12 years ago[PATCH] Kdump documentation update: ia64 portion
Horms [Tue, 23 Jan 2007 04:40:49 +0000]
[PATCH] Kdump documentation update: ia64 portion

this patch fills in the portions for ia64 kexec.

Signed-off-by: Simon Horman <horms@verge.net.au>
Cc: "Zou, Nanhai" <nanhai.zou@intel.com>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] Kdump documentation update: kexec-tools update
Horms [Tue, 23 Jan 2007 04:40:48 +0000]
[PATCH] Kdump documentation update: kexec-tools update

Mohan Kumar suggested making kexec-tools-testing.tar.gz a link to the
latest version.  I have done this and this patch updates the documentation
accordingly.

Signed-off-by: Simon Horman <horms@verge.net.au>
Cc: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] resierfs: avoid tail packing if an inode was ever mmapped
Vladimir Saveliev [Tue, 23 Jan 2007 04:40:46 +0000]
[PATCH] resierfs: avoid tail packing if an inode was ever mmapped

This patch fixes a confusion reiserfs has for a long time.

On release file operation reiserfs used to try to pack file data stored in
last incomplete page of some files into metadata blocks.  After packing the
page got cleared with clear_page_dirty.  It did not take into account that
the page may be mmaped into other process's address space.  Recent
replacement for clear_page_dirty cancel_dirty_page found the confusion with
sanity check that page has to be not mapped.

The patch fixes the confusion by making reiserfs avoid tail packing if an
inode was ever mmapped.  reiserfs_mmap and reiserfs_file_release are
serialized with mutex in reiserfs specific inode.  reiserfs_mmap locks the
mutex and sets a bit in reiserfs specific inode flags.
reiserfs_file_release checks the bit having the mutex locked.  If bit is
set - tail packing is avoided.  This eliminates a possibility that mmapped
page gets cancel_page_dirty-ed.

Signed-off-by: Vladimir Saveliev <vs@namesys.com>
Cc: Jeff Mahoney <jeffm@suse.com>
Cc: Chris Mason <mason@suse.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] mbind: restrict nodes to the currently allowed cpuset
Christoph Lameter [Tue, 23 Jan 2007 04:40:45 +0000]
[PATCH] mbind: restrict nodes to the currently allowed cpuset

Currently one can specify an arbitrary node mask to mbind that includes
nodes not allowed.  If that is done with an interleave policy then we will
go around all the nodes.  Those outside of the currently allowed cpuset
will be redirected to the border nodes.  Interleave will then create
imbalances at the borders of the cpuset.

This patch restricts the nodes to the currently allowed cpuset.

The RFC for this patch was discussed at
http://marc.theaimsgroup.com/?t=116793842100004&r=1&w=2

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Paul Jackson <pj@sgi.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] tlclk: bug fix + misc fixes
Mark Gross [Tue, 23 Jan 2007 04:40:44 +0000]
[PATCH] tlclk: bug fix + misc fixes

The following patch fixes a few problems with the tlclk driver.
* bug in the select_amcb1_transmit_clock
* racy read sys call
* racy open sys call
* use of add_timer where mod_timer would be better
* change to the timer data parameter use

Signed-off-by: Mark Gross <mark.gross@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] fix blk_direct_IO bio preparation
Chen, Kenneth W [Tue, 23 Jan 2007 04:40:43 +0000]
[PATCH] fix blk_direct_IO bio preparation

For large size DIO that needs multiple bio, one full page worth of data was
lost at the boundary of bio's maximum sector or segment limits.  After a
bio is full and got submitted.  The outer while (nbytes) { ...  } loop will
allocate a new bio and just march on to index into next page.  It just
forgets about the page that bio_add_page() rejected when previous bio is
full.  Fix it by put the rejected page back to pvec so we pick it up again
for the next bio.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] rtc-sh: act on rtc_wkalrm.enabled when setting an alarm
Jamie Lenehan [Tue, 23 Jan 2007 04:40:41 +0000]
[PATCH] rtc-sh: act on rtc_wkalrm.enabled when setting an alarm

This fixes the SH rtc driver correctly act on the "enabled" flag when
setting an alarm.

Signed-off-by: Jamie Lenehan <lenehan@twibble.org>
Cc: David Brownell <david-b@pacbell.net>
Cc: Alessandro Zummo <alessandro.zummo@towertech.it>
Cc: Paul Mundt <lethal@linux-sh.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] KVM: fix bogus pagefault on writable pages
Avi Kivity [Tue, 23 Jan 2007 04:40:40 +0000]
[PATCH] KVM: fix bogus pagefault on writable pages

If a page is marked as dirty in the guest pte, set_pte_common() can set the
writable bit on newly-instantiated shadow pte.  This optimization avoids
a write fault after the initial read fault.

However, if a write fault instantiates the pte, fix_write_pf() incorrectly
reports the fault as a guest page fault, and the guest oopses on what appears
to be a correctly-mapped page.

Fix is to detect the condition and only report a guest page fault on a user
access to a kernel page.

With the fix, a kvm guest can survive a whole night of running the kernel
hacker's screensaver (make -j9 in a loop).

Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] KVM: x86 emulator: fix bit string instructions
Avi Kivity [Tue, 23 Jan 2007 04:40:40 +0000]
[PATCH] KVM: x86 emulator: fix bit string instructions

The various bit string instructions (bts, btc, etc.) fail to adjust the
address correctly if the bit address is beyond BITS_PER_LONG.

This bug creeped in as the emulator originally relied on cr2 to contain the
memory address; however we now decode it from the mod r/m bits, and must
adjust the offset to account for large bit indices.

The patch is rather large because it switches src and dst decoding around, so
that the bit index is available when decoding the memory address.

This fixes workloads like the FC5 installer.

Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] KVM: fix race between mmio reads and injected interrupts
Avi Kivity [Tue, 23 Jan 2007 04:40:39 +0000]
[PATCH] KVM: fix race between mmio reads and injected interrupts

The kvm mmio read path looks like:

 1. guest read faults
 2. kvm emulates read, calls emulator_read_emulated()
 3. fails as a read requires userspace help
 4. exit to userspace
 5. userspace emulates read, kvm sets vcpu->mmio_read_completed
 6. re-enter guest, fault again
 7. kvm emulates read, calls emulator_read_emulated()
 8. succeeds as vcpu->mmio_read_emulated is set
 9. instruction completes and guest is resumed

A problem surfaces if the userspace exit (step 5) also requests an interrupt
injection.  In that case, the guest does not re-execute the original
instruction, but the interrupt handler.  The next time an mmio read is
exectued (likely for a different address), step 3 will find
vcpu->mmio_read_completed set and return the value read for the original
instruction.

