I2O: remove Markus from MAINTAINERS
Markus Lidel [Tue, 27 Mar 2007 05:32:32 +0000 (21:32 -0800)]
[PATCH] I2O: remove Markus from MAINTAINERS

Because i don't have much time lately and my responses are pretty slow
it's probably best to remove me from MAINTAINERS to give someone else
the chance to jump in.

Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ipcns: fix !CONFIG_IPC_NS behavior
Serge E. Hallyn [Tue, 27 Mar 2007 05:32:31 +0000 (21:32 -0800)]
[PATCH] ipcns: fix !CONFIG_IPC_NS behavior

When CONFIG_IPC_NS=n, clone(CLONE_NEWIPC) claims success, but did not actually
clone a new IPC namespace.

Fix this to return -EINVAL so the caller knows his request was denied.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] i2o: block IO errors on i2o disk
Vasily Averin [Tue, 27 Mar 2007 05:32:30 +0000 (21:32 -0800)]
[PATCH] i2o: block IO errors on i2o disk

I2O subsystem has been broken in mainstream several months ago (after
2.6.18).  Commit 4aff5e2333c9a1609662f2091f55c3f6fffdad36 from Jens
Axboe split struct request ->flags into two parts: cmd_type and
cmd_flags.

In i2o layer this patch has replaced flag REQ_SPECIAL by the according
cmd_type.  However i2o has used REQ_SPECIAL not as command type but as
driver-specific flag for the debug purposes.  As result all i2o requests
have type "special" now, are not processed to the hardware and fail with
I/O error:

   i2o/hda:<3>Buffer I/O error on device i2o/hda, logical block 0
  Buffer I/O error on device i2o/hda, logical block 0
  Buffer I/O error on device i2o/hda, logical block 0
   unable to read partition table
  block-osm: device added (TID: 207): i2o/hda

The following patch removes the extra debug checks without any drawbacks and
restores the normal driver's work.

Signed-off-by: Vasily Averin <vvs@sw.ru>
Acked-by: Markus Lidel <Markus.Lidel@shadowconnect.com>
Cc: Jens Axboe <jens.axboe@oracle.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Fix struct device member name in PCMCIA au1000_generic
Yoichi Yuasa [Tue, 27 Mar 2007 05:32:28 +0000 (21:32 -0800)]
[PATCH] Fix struct device member name in PCMCIA au1000_generic

  drivers/pcmcia/au1000_generic.c: In function 'au1x00_pcmcia_socket_probe':
  drivers/pcmcia/au1000_generic.c:375: error: 'struct device' has no member named 'dev'

Signed-off-by: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Jaroslav Kysela <perex@suse.cz>
Cc: Takashi Iwai <tiwai@suse.de>
Cc: Greg KH <greg@kroah.com>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] drivers/eisa/pci_eisa.c:pci_eisa_init() should be init
Adrian Bunk [Tue, 27 Mar 2007 05:32:27 +0000 (21:32 -0800)]
[PATCH] drivers/eisa/pci_eisa.c:pci_eisa_init() should be init

  WARNING: drivers/built-in.o - Section mismatch: reference to .init.text:eisa_root_register from .text between 'pci_eisa_init' (at offset 0xabf670) and 'virtual_eisa_release'

AFAIK a PCI to EISA bridge isn't anything hotpluggable, so
pci_eisa_init() can become __init.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ntp: avoid time_offset overflows
john stultz [Tue, 27 Mar 2007 05:32:26 +0000 (21:32 -0800)]
[PATCH] ntp: avoid time_offset overflows

I've been seeing some odd NTP behavior recently on a few boxes and
finally narrowed it down to time_offset overflowing when converted to
SHIFT_UPDATE units (which was a side effect from my HZfreeNTP patch).

This patch converts time_offset from a long to a s64 which resolves the
issue.

[tglx@linutronix.de: signedness fixes]
Signed-off-by: John Stultz <johnstul@us.ibm.com>
Cc: Roman Zippel <zippel@linux-m68k.org>
Cc: john stultz <johnstul@us.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: use correct register file size everywhere
Jeff Dike [Tue, 27 Mar 2007 05:32:25 +0000 (21:32 -0800)]
[PATCH] uml: use correct register file size everywhere

This patch uses MAX_REG_NR consistently to refer to the register file size.
 FRAME_SIZE isn't sufficient because on x86_64, it is smaller than the
ptrace register file size.  MAX_REG_NR was introduced as a consistent way
to get the number of registers, but wasn't used everywhere it should be.