The problem manifested itself in a few annoying ways:
- little squares appear randomly on console when switching virtual terminals
- ne2000 fails under nfs read load
- rtl8139 complains about "pci errors" even though the device model is
  incapable of issuing them.

Fix by skipping interrupt injection if an mmio read is pending.

A better fix is to avoid re-entry into the guest, and re-emulating immediately
instead.  However that's a bit more complex.

Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] KVM: make sure there is a vcpu context loaded when destroying the mmu
Avi Kivity [Tue, 23 Jan 2007 04:40:38 +0000]
[PATCH] KVM: make sure there is a vcpu context loaded when destroying the mmu

This makes the vmwrite errors on vm shutdown go away.

Signed-off-by: Avi Kivity <avi@qumranet.com>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] paravirt: mark the paravirt_ops export internal
Ingo Molnar [Tue, 23 Jan 2007 04:40:36 +0000]
[PATCH] paravirt: mark the paravirt_ops export internal

The paravirt subsystem is still in flux so all exports from it are
definitely internal use only.  The APIs around this /will/ change.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Andi Kleen <ak@suse.de>
Cc: Zachary Amsden <zach@vmware.com>
Cc: Jeremy Fitzhardinge <jeremy@xensource.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] SubmitChecklist update
Andrew Morton [Tue, 23 Jan 2007 04:40:36 +0000]
[PATCH] SubmitChecklist update

Sing the praises of `gcc -W'.  Would have prevented that blockdev direct-IO
bug.

Cc: "Chen, Kenneth W" <kenneth.w.chen@intel.com>
Cc: "Randy.Dunlap" <rdunlap@xenotime.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] blockdev direct_io: fix signedness bug
Andrew Morton [Tue, 23 Jan 2007 04:40:35 +0000]
[PATCH] blockdev direct_io: fix signedness bug

size_t is unsigned.  IO errors aren't getting through.

Cc: "Chen, Kenneth W" <kenneth.w.chen@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] Revert nmi_known_cpu() check during boot option parsing
Venkatesh Pallipadi [Tue, 23 Jan 2007 04:40:34 +0000]
[PATCH] Revert nmi_known_cpu() check during boot option parsing

Commit f2802e7f571c05f9a901b1f5bd144aa730ccc88e and its x86 version
(b7471c6da94d30d3deadc55986cc38d1ff57f9ca) adds nmi_known_cpu() check
while parsing boot options in x86_64 and i386.

With that, "nmi_watchdog=2" stops working for me on Intel Core 2 CPU
based system.

The problem is, setup_nmi_watchdog is called while parsing the boot
option and identify_cpu is not done yet.  So, the return value of
nmi_known_cpu() is not valid at this point.

So revert that check.  This should not have any adverse effect as the
nmi_known_cpu() check is done again later in enable_lapic_nmi_watchdog().

Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Don Zickus <dzickus@redhat.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] fix "kvm: add vm exit profiling"
Andrew Morton [Tue, 23 Jan 2007 04:40:33 +0000]
[PATCH] fix "kvm: add vm exit profiling"

export profile_hits() on !SMP too.

Cc: Ingo Molnar <mingo@elte.hu>
Cc: Avi Kivity <avi@qumranet.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[PATCH] x86: fix PDA variables to work during boot
James Bottomley [Mon, 22 Jan 2007 15:18:31 +0000]
[PATCH] x86: fix PDA variables to work during boot

The current PDA code, which went in in post 2.6.19 has a flaw in that it
doesn't correctly cycle the GDT and %GS segment through the boot PDA,
the CPU PDA and finally the per-cpu PDA.

The bug generally doesn't show up if the boot CPU id is zero, but
everything falls apart for a non zero boot CPU id.  The basically kills
voyager which is perfectly capable of doing non zero CPU id boots, so
voyager currently won't boot without this.

The fix is to be careful and actually do the GDT setups correctly.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
Cc: Andi Kleen <ak@suse.de>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years agoMerge git://git.infradead.org/mtd-2.6
Linus Torvalds [Tue, 23 Jan 2007 03:32:13 +0000]
Merge git://git.infradead.org/mtd-2.6

* git://git.infradead.org/mtd-2.6: (84 commits)
  [JFFS2] debug.h: include <linux/sched.h> for current->pid
  [MTD] OneNAND: Handle DDP chip boundary during read-while-load
  [MTD] OneNAND: return ecc error code only when 2-bit ecc occurs
  [MTD] OneNAND: Implement read-while-load
  [MTD] OneNAND: fix onenand_wait bug in read ecc error
  [MTD] OneNAND: release CPU in cycles
  [MTD] OneNAND: add subpage write support
  [MTD] OneNAND: fix onenand_wait bug
  [JFFS2] use the ref_offset macro
  [JFFS2] Reschedule in loops
  [JFFS2] Fix error-path leak in summary scan
  [JFFS2] add cond_resched() when garbage collecting deletion dirent
  [MTD] Nuke IVR leftovers
  [MTD] OneNAND: fix oob handling in recent oob patch
  [MTD] Fix ssfdc blksize typo
  [JFFS2] replace kmalloc+memset with kzalloc
  [MTD] Fix SSFDC build for variable blocksize.
  [MTD] ESB2ROM uses PCI
  [MTD] of_device-based physmap driver
  [MTD] Support combined RedBoot FIS directory and configuration area
  ...

12 years ago[PATCH] vmx: Fix register constraint in launch code
Herbert Xu [Tue, 23 Jan 2007 03:10:00 +0000]
[PATCH] vmx: Fix register constraint in launch code

Both "=r" and "=g" breaks my build on i386:

  $ make
CC [M] drivers/kvm/vmx.o
  {standard input}: Assembler messages:
  {standard input}:3318: Error: bad register name `%sil'
  make[1]: *** [drivers/kvm/vmx.o] Error 1
  make: *** [_module_drivers/kvm] Error 2

The reason is that setbe requires an 8-bit register but "=r" does not
constrain the target register to be one that has an 8-bit version on
i386.

According to

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=10153

the correct constraint is "=q".

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6
Linus Torvalds [Mon, 22 Jan 2007 21:14:37 +0000]
Merge /pub/scm/linux/kernel/git/gregkh/pci-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6:
  PCI: fix pci-driver kernel-doc
  PCI: rework Documentation/pci.txt
  PCI: Unhide the SMBus on the Asus P4P800-X

12 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Mon, 22 Jan 2007 21:14:16 +0000]
Merge /pub/scm/linux/kernel/git/gregkh/usb-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6:
  USB: unusual_devs.h entry for nokia 6233
  USB: Fix for typo in ohci-ep93xx.c
  USB: disable USB_MULTITHREAD_PROBE
  USB: add vendor/device id for Option GT Max 3.6 cards
  USB: unusual_devs.h for 0x046b:ff40
  USB: make usbhid ignore Imation Disc Stakka
  USB: rndis_host: fix crash while probing a Nokia S60 mobile
  USB: asix: Detect internal PHY and enable/use accordingly
  usbtouchscreen: make ITM screens report BTN_TOUCH as zero when not touched

12 years agoPCI: fix pci-driver kernel-doc
Randy Dunlap [Wed, 10 Jan 2007 05:41:01 +0000]
PCI: fix pci-driver kernel-doc

Function short description should be on only one line.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoPCI: rework Documentation/pci.txt
Grant Grundler [Mon, 25 Dec 2006 08:06:35 +0000]
PCI: rework Documentation/pci.txt

Rewrite Documentation/pci.txt:
o restructure document to match how API is used when writing init code.
o update to reflect changes in struct pci_driver function pointers.
o removed language on "new style vs old style" device discovery.
  "Old style" is now deprecated. Don't use it. Left description in
  to document existing driver behaviors.
o add section "Legacy I/O Port free driver" by Kenji Kaneshige
  http://lkml.org/lkml/2006/11/22/25
  (renamed to "pci_enable_device_bars() and Legacy I/O Port space")
o add "MMIO space and write posting" section to help avoid common pitfall
  when converting drivers from IO Port space to MMIO space.
  Orignally posted http://lkml.org/lkml/2006/2/27/24
o many typo/grammer/spelling corrections from Randy Dunlap
o two more spelling corrections from Stephan Richter
o fix CodingStyle as per Randy Dunlap

Signed-off-by: Grant Grundler <grundler@parisc-linux.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoPCI: Unhide the SMBus on the Asus P4P800-X
Jean Delvare [Fri, 5 Jan 2007 08:17:56 +0000]
PCI: Unhide the SMBus on the Asus P4P800-X

Unhide the SMBus on the Asus P4P800-X (and probably some other
models of the family.) This gives access to the memory module SPD
EEPROMs.

Thanks to Winbond for supporting the lm-sensors project with the
donation of this motherboard.

Signed-off-by: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoUSB: unusual_devs.h entry for nokia 6233
Manuel Osdoba [Thu, 18 Jan 2007 20:28:45 +0000]
USB: unusual_devs.h entry for nokia 6233

In appendix a patch for the nokia 6233 mobile phone is included.
The patch is against 2.6.20-rc5. It is my first patch. Hopefully it has
the right format. The code makes my nokia 6233 on my computer work.

From: Manuel Osdoba <manuel.osdoba@tu-ilmenau.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoUSB: Fix for typo in ohci-ep93xx.c
Petr Stetiar [Wed, 17 Jan 2007 14:30:39 +0000]
USB: Fix for typo in ohci-ep93xx.c

Attached patch fixes typo in USB driver reported by Chase Douglas on linux-cirrus mailing
list. http://www.freelists.org/archives/linux-cirrus/12-2006/msg00003.html

Signed-off-by: Petr Stetiar <ynezz@true.cz>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoUSB: disable USB_MULTITHREAD_PROBE
Greg Kroah-Hartman [Tue, 16 Jan 2007 23:41:03 +0000]
USB: disable USB_MULTITHREAD_PROBE

Disable the USB_MULTITHREAD_PROBE option because it causes crashes on
people's machines and they never remember to actually read the config
help files.

No one likes this, everyone hates it, I'm going to go eat worms...

The full logic will be ripped out later.

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

12 years agoUSB: add vendor/device id for Option GT Max 3.6 cards
garrett_damore@tadpole.com [Sat, 13 Jan 2007 01:10:28 +0000]
USB: add vendor/device id for Option GT Max 3.6 cards

This fixes http://bugzilla.kernel.org/show_bug.cgi?id=7814

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

12 years agoUSB: unusual_devs.h for 0x046b:ff40
Pete Zaitcev [Sat, 13 Jan 2007 23:17:46 +0000]
USB: unusual_devs.h for 0x046b:ff40

American Megatrends did something wrong in their floppy emulator. It breaks
with both kinds of MODE SENSE which our stack sends. Alan and I tried a few
tweaks, and got LUNs sensed right, but US_FL_NO_WP_DETECT is still needed.

I set the firmware bracket to 1.00 exactly, in case AMI or Sun fix it with a
firmware update. Hey, you never know.

Signed-off-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Phil Dibowitz <phil@ipom.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoUSB: make usbhid ignore Imation Disc Stakka
Oliver Neukum [Thu, 11 Jan 2007 09:14:33 +0000]
USB: make usbhid ignore Imation Disc Stakka

on request of the sourceforge project for this device, a kind of
robotized CD storage, it should be ignored by the generic driver.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoUSB: rndis_host: fix crash while probing a Nokia S60 mobile
Daniel Gollub [Tue, 16 Jan 2007 10:03:01 +0000]
USB: rndis_host: fix crash while probing a Nokia S60 mobile

Bug fix for driver rndis_host which fixes rndis_host probing certain
Nokia S60 (Series 60) mobiles. While the rndis_host get probed by usbnet
and tries to bind the Nokia mobile the bind is going to fail. The
rndis_host module tries to release the device, in a wrong way, which
cause the oops.

Fixes Bugzilla #7201

Signed-off-by: Daniel Gollub <dgollub@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoUSB: asix: Detect internal PHY and enable/use accordingly
Andres Salomon [Thu, 11 Jan 2007 23:39:16 +0000]
USB: asix: Detect internal PHY and enable/use accordingly

Different AX88772 dongles use different PHYs; the chip is capable of using
both a primary and secondary PHY, and supports an internal and external PHY.

It appears that some DUB-E100 devices use the internal PHY, so trying to use
an external one will not work (note that this is different across revisions,
as well; the "A" and "B" revs of the DUB-E100 use different PHYs!).  The data
sheet for the AX88772 chip specifies that the internal PHY id will be 0x10,
so if that's read from the EEPROM, we should use that rather than attempting
to use an external PHY.

Thanks to Mitch Bradley for pointing this out!

Signed-off-by: Andres Salomon <dilinger@debian.org>
Cc: David Hollis <dhollis@davehollis.com>
Cc: Chris Ball <cjb@laptop.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agousbtouchscreen: make ITM screens report BTN_TOUCH as zero when not touched
Daniel Ritz [Mon, 11 Dec 2006 17:17:45 +0000]
usbtouchscreen: make ITM screens report BTN_TOUCH as zero when not touched

ITM screens send invalid x/y data when not touched. this was fixes a while ago
but the problem is if the screen is not touched anymore the driver never does
not report BTN_TOUCH as zero. fix it by sending the report with the last valid
coordinates when pressure is released.