When this causes a problem, it makes PTRACE_SETREGS fail on x86_64 because
of a corrupted segment register value in the known-good register file.  The
patch also adds a register dump at that point in case there are any future
problems here.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: fix static linking
Jeff Dike [Tue, 27 Mar 2007 05:32:24 +0000 (21:32 -0800)]
[PATCH] uml: fix static linking

During a static link, ld has started putting a .note section in the
.uml.setup.init section.  This has the result that the UML setups begin
with 32 bytes of garbage and UML crashes immediately on boot.

This patch creates a specific .note section for ld to drop this stuff
into.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] drivers/spi/: fix section mismatches
Adrian Bunk [Tue, 27 Mar 2007 05:32:23 +0000 (21:32 -0800)]
[PATCH] drivers/spi/: fix section mismatches

  WARNING: drivers/built-in.o - Section mismatch: reference to .init.text:spi_register_master from .text between 'spi_bitbang_start' (at offset 0x84e11a) and 'bitbang_work'
  WARNING: drivers/built-in.o - Section mismatch: reference to .init.text:spi_alloc_master from .text between 'butterfly_attach' (at offset 0x84e681) and 'at25_remove'
  WARNING: drivers/built-in.o - Section mismatch: reference to .init.text:spi_new_device from .text between 'butterfly_attach' (at offset 0x84e7e4) and 'at25_remove'

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] utsns: fix !CONFIG_UTS_NS behavior
Serge E. Hallyn [Tue, 27 Mar 2007 05:32:22 +0000 (21:32 -0800)]
[PATCH] utsns: fix !CONFIG_UTS_NS behavior

When CONFIG_UTS_NS=n, clone(CLONE_NEWUTS) quietly refuses.  So correctly does
not unshare a new uts namespace, but also does not return -EINVAL.

Fix this to return -EINVAL so the caller knows his request was denied.

Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] tty: minor merge correction
Alan Cox [Tue, 27 Mar 2007 05:32:21 +0000 (21:32 -0800)]
[PATCH] tty: minor merge correction

Its now used.. because we added the new definitions so enabled all the
goodies on i386

Signed-off-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] uml: fix epoll
Jeff Dike [Tue, 27 Mar 2007 05:32:20 +0000 (21:32 -0800)]
[PATCH] uml: fix epoll

UML/x86_64 needs the same packing of struct epoll_event as x86_64.

Signed-off-by: Jeff Dike <jdike@linux.intel.com>
Cc: Davide Libenzi <davidel@xmailserver.org>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] ISDN: fix BChannel_bh() call from isar_bh()
Jiri Kosina [Tue, 27 Mar 2007 05:32:19 +0000 (21:32 -0800)]
[PATCH] ISDN: fix BChannel_bh() call from isar_bh()

isar_bh() bh handler calls another (compatible) bh handler - BChannel_bh()
- but passes struct BCState* instead of struct work_struct*, which seems
wrong.

Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] correct slow acpi_pm rating
john stultz [Tue, 27 Mar 2007 05:32:19 +0000 (21:32 -0800)]
[PATCH] correct slow acpi_pm rating

On Bob's machine clocksource is selecting PIT over the ACPI PM timer,
because he has the PIIX4 bug.  That bug drops the ACPI PM timers rating
to the same as the PIT, so that's why you're getting the PIT.

Realistically, the PIT is much slower then even the triple read ACPI PM,
so the de-ranking code is probably dropping it too far.

So don't drop ACPI PM quite so low if we see the PIIX4 bug.

Signed-off-by: John Stultz <johnstul@us.ibm.com>
Cc: Bob Tracy <rct@gherkin.frus.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] generic_serial: fix decoding of baud rate
Daniel Drake [Tue, 27 Mar 2007 05:32:15 +0000 (21:32 -0800)]
[PATCH] generic_serial: fix decoding of baud rate

Commit d720bc4b8fc5d6d179ef094908d4fbb5e436ffad partially removed a
private implementation of baud speed decoding.  However it doesn't seem
to be complete: after the speed is decoded, it is still being used as an
index to a local speed table (array overrun, no doubt).

This was found by Graham Murray who noticed it caused a 2.6.19 regression
with the SX driver: https://bugs.gentoo.org/170554

Signed-off-by: Daniel Drake <dsd@gentoo.org>
Acked-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: convert compile time warnings into runtime warnings
NeilBrown [Tue, 27 Mar 2007 05:32:14 +0000 (21:32 -0800)]
[PATCH] md: convert compile time warnings into runtime warnings

...  still not sure why we need this ....