Signed-off-by: Daniel Ritz <daniel.ritz@gmx.ch>
Cc: J.P. Delport <jpdelport@csir.co.za>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

12 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mfashe...
Linus Torvalds [Mon, 22 Jan 2007 19:33:40 +0000]
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: Add backup superblock info to ocfs2_fs.h
  ocfs2: cleanup ocfs2_iget() errors
  ocfs2: Directory c/mtime update fixes
  ocfs2: Don't print errors when following symlinks

12 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Mon, 22 Jan 2007 19:31:11 +0000]
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:
  sis190: failure to set the MAC address from EEPROM
  Fix phy_read/write redefinition errors in ucc_geth_phy.c
  Update ucc_geth.c for new workqueue structure
  myri10ge: update driver version to 1.2.0
  myri10ge: check that we can get an irq
  myri10ge: make wc_fifo usage load-time tunable
  8139cp: Don't blindly enable interrupts
  [PATCH] bcm43xx: Fix failure to deliver PCI-E interrupts

12 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Mon, 22 Jan 2007 19:30:47 +0000]
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:
  sata_mv HighPoint 2310 support (88SX7042)
  libata: fix handling of port actions in per-dev action mask
  libata: initialize qc->dma_dir to DMA_NONE
  sata_via: add PCI ID 0x5337
  libata doc: "error : unterminated entity reference exceptions"

12 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6
Linus Torvalds [Mon, 22 Jan 2007 18:35:34 +0000]
Merge /linux/kernel/git/jejb/scsi-rc-fixes-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6: (30 commits)
  [SCSI] qla1280: set residual correctly
  [SCSI] fusion: bump version
  [SCSI] fusion: MODULE_VERSION support
  [SCSI] fusion: power pc and miscellaneous bug fixs
  [SCSI] fusion: fibre channel: return DID_ERROR for MPI_IOCSTATUS_SCSI_IOC_TERMINATED
  [SCSI] megaraid_sas: Update module author
  [SCSI] 3ware 8000 serialize reset code
  [SCSI] sr: fix error code check in sr_block_ioctl()
  [SCSI] scsi: lpfc error path fix
  [SCSI] aacraid: Product List Update
  [SCSI] libiscsi: fix senselen calculation
  [SCSI] iscsi: simplify IPv6 and IPv4 address printing
  [SCSI] iscsi: newline in printk
  [SCSI] iscsi: fix crypto_alloc_hash() error check
  [SCSI] iscsi: fix 2.6.19 data digest calculation bug
  [SCSI] scsi_scan: fix report lun problems with CDROM or RBC devices
  [SCSI] qla2xxx: Update version number to 8.01.07-k4.
  [SCSI] qla2xxx: Use generic isp_ops.fw_dump() function.
  [SCSI] qla2xxx: Perform a fw-dump when an ISP23xx RISC-paused state is detected.
  [SCSI] qla2xxx: Correct reset handling logic.
  ...

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Mon, 22 Jan 2007 17:23:23 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  hid-core.c: Adds GTCO CalComp Interwrite IPanel PIDs to blacklist
  HID: put usb_interface instead of usb_device into hid->dev to fix udevinfo breakage
  HID: add missing RX, RZ and RY enum values to hid-debug output
  HID: hid/hid-input.c doesn't need to include linux/usb/input.h
  HID: compilation fix when DEBUG_DATA is defined
  HID: proper LED-mapping for SpaceNavigator
  HID: update MAINTAINERS entry for USB-HID
  HID: GEYSER4_ISO needs quirk
  HID: fix some ARM builds due to HID brokenness - make USB_HID depend on INPUT

12 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc
Linus Torvalds [Mon, 22 Jan 2007 17:21:02 +0000]
Merge branch 'merge' of git://git./linux/kernel/git/paulus/powerpc

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/paulus/powerpc:
  [POWERPC] Update defconfigs
  [POWERPC] atomic_dec_if_positive sign extension fix
  [POWERPC] Fix OF node refcnt underflow in 836x and 832x platform code
  [POWERPC] Make it blatantly clear; mpc5200 device tree is not yet stable
  [POWERPC] Fix broken DMA on non-LPAR pSeries
  [POWERPC] Fix cell's mmio nvram to properly parse device tree
  [POWERPC] Remove bogus sanity check in pci -> OF node code

12 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Mon, 22 Jan 2007 17:07:00 +0000]
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Delete duplicate call to load_irq_save.
  [MIPS] SMTC: Fix cp0 hazard.

12 years agoMerge branch 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6
Linus Torvalds [Mon, 22 Jan 2007 16:57:03 +0000]
Merge branch 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6

* 'hwmon-for-linus' of git://jdelvare.pck.nerim.net/jdelvare-2.6:
  hwmon/w83793: Hide invalid VID readings
  hwmon/w83793: Fix the fan input detection
  hwmon/w83793: Ignore disabled temperature channels
  hwmon: Fix the VRD 11 decoding
  hwmon/w83793: Remove the description of AMDSI and update the voltage formula

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/aia21/ntfs-2.6
Linus Torvalds [Mon, 22 Jan 2007 16:55:56 +0000]
Merge git://git./linux/kernel/git/aia21/ntfs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/aia21/ntfs-2.6:
  NTFS: Forgot to bump version number in makefile to 2.1.28...
  NTFS: 2.1.28 - Fix deadlock reported by Sergey Vlasov due to ntfs_put_inode().

12 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb
Linus Torvalds [Mon, 22 Jan 2007 16:55:28 +0000]
Merge branch 'master' of git://git./linux/kernel/git/mchehab/v4l-dvb

* 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb:
  V4L/DVB (5023): Fix compilation on ppc32 architecture
  V4L/DVB (5071): Tveeprom: autodetect LG TAPC G701D as tuner type 37
  V4L/DVB (5069): Fix bttv and friends on 64bit machines with lots of memory
  V4L/DVB (5033): MSI TV@nywhere Plus fixes
  V4L/DVB (5029): Ks0127 status flags
  V4L/DVB (5024): Fix quickcam communicator driver for big endian architectures
  V4L/DVB (5021): Cx88xx: Fix lockup on suspend
  V4L/DVB (5020): Fix: disable interrupts while at KM_BOUNCE_READ
  V4L/DVB (5019): Fix the frame->grabstate update in read() entry point.

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/drzeus/mmc
Linus Torvalds [Mon, 22 Jan 2007 16:54:48 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/drzeus/mmc

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/drzeus/mmc:
  mmc: Correct definition of R6
  omap: Update MMC response types

12 years agoRevert "[PATCH] Fix up mmap_kmem"
Linus Torvalds [Mon, 22 Jan 2007 16:53:24 +0000]
Revert "[PATCH] Fix up mmap_kmem"

This reverts commit 99a10a60ba9bedcf5d70ef81414d3e03816afa3f.