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: clear the congested_fn when stopping a raid5
NeilBrown [Tue, 27 Mar 2007 05:32:14 +0000 (21:32 -0800)]
[PATCH] md: clear the congested_fn when stopping a raid5

If this mddev and queue got reused for another array that doesn't register a
congested_fn, this function would get called incorretly.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] md: allow raid4 arrays to be reshaped
NeilBrown [Tue, 27 Mar 2007 05:32:13 +0000 (21:32 -0800)]
[PATCH] md: allow raid4 arrays to be reshaped

All that is missing the the function pointers in raid4_pers.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: nfsd4: remove superfluous cancel_delayed_work() call
J. Bruce Fields [Tue, 27 Mar 2007 05:32:10 +0000 (21:32 -0800)]
[PATCH] knfsd: nfsd4: remove superfluous cancel_delayed_work() call

This cancel_delayed_work call is called from a function that is only called
from a piece of code that immediate follows a cancel and destruction of the
workqueue, so it's clearly a mistake.

Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: nfsd4: demote "clientid in use" printk to a dprintk
Bruce Fields [Tue, 27 Mar 2007 05:32:09 +0000 (21:32 -0800)]
[PATCH] knfsd: nfsd4: demote "clientid in use" printk to a dprintk

The reused clientid here is a more of a problem for the client than the
server, and the client can report the problem itself if it's serious.

Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: nfsd4: fix inheritance flags on v4 ace derived from posix default ace
Bruce Fields [Tue, 27 Mar 2007 05:32:09 +0000 (21:32 -0800)]
[PATCH] knfsd: nfsd4: fix inheritance flags on v4 ace derived from posix default ace

A regression introduced in the last set of acl patches removed the
INHERIT_ONLY flag from aces derived from the posix acl.  Fix.

Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] knfsd: allow nfsd READDIR to return 64bit cookies
NeilBrown [Tue, 27 Mar 2007 05:32:08 +0000 (21:32 -0800)]
[PATCH] knfsd: allow nfsd READDIR to return 64bit cookies

->readdir passes lofft_t offsets (used as nfs cookies) to
nfs3svc_encode_entry{,_plus}, but when they pass it on to encode_entry it
becomes an 'off_t', which isn't good.

So filesystems that returned 64bit offsets would lose.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[PATCH] Fix maxcpus=1 trigerring BUG() in cpufreq
Venki Pallipadi [Mon, 26 Mar 2007 19:03:19 +0000 (12:03 -0700)]
[PATCH] Fix maxcpus=1 trigerring BUG() in cpufreq

Ingo reported it on lkml in the thread
  "2.6.21-rc5: maxcpus=1 crash in cpufreq: kernel BUG at drivers/cpufreq/cpufreq.c:82!"

This check added to remove_dev  is symmetric to one in add_dev and handles
callbacks for offline cpus cleanly.

Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoKVM: always reload segment selectors
Ingo Molnar [Sun, 11 Mar 2007 12:52:33 +0000 (13:52 +0100)]
KVM: always reload segment selectors

failed VM entry on VMX might still change %fs or %gs, thus make sure
that KVM always reloads the segment selectors. This is crutial on both
x86 and x86_64: x86 has __KERNEL_PDA in %fs on which things like
'current' depends and x86_64 has 0 there and needs MSR_GS_BASE to work.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
17 years agoKVM: Prevent system selectors leaking into guest on real->protected mode transition...
Avi Kivity [Mon, 19 Mar 2007 11:18:10 +0000 (13:18 +0200)]
KVM: Prevent system selectors leaking into guest on real->protected mode transition on vmx

Intel virtualization extensions do not support virtualizing real mode.  So
kvm uses virtualized vm86 mode to run real mode code.  Unfortunately, this
virtualized vm86 mode does not support the so called "big real" mode, where
the segment selector and base do not agree with each other according to the
real mode rules (base == selector << 4).

To work around this, kvm checks whether a selector/base pair violates the
virtualized vm86 rules, and if so, forces it into conformance.  On a
transition back to protected mode, if we see that the guest did not touch
a forced segment, we restore it back to the original protected mode value.

This pile of hacks breaks down if the gdt has changed in real mode, as it
can cause a segment selector to point to a system descriptor instead of a
normal data segment.  In fact, this happens with the Windows bootloader
and the qemu acpi bios, where a protected mode memcpy routine issues an
innocent 'pop %es' and traps on an attempt to load a system descriptor.