As per Hugh Dickins:

  "Nadia Derbey has reported that mmap of /dev/kmem no longer works with
   the kernel virtual address as offset, and Franck has confirmed that
   his patch came from a misunderstanding of what an offset means to
   /dev/kmem - whereas his patch description seems to say that he was
   correcting the offset on a few plaforms, there was no such problem to
   correct, and his patch was in fact changing its API on all platforms."

Suggested-by: Hugh Dickins <hugh@veritas.com>
Cc: Franck Bui-Huu <fbuihuu@gmail.com>
Cc: Nadia Derbey <Nadia.Derbey@bull.net>
Cc: Andi Kleen <ak@suse.de>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

12 years ago[POWERPC] Update defconfigs
Paul Mackerras [Mon, 22 Jan 2007 11:39:24 +0000]
[POWERPC] Update defconfigs

Mostly took the defaults, except tried to get the netfilter options
more or less as they were before.

Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] atomic_dec_if_positive sign extension fix
Robert Jennings [Wed, 17 Jan 2007 16:50:20 +0000]
[POWERPC] atomic_dec_if_positive sign extension fix

On 64-bit machines, if an atomic counter is explicitly set to a
negative value, the atomic_dec_if_positive function will decrement and
store the next smallest value in the atomic counter, contrary to its
intended operation.

The comparison to determine if the decrement will make the result
negative was done by the "addic." instruction, which operates on a
64-bit value, namely the zero-extended word loaded from the atomic
variable.  This patch uses an explicit word compare (cmpwi) and
changes the addic. to an addi (also changing "=&r" to "=&b" so that r0
isn't used, and addi doesn't become li).

This also fixes a bug for both 32-bit and 64-bit in that previously
0x80000000 was considered positive, since the result after
decrementing is positive.  Now it is considered negative.

Also, I clarify the return value in the comments just to make it clear
that the value returned is always the decremented value, even if that
value is not stored back to the atomic counter.

Signed-off-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Fix OF node refcnt underflow in 836x and 832x platform code
Li Yang [Wed, 17 Jan 2007 06:42:22 +0000]
[POWERPC] Fix OF node refcnt underflow in 836x and 832x platform code

Incorrect use of of_find_node_by_name() causes of_node_put()
on a node which has already been put.  It causes the refcount of
the node to underflow, which triggers the WARN_ON in kref_get
for 836x and 832x.  This fixes it.

Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Make it blatantly clear; mpc5200 device tree is not yet stable
Grant Likely [Fri, 19 Jan 2007 07:00:14 +0000]
[POWERPC] Make it blatantly clear; mpc5200 device tree is not yet stable

Documentation-only change.  The 5200 device tree layout has not yet
stablized, so nobody should depend on the layout of the tree.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Sylvain Munaut <tnt@246tNt.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Fix broken DMA on non-LPAR pSeries
Linas Vepstas [Thu, 11 Jan 2007 01:16:29 +0000]
[POWERPC] Fix broken DMA on non-LPAR pSeries

It appears that the iommu table address is never stored, and thus
never found, on non-lpar systems. Thus, for example, during boot:

<7>[   93.067916] PCI: Scanning bus 0001:41
<7>[   93.068542] PCI: Found 0001:41:01.0 [8086/100f] 000200 00
<7>[   93.068550] PCI: Calling quirk c0000000007822e0 for 0001:41:01.0
<7>[   93.069815] PCI: Fixups for bus 0001:41
<4>[   93.070167] iommu: Device 0001:41:01.0 has no iommu table
<7>[   93.070251] PCI: Bus scan for 0001:41 returning with max=41

No iommu table? How can that be? Well, circa line 471 of
arch/powerpc/platforms/pseries/iommu.c we see the code:

   while (dn && PCI_DN(dn) && PCI_DN(dn)->iommu_table == NULL)
      dn = dn->parent;

and a few lines later is the surprising print statement about
the missing table.  Seems that this loop ran unto the end, never
once finding a non-null PCI_DN(dn)->iommu_table.

The problem can be found a few lines earlier: it sems that the
value of PCI_DN(dn)->iommu_table is never ever set. Thus, the
patch sets it.

The patch was tested on a Power4 system running in full system
partition mode, which is where I saw the problem. It works; I've
not done any wider testing. Had a brief discussion on this on irc.

Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Fix cell's mmio nvram to properly parse device tree
Benjamin Herrenschmidt [Thu, 11 Jan 2007 05:08:41 +0000]
[POWERPC] Fix cell's mmio nvram to properly parse device tree

The mmio nvram driver (used by cell only atm) isn't properly parsing
the device-tree, meaning that nvram isn't found correctly on the new
Cell blades.  It works ok for old blades where the nvram is at the
root of the device tree but fails on Malta and CAB when it's hanging
off axon.  This fixes it by using the proper OF parsing functions.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Remove bogus sanity check in pci -> OF node code
Benjamin Herrenschmidt [Thu, 11 Jan 2007 05:38:22 +0000]
[POWERPC] Remove bogus sanity check in pci -> OF node code

The new implementation of pci_device_to_OF_node() on ppc32 has a bogus
sanity check in it that can cause oopses at boot when no device node is
present, and might hit correct cases with older/weird apple device-trees
where they have the type "vci" for the chaos bridge.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years agoocfs2: Add backup superblock info to ocfs2_fs.h
Mark Fasheh [Sun, 21 Jan 2007 22:44:59 +0000]
ocfs2: Add backup superblock info to ocfs2_fs.h

This synchronizes us with recent ocfs2-tools changes.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>

12 years agoocfs2: cleanup ocfs2_iget() errors
Mark Fasheh [Thu, 4 Jan 2007 01:06:59 +0000]
ocfs2: cleanup ocfs2_iget() errors

Get rid of some error prints in the ocfs2_iget() path from
ocfs2_get_dentry(). NFSD can easily cause us to read stale inodes.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>

12 years agoocfs2: Directory c/mtime update fixes
Mark Fasheh [Wed, 3 Jan 2007 01:59:40 +0000]
ocfs2: Directory c/mtime update fixes

ocfs2 wasn't updating c/mtime on directories during dirent
creation/deletion. Fix ocfs2_unlink(), ocfs2_rename() and
__ocfs2_add_entry() by adding the proper code to update the struct inode and
push the change out to disk.

This helps rename/unlink on nfs exported file systems in particular as those
clients compare directory time values to avoid a full re-reading a directory
which hasn't changed.

ocfs2_rename() loses some superfluous error handling as a result of this
patch.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>

12 years agoocfs2: Don't print errors when following symlinks
Mark Fasheh [Thu, 4 Jan 2007 01:25:40 +0000]
ocfs2: Don't print errors when following symlinks

We shouldn't print errors returned from vfs_follow_link(). This was causing
spurious errors to show up in the logs.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>

12 years agohid-core.c: Adds GTCO CalComp Interwrite IPanel PIDs to blacklist
Jeremy Roberson [Fri, 19 Jan 2007 22:44:18 +0000]
hid-core.c: Adds GTCO CalComp Interwrite IPanel PIDs to blacklist

Adds GTCO CalComp Interwrite IPanels to the hid-core.c blacklist.