"Fix" by checking if the to-be-restored selector points at a system segment,
and if so, coercing it into a normal data segment.  The long term solution,
of course, is to abandon vm86 mode and use emulation for big real mode.

Signed-off-by: Avi Kivity <avi@qumranet.com>
17 years agoV4L/DVB (5472): Isl6421: don't reference freed memory
Thomas Viehweger [Thu, 22 Mar 2007 14:20:32 +0000 (11:20 -0300)]
V4L/DVB (5472): Isl6421: don't reference freed memory

After freeing a block there should be no reference to this block.

Signed-off-by: Thomas Viehweger <Thomas.Viehweger@marconi.com>
Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5441): Saa7146: Fix allocation of clipping memory
Oliver Endriss [Wed, 14 Mar 2007 02:44:57 +0000 (23:44 -0300)]
V4L/DVB (5441): Saa7146: Fix allocation of clipping memory

Olaf Hering pointed out that SAA7146_CLIPPING_MEM would become
very large for PAGE_SIZE > 4K.
In fact, the number of clipping windows is limited to 16,
and calculate_clipping_registers_rect() does not use more
than 256 bytes. SAA7146_CLIPPING_MEM adjusted accordingly.

Thanks-to: Olaf Hering <olaf@aepfle.de>
Acked-by: Michael Hunold <hunold@linuxtv.org>
Signed-off-by: Oliver Endriss <o.endriss@gmx.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5421): Fix suspend/resume in msp3400 and tuner
Jean Delvare [Mon, 12 Mar 2007 22:20:15 +0000 (19:20 -0300)]
V4L/DVB (5421): Fix suspend/resume in msp3400 and tuner

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5415): Msp_attach must return 0 if no msp3400 was found.
Hans Verkuil [Sat, 10 Mar 2007 23:21:07 +0000 (20:21 -0300)]
V4L/DVB (5415): Msp_attach must return 0 if no msp3400 was found.

Returning -1 causes the probe to stop, but it should just continue
instead.

Signed-off-by: Hans Verkuil <hverkuil@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5408): Fix SECAM handling on saa7115
Mauro Carvalho Chehab [Tue, 13 Mar 2007 16:41:49 +0000 (13:41 -0300)]
V4L/DVB (5408): Fix SECAM handling on saa7115

Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5400): Core: fix several locking related problems
Simon Arlott [Sat, 10 Mar 2007 09:21:25 +0000 (06:21 -0300)]
V4L/DVB (5400): Core: fix several locking related problems

Fix several instances of dvb-core functions using mutex_lock_interruptible
and returning -ERESTARTSYS where the calling function will either never
retry or never check the return value.
These cause a race condition with dvb_dmxdev_filter_free and
dvb_dvr_release, both of which are filesystem release functions whose
return value is ignored and will never be retried.  When this happens it
becomes impossible to open dvr0 again (-EBUSY) since it has not been
released properly.

Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-By: Johannes Stezenbach <js@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5390): Radio: Fix error in Kbuild file
Trent Piepho [Wed, 7 Mar 2007 21:19:46 +0000 (18:19 -0300)]
V4L/DVB (5390): Radio: Fix error in Kbuild file

All the radio drivers need video_dev, but they were depending on
VIDEO_DEV!=n.  That meant that one could try to compile the driver into
the kernel when VIDEO_DEV=m, which will not work.  If video_dev is a
module, then the radio drivers must be modules too.

Signed-off-by: Trent Piepho <xyzzy@speakeasy.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years agoV4L/DVB (5332): Ir_rc5_timer_end decoder lockup fix
Vincent Penne [Sun, 25 Mar 2007 14:58:23 +0000 (11:58 -0300)]
V4L/DVB (5332): Ir_rc5_timer_end decoder lockup fix

ir_rc5_timer_end decoder lockup fix

Signed-off-by: Vincent Penne <ziggy@sashipa.com>
Signed-off-by: Hermann Pitton <hermann-pitton@arcor.de>
Signed-off-by: Hartmut Hackmann <hartmut.hackmann@t-online.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>
17 years ago[SPARC]: sparc64 gcc-4.2.0 20070317 -Werror failure
Mikael Pettersson [Tue, 27 Mar 2007 08:13:55 +0000 (01:13 -0700)]
[SPARC]: sparc64 gcc-4.2.0 20070317 -Werror failure

Compiling 2.6.21-rc5 with gcc-4.2.0 20070317 (prerelease)
for sparc64 fails as follows:

  gcc -Wp,-MD,arch/sparc64/kernel/.time.o.d  -nostdinc -isystem /home/mikpe/pkgs/linux-sparc64/gcc-4.2.0/lib/gcc/sparc64-unknown-linux-gnu/4.2.0/include -D__KERNEL__ -Iinclude  -include include/linux/autoconf.h -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing -fno-common -Os -m64 -pipe -mno-fpu -mcpu=ultrasparc -mcmodel=medlow -ffixed-g4 -ffixed-g5 -fcall-used-g7 -Wno-sign-compare -Wa,--undeclared-regs -fomit-frame-pointer  -fno-stack-protector -Wdeclaration-after-statement -Wno-pointer-sign -Werror   -D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_STR(time)"  -D"KBUILD_MODNAME=KBUILD_STR(time)" -c -o arch/sparc64/kernel/time.o arch/sparc64/kernel/time.c
cc1: warnings being treated as errors
arch/sparc64/kernel/time.c: In function 'kick_start_clock':
arch/sparc64/kernel/time.c:559: warning: overflow in implicit constant conversion
make[1]: *** [arch/sparc64/kernel/time.o] Error 1
make: *** [arch/sparc64/kernel] Error 2

gcc gets unhappy when the MSTK_SET macro's u8 __val variable
is updated with &= ~0xff (MSTK_YEAR_MASK). Making the constant
unsigned fixes the problem.

[ I fixed up the sparc32 side as well -DaveM ]

Signed-off-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoExport __splice_from_pipe()
Mark Fasheh [Wed, 21 Mar 2007 12:11:02 +0000 (13:11 +0100)]
Export __splice_from_pipe()

Ocfs2 wants to implement it's own splice write actor so that it can better
manage cluster / page locks. This lets us re-use the rest of splice write
while only providing our own code where it's actually important.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years ago2/2 splice: dont readpage
Nick Piggin [Tue, 27 Mar 2007 06:55:39 +0000 (08:55 +0200)]
2/2 splice: dont readpage

Splice does not need to readpage to bring the page uptodate before writing
to it, because prepare_write will take care of that for us.

Splice is also wrong to SetPageUptodate before the page is actually uptodate.
This results in the old uninitialised memory leak. This gets fixed as a
matter of course when removing the readpage logic.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years ago1/2 splice: dont steal
Nick Piggin [Tue, 27 Mar 2007 06:55:08 +0000 (08:55 +0200)]
1/2 splice: dont steal

Stealing pages with splice is problematic because we cannot just insert
an uptodate page into the pagecache and hope the filesystem can take care
of it later.

We also cannot just ClearPageUptodate, then hope prepare_write does not
write anything into the page, because I don't think prepare_write gives
that guarantee.

Remove support for SPLICE_F_MOVE for now. If we really want to bring it
back, we might be able to do so with a the new filesystem buffered write
aops APIs I'm working on. If we really don't want to bring it back, then
we should decide that sooner rather than later, and remove the flag and
all the stealing infrastructure before anybody starts using it.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agomake elv_register() output atomic
Thibaut VARENE [Thu, 15 Mar 2007 11:59:19 +0000 (12:59 +0100)]
make elv_register() output atomic

Booting 2.6.21-rc3-g45592145 I noticed the following on one of my
machines in the bootlog:

io scheduler noop registered<6>Time: jiffies clocksource has been installed.

io scheduler deadline registered (default)

Looking at block/elevator.c, it appears that elv_register() uses two
consecutive printks in a non-atomic way, leading to the above glitch. The
attached trivial patch fixes this issue, by using a single printk.

Signed-off-by: Thibaut VARENE <varenet@parisc-linux.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years agoblock: blk_max_pfn is somtimes wrong
Vasily Tarasov [Tue, 27 Mar 2007 06:52:47 +0000 (08:52 +0200)]
block: blk_max_pfn is somtimes wrong

There is a small problem in handling page bounce.

At the moment blk_max_pfn equals max_pfn, which is in fact not maximum
possible _number_ of a page frame, but the _amount_ of page frames.  For
example for the 32bit x86 node with 4Gb RAM, max_pfn = 0x100000, but not
0xFFFF.

request_queue structure has a member q->bounce_pfn and queue needs bounce
pages for the pages _above_ this limit.  This routine is handled by
blk_queue_bounce(), where the following check is produced:

if (q->bounce_pfn >= blk_max_pfn)
return;

Assume, that a driver has set q->bounce_pfn to 0xFFFF, but blk_max_pfn
equals 0x10000.  In such situation the check above fails and for each bio
we always fall down for iterating over pages tied to the bio.