Signed-off-by: Jeremy A. Roberson <jroberson@gtcocalcomp.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: put usb_interface instead of usb_device into hid->dev to fix udevinfo breakage
Anssi Hannula [Fri, 19 Jan 2007 17:28:17 +0000]
HID: put usb_interface instead of usb_device into hid->dev to fix udevinfo breakage

The commit 4916b3a57fc94664677d439b911b8aaf86c7ec23 introduced a
hid regression between 2.6.19 and 2.6.20-rc1. The device put in
input_dev->cdev is now of type usb_device instead of usb_interface.

Before:
> # readlink -f /sys/class/input/input6/event4/device
> /sys/devices/pci0000:00/0000:00:10.0/usb2/2-1/2-1:1.1
After:
> # readlink -f /sys/class/input/input3/event3/device
> /sys/devices/pci0000:00/0000:00:10.0/usb1/1-1

This causes breakage:
> # udevinfo -q all -n /dev/input/event3
> P: /class/input/input3/event3
> N: input/event3
> S: input/by-path/pci-1-1--event-
> E: ID_SERIAL=noserial
> E: ID_PATH=pci-1-1-

No ID_MODEL, ID_VENDOR, ID_REVISION, ID_TYPE etc etc.

Fix this by assigning the intf->dev into hid->dev, and fixing
all the users.

Signed-off-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: add missing RX, RZ and RY enum values to hid-debug output
Simon Budig [Mon, 15 Jan 2007 17:34:32 +0000]
HID: add missing RX, RZ and RY enum values to hid-debug output

This trivial change adds some missing enum values to the hid-debug output.

Signed-off-by: Simon Budig <simon@budig.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: hid/hid-input.c doesn't need to include linux/usb/input.h
Jiri Kosina [Mon, 15 Jan 2007 23:30:42 +0000]
HID: hid/hid-input.c doesn't need to include linux/usb/input.h

There is no reason for generic hid-input.c to include usb-specific
input.h.

As a sideeffect, this also fixes warning of redefinition of dbg()
macro, when hid-input.c is compiled with DEBUG (as there is a clash
between dbg() from hid.h and usb/input.h).

Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: compilation fix when DEBUG_DATA is defined
Jiri Kosina [Mon, 15 Jan 2007 22:53:05 +0000]
HID: compilation fix when DEBUG_DATA is defined

hid/hid-core.c references 'len' variable when DEBUG_DATA is defined,
but the actual name of the variable is 'size'. Fix it.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: proper LED-mapping for SpaceNavigator
Simon Budig [Mon, 15 Jan 2007 16:28:47 +0000]
HID: proper LED-mapping for SpaceNavigator

This change introduces a mapping for LED indicators between the HID
specification and the Linux input subsystem. The previous code properly
mapped the LEDs relevant for Keyboards, but garbeled the remaining ones.
With this change all LED enums from the input system get mapped to more
or less equivalent LED numbers from the HID specification.

This patch also ensures that the unused bits in a HID report to the
device are zeroed out. This makes the 3Dconnexion SpaceNavigator fully
usable with the linux input system.

Signed-off-by: Simon Budig <simon@budig.de>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: update MAINTAINERS entry for USB-HID
Jiri Kosina [Mon, 15 Jan 2007 08:56:21 +0000]
HID: update MAINTAINERS entry for USB-HID

Change USB-HID maintainer from Vojtech Pavlik to Jiri Kosina.

Acked-by: Vojtech Pavlik <vojtech@suse.cz>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: GEYSER4_ISO needs quirk
Adrian Friedli [Sun, 14 Jan 2007 21:12:43 +0000]
HID: GEYSER4_ISO needs quirk

I've got a newer MacBook with core2duo. Two keys on the keyboard are
swapped, "unswaping" works with the same trick as GEYSER3_ISO.

Signed-off-by: Adrian Friedli <masteradi@gmx.ch>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agoHID: fix some ARM builds due to HID brokenness - make USB_HID depend on INPUT
Russell King [Fri, 12 Jan 2007 21:53:51 +0000]
HID: fix some ARM builds due to HID brokenness - make USB_HID depend on INPUT

We have USB_HID _newly_ selected in configurations which didn't
have it before, which overrides CONFIG_HID and builds HID without
input support.

Nevertheless, here's a patch to solve more of the same that my original
patch attempted to solve.  The original patch is still required.  Seems
to solve the final instance of this problem here.

Signed-off-by: Russell King <rmk+lkml@arm.linux.org.uk>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>

12 years agosata_mv HighPoint 2310 support (88SX7042)
Olof Johansson [Fri, 19 Jan 2007 00:39:59 +0000]
sata_mv HighPoint 2310 support (88SX7042)

With the following patch, my HighPoint 2310 with a Marvell 88SX7042 on
it seems to work OK.

The controller only has 4 ports, with MV_FLAG_DUAL_HC it seems to init 8
ports and fails miserably at probe time. There are no other devices mapped
to that chip, maybe it was just incorrectly specified in the first place?

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agolibata: fix handling of port actions in per-dev action mask
Tejun Heo [Thu, 18 Jan 2007 08:22:18 +0000]
libata: fix handling of port actions in per-dev action mask

libata EH ignores port-wide actions in per-dev action mask.  However,
device resume requests EH_SOFTRESET using per-dev action mask.  Under
certain circumstances, this results in not resetting frozen port after
resuming which causes failure of all commands.

This patch allows port-wide actions to be requested in per-dev action
mask.  Before EH recovery starts, port-wide actions will be collected.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agolibata: initialize qc->dma_dir to DMA_NONE
Tejun Heo [Wed, 17 Jan 2007 02:34:02 +0000]
libata: initialize qc->dma_dir to DMA_NONE

libata didn't used to init qc->dma_dir to any specific value on qc
initialization and command translation path didn't set qc->dma_dir if
the command doesn't need data transfer.  This made non-data commands
to have random qc->dma_dir.

This usually doesn't cause problem because LLDs usually check
qc->protocol first and look at qc->dma_dir iff the command needs data
transfer but this doesn't hold for all LLDs.

It might be worthwhile to rename qc->dma_dir to qc->data_dir as we use
the field to tag data direction for both PIO and DMA protocols.

This problem has been spotted by James Bottomley.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Cc: James Bottomley <James.Bottomley@SteelEye.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agosata_via: add PCI ID 0x5337
Luca Pedrielli [Tue, 16 Jan 2007 03:55:04 +0000]
sata_via: add PCI ID 0x5337

Add PCI ID 0x5337 to supported PCI ID.  This is VT8237 in IDE mode.

Signed-off-by: Luca Pedrielli <luca.pedrielli@barradev.it>
Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agolibata doc: "error : unterminated entity reference exceptions"
Robert P. J. Day [Mon, 8 Jan 2007 07:22:03 +0000]
libata doc: "error : unterminated entity reference exceptions"

  Fix libata.tmpl to not generate "error : unterminated entity
reference  exceptions" errors anymore when running "make htmldocs".

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agosis190: failure to set the MAC address from EEPROM
Francois Romieu [Thu, 18 Jan 2007 22:22:23 +0000]
sis190: failure to set the MAC address from EEPROM

Fix from http://bugzilla.kernel.org/show_bug.cgi?id=7747

Signed-off-by: Andrew Morton <akpm@osdl.org>
Cc: <sleepy@mike-neko.net>
Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years ago[MIPS] Delete duplicate call to load_irq_save.
Ralf Baechle [Tue, 19 Dec 2006 01:12:42 +0000]
[MIPS] Delete duplicate call to load_irq_save.

This call may have resulted to local_tlb_flush_range returning with
interrupts disabled resulting in excessive interrupt latency.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

12 years ago[MIPS] SMTC: Fix cp0 hazard.
Ralf Baechle [Wed, 17 Jan 2007 18:58:44 +0000]
[MIPS] SMTC: Fix cp0 hazard.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>

12 years agohwmon/w83793: Hide invalid VID readings
Gong Jun [Thu, 18 Jan 2007 21:14:24 +0000]
hwmon/w83793: Hide invalid VID readings

Ignore the VID readings when the motherboard has not designed
the function.

Signed-off-by: Gong Jun <jgong@winbond.com>
Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

12 years agohwmon/w83793: Fix the fan input detection
Rudolf Marek [Thu, 18 Jan 2007 21:14:24 +0000]
hwmon/w83793: Fix the fan input detection

Catch the cases when alternative pins are used to route the
fan9-12 input.

Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Gong Jun <jgong@winbond.com>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

12 years agohwmon/w83793: Ignore disabled temperature channels
Gong Jun [Thu, 18 Jan 2007 21:14:24 +0000]
hwmon/w83793: Ignore disabled temperature channels

Ignore the temperature readings when its channel is disabled,
ignore AMDSI readings.

Signed-off-by: Gong Jun <jgong@winbond.com>
Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

12 years agohwmon: Fix the VRD 11 decoding
Jean Delvare [Thu, 18 Jan 2007 21:14:23 +0000]
hwmon: Fix the VRD 11 decoding

I wonder how we came up with such a broken test in the first place.

Signed-off-by: Jean Delvare <khali@linux-fr.org>

12 years agohwmon/w83793: Remove the description of AMDSI and update the voltage formula
Gong Jun [Thu, 18 Jan 2007 21:14:23 +0000]
hwmon/w83793: Remove the description of AMDSI and update the voltage formula

Fix the driver to match the information in datasheet 1.0. AMD
SI interface is marked as reserved, computing formula for 5VDD
and 5VSB is updated.

Signed-off-by: Gong Jun <jgong@winbond.com>
Signed-off-by: Rudolf Marek <r.marek@assembler.cz>
Signed-off-by: Jean Delvare <khali@linux-fr.org>

12 years agoFix phy_read/write redefinition errors in ucc_geth_phy.c
Timur Tabi [Thu, 18 Jan 2007 17:09:35 +0000]
Fix phy_read/write redefinition errors in ucc_geth_phy.c

The local versions of phy_read() and phy_write() in ucc_geth_phy.c conflict
with the prototypes in include/linux/phy.h, so this patch renames them,
moves them to the top of the file (while eliminating the redundant prototype),
and makes them static.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agoUpdate ucc_geth.c for new workqueue structure
Timur Tabi [Tue, 9 Jan 2007 18:31:38 +0000]
Update ucc_geth.c for new workqueue structure

The workqueue interface changed with David Howell's patch on 11/22/2006
(SHA 65f27f38446e1976cc98fd3004b110fedcddd189).  Several drivers were
updated with that patch to handle the new interface, but ucc_geth.c
was not one of them.  This patch updates ucc_geth.c to support the new
model.

A compiler warning in set_mac_addr() was also fixed.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agomyri10ge: update driver version to 1.2.0
Brice Goglin [Tue, 9 Jan 2007 20:05:35 +0000]
myri10ge: update driver version to 1.2.0

Driver is now at version 1.2.0.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agomyri10ge: check that we can get an irq
Brice Goglin [Tue, 9 Jan 2007 20:05:04 +0000]
myri10ge: check that we can get an irq

Now that IRQ allocation is done in myri10ge_open(), we want to still
check when loading the driver that IRQ allocation could succeed later.

Additionaly, we fix the initialization and printing of netdev->irq.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agomyri10ge: make wc_fifo usage load-time tunable
Brice Goglin [Tue, 9 Jan 2007 20:04:25 +0000]
myri10ge: make wc_fifo usage load-time tunable

Under some circumstances, using WC without the WC fifo is faster.
So we make it possible to tune wc_fifo with a module parameter.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years ago8139cp: Don't blindly enable interrupts
Chris Lalancette [Tue, 16 Jan 2007 21:41:44 +0000]
8139cp: Don't blindly enable interrupts

Similar to this commit:

http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=d15e9c4d9a75702b30e00cdf95c71c88e3f3f51e

It's not safe in cp_start_xmit to blindly call spin_lock_irq and then
spin_unlock_irq, since it may very well be the case that cp_start_xmit
was called with interrupts already disabled (I came across this bug in
the context of netdump in RedHat kernels, but the same issue holds, for
example, in netconsole). Therefore, replace all instances of
spin_lock_irq and spin_unlock_irq with spin_lock_irqsave and
spin_unlock_irqrestore, respectively, in cp_start_xmit(). I tested this
against a fully-virtualized Xen guest using netdump, which happens to
use the 8139cp driver to talk to the emulated hardware. I don't have a
real piece of 8139cp hardware to test on, so someone else will have to
do that.

Signed-off-by: Chris Lalancette <clalance@redhat.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>

12 years agoNTFS: Forgot to bump version number in makefile to 2.1.28...
Anton Altaparmakov [Thu, 18 Jan 2007 10:28:18 +0000]
NTFS: Forgot to bump version number in makefile to 2.1.28...

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

12 years agoNTFS: 2.1.28 - Fix deadlock reported by Sergey Vlasov due to ntfs_put_inode().
Anton Altaparmakov [Thu, 18 Jan 2007 09:42:48 +0000]
NTFS: 2.1.28 - Fix deadlock reported by Sergey Vlasov due to ntfs_put_inode().