I want to notice, that for quite a big range of device drivers (ide, md,
...) such problem doesn't happen because they use BLK_BOUNCE_ANY for
bounce_pfn.  BLK_BOUNCE_ANY is defined as blk_max_pfn << PAGE_SHIFT, and
then the check above doesn't fail.  But for other drivers, which obtain
reuired value from drivers, it fails.  For example sata_nv uses
ATA_DMA_MASK or dev->dma_mask.

I propose to use (max_pfn - 1) for blk_max_pfn.  And the same for
blk_max_low_pfn.  The patch also cleanses some checks related with
bounce_pfn.

Signed-off-by: Vasily Tarasov <vtaras@openvz.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
17 years ago[VIDEO] ffb: Fix two DAC handling bugs.
David S. Miller [Tue, 27 Mar 2007 06:18:09 +0000 (23:18 -0700)]
[VIDEO] ffb: Fix two DAC handling bugs.

The determination of whether the DAC has inverted cursor logic is
broken, import the version checks the X.org driver uses to fix this.

Next, when we change the timing generator, borrow code from X.org that
does 10 NOP reads of the timing generator register afterwards to make
sure the video-enable transition occurs cleanly.

Finally, use macros for the DAC registers and fields in order to
provide documentation for the next person who reads this code.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[SPARC32]: Fix SMP build regression
Robert Reif [Tue, 27 Mar 2007 02:10:43 +0000 (19:10 -0700)]
[SPARC32]: Fix SMP build regression

commit b19cbe2a1695c09c74f83646c4b82b51123b3690 [BRIDGE]: Fix fdb RCU
race

breaks sparc SMP build because atomic_add_unless is not exported.

This patch exports atomic_add_unless and atomic_cmpxchg.

Signed-off-by: Robert Reif <reif@earthlink.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[DRM]: Delete sparc64 FFB driver code that never gets built.
David S. Miller [Mon, 26 Mar 2007 23:54:39 +0000 (16:54 -0700)]
[DRM]: Delete sparc64 FFB driver code that never gets built.

The Kconfig bits were removed long ago, so we should kill off the
driver too.

Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoocfs2_dlm: Check for migrateable lockres in dlm_empty_lockres()
Sunil Mushran [Fri, 23 Mar 2007 00:08:32 +0000 (17:08 -0700)]
ocfs2_dlm: Check for migrateable lockres in dlm_empty_lockres()

In dlm_migrate_lockres(), we check upfront whether the lockres is a
candidate for migration. This patch encapsulates that code in a separate
function so that dlm_empty_lockres() can also use it during umount. This
patch addresses the umount process spinning problem.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
17 years agoocfs2_dlm: Fix lockres ref counting bug
Sunil Mushran [Fri, 23 Mar 2007 00:01:07 +0000 (17:01 -0700)]
ocfs2_dlm: Fix lockres ref counting bug

During umount, the umount thread migrates the lockres' and the dlm_thread
frees the empty lockres'. Due to a race, the reference counting on the
lockres goes awry leading to extra puts.

Signed-off-by: Sunil Mushran <sunil.mushran@oracle.com>
Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
17 years agoIB/iser: Handle aborting a command after it is sent
Erez Zilber [Sun, 25 Mar 2007 10:07:10 +0000 (12:07 +0200)]
IB/iser: Handle aborting a command after it is sent

The SCSI midlayer may abort a command that was already sent.  If the
initiator is still trying to send the command (or data-out PDUs for
that command), the QP may time out after the midlayer times
out. Therefore, when aborting the command, iSER may still have
references for the command's buffers.  When sending these PDUs, the
sends will complete with an error and their resources will be released
then.

Signed-off-by: Erez Zilber <erezz@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
17 years ago[NET]: Remove dead net/sched/Makefile entry for sch_hpfq.o.
Robert P. J. Day [Mon, 26 Mar 2007 23:20:34 +0000 (16:20 -0700)]
[NET]: Remove dead net/sched/Makefile entry for sch_hpfq.o.

Remove the worthless net/sched/Makefile entry for the non-existent
source file sch_hpfq.c.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: David S. Miller <davem@davemloft.net>