- Fix deadlock in fs/ntfs/inode.c::ntfs_put_inode().  Thanks to Sergey
  Vlasov for the report and detailed analysis of the deadlock.  The fix
  involved getting rid of ntfs_put_inode() altogether and hence NTFS no
  longer has a ->put_inode super operation.

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>

12 years ago[PATCH] bcm43xx: Fix failure to deliver PCI-E interrupts
Larry Finger [Fri, 12 Jan 2007 18:08:50 +0000]
[PATCH] bcm43xx: Fix failure to deliver PCI-E interrupts

The PCI-E modifications to bcm43xx do not set up the interrupt vector
correctly. Tested with BCM4311 (PCI-E) on x86_64 and BCM4306 (PCI) on i386.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

12 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux-2.6
David Woodhouse [Wed, 17 Jan 2007 23:34:51 +0000]
Merge branch 'master' of git://git./linux/kernel/git/torvalds/linux-2.6

12 years agoV4L/DVB (5023): Fix compilation on ppc32 architecture
Mauro Carvalho Chehab [Sun, 7 Jan 2007 11:51:05 +0000]
V4L/DVB (5023): Fix compilation on ppc32 architecture

There's a problem, pointed by Meelis Roos <mroos@linux.ee>, that, on ppc32 arch,
with some gcc versions (noticed with prerelease 4.1.2 20061115), compilation
fails, due the lack of __ucmpdi2 to do the required 64-bit comparision.
This patch takes some sugestions made by Andrew Morton <akpm@osdl.org>,
Stelian Pop <stelian@popies.net> and Segher Boessenkool <segher@kernel.crashing.org>

Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5071): Tveeprom: autodetect LG TAPC G701D as tuner type 37
Michael Krufky [Fri, 12 Jan 2007 20:38:05 +0000]
V4L/DVB (5071): Tveeprom: autodetect LG TAPC G701D as tuner type 37

autodetect LG TAPC G701D as tuner type 37.
Thanks to Adonis Papas, for pointing out the missing autodetection
for this tuner.

Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5069): Fix bttv and friends on 64bit machines with lots of memory
Gerd Hoffmann [Sat, 13 Jan 2007 00:26:32 +0000]
V4L/DVB (5069): Fix bttv and friends on 64bit machines with lots of memory

We have a DMA32 zone now, lets use it to make sure the card
can reach the memory we have allocated for the video frame
buffers.

Signed-off-by: Gerds Hoffmann <kraxel@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5033): MSI TV@nywhere Plus fixes
hermann pitton [Mon, 8 Jan 2007 00:19:48 +0000]
V4L/DVB (5033): MSI TV@nywhere Plus fixes

- MSI TV@nywhere Plus. Fix radio, S-Video and external analog audio in
  as far we can know currently.

Signed-off-by: Hermann Pitton <hermann-pitton@arcor.de>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5029): Ks0127 status flags
Martin Samuelsson [Sun, 7 Jan 2007 23:50:27 +0000]
V4L/DVB (5029): Ks0127 status flags

Or status flags together in DECODER_GET_STATUS instead of and-zapping them.

Signed-off-by: Martin Samuelsson <sam@home.se>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5024): Fix quickcam communicator driver for big endian architectures
Grant Likely [Sun, 7 Jan 2007 13:33:30 +0000]
V4L/DVB (5024): Fix quickcam communicator driver for big endian architectures

Host endianess does not affect the order that pixel rgb data comes
in from the quickcam (the values are bytes, not words or longs).  The
driver is erroniously swapping the order of rgb values for big endian
machines.  This patch is needed get the Quickcam communicator working
on big endian machines (tested on powerpc)

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5021): Cx88xx: Fix lockup on suspend
Robert Hancock [Sun, 7 Jan 2007 11:12:19 +0000]
V4L/DVB (5021): Cx88xx: Fix lockup on suspend

Suspending with the cx88xx module loaded causes the system to lock up
because the cx88_audio_thread kthread was missing a try_to_freeze()
call, which caused it to go into a tight loop and result in softlockup
when suspending. Fix that.

Signed-off-by: Robert Hancock <hancockr@shaw.ca>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5020): Fix: disable interrupts while at KM_BOUNCE_READ
Mauro Carvalho Chehab [Sun, 14 Jan 2007 12:14:17 +0000]
V4L/DVB (5020): Fix: disable interrupts while at KM_BOUNCE_READ

vivi.c uses the KM_BOUNCE_READ with local interrupts enabled.
This means that if a disk interrupt occurs while vivi.c is using this
fixmap slot, the vivi.c driver will, upon return from that interrupt, find
that the fixmap slot now points at a different physical page.
The net result will probably be rare corruption of disk file contents,
because viv.c will now be altering the page which the disk code was
recently using.

Thanks to Andrew Morton for pointing this.
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agoV4L/DVB (5019): Fix the frame->grabstate update in read() entry point.
Thierry MERLE [Fri, 29 Dec 2006 23:32:08 +0000]
V4L/DVB (5019): Fix the frame->grabstate update in read() entry point.

The Coverity checker spotted that in usbvision_v4l2_read(), the variable
"frmx" is never assigned any value different from -1, but it's used an
an array index in "usbvision->frame[frmx]".
Thanks to Adrian Bunk <bunk@stusta.de> for warning about that.

Signed-off-by: Thierry MERLE <thierry.merle@free.fr>
Signed-off-by: Mauro Carvalho Chehab <mchehab@infradead.org>

12 years agommc: Correct definition of R6
Philip Langdale [Thu, 4 Jan 2007 15:04:47 +0000]
mmc: Correct definition of R6

During development of SDHC support, it was discovered that the definition
for R6 was incorrect. This patch fixes that and patches the drivers that
do switch on the response type.

Signed-off-by: Philip Langdale <philipl@overt.org>
Cc: Alex Dubov <oakad@yahoo.com>
Cc: Pavel Pisa <ppisa@pikron.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>

12 years agoomap: Update MMC response types
Carlos Eduardo Aguiar [Mon, 15 Jan 2007 05:38:15 +0000]
omap: Update MMC response types

This patch is a fix in order to update MMC response types. This modification is
needed to allow SD card support on OMAP platforms.

Signed-off-by: Carlos Eduardo Aguiar <carlos.aguiar@indt.org.br>
Signed-off-by: Yuha Yrjola <juha.yrjola@solidboot.com>
Signed-off-by: Pierre Ossman <drzeus@drzeus.cx>

12 years ago[JFFS2] debug.h: include <linux/sched.h> for current->pid
David Woodhouse [Sat, 13 Jan 2007 00:19:03 +0000]
[JFFS2] debug.h: include <linux/sched.h> for current->pid

Signed-off-by: David Woodhouse <dwmw2@infradead.org>