GitHub/LineageOS/android_kernel_samsung_universal7580.git
19 years ago[SCSI] Legacy MegaRAID: Fix READ CAPACITY
James Bottomley [Sat, 1 Oct 2005 14:38:05 +0000 (09:38 -0500)]
[SCSI] Legacy MegaRAID: Fix READ CAPACITY

Some Legacy megaraid cards can't actually cope with the scatter/gather
version of the READ CAPACITY command (which is what we now send them
since altering all SCSI internal I/O to go via the block layer).  Fix
this (and a few other broken megaraid driver assumptions) by sending
the non-sg version of the command if the sg list only has a single
element.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aic7xxx/aic79xx: fix module removal path not to panic
James Bottomley [Sun, 2 Oct 2005 20:22:35 +0000 (15:22 -0500)]
[SCSI] aic7xxx/aic79xx: fix module removal path not to panic

In these drivers, scsi_remove_host() is called too late, at the point
it is called, the driver has already shut down too far to accept any
I/O that the shutdown might generate.  Any generated I/O actually
triggers a panic.

Fix this by calling scsi_remove_host() as early as possible and not
calling scsi_host_put() until just before we kfree the ahc_softc.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] fix potential panic with proc on module removal
James Bottomley [Sun, 2 Oct 2005 17:59:49 +0000 (12:59 -0500)]
[SCSI] fix potential panic with proc on module removal

There's a problem in our host release in that it calls
scsi_proc_hostdir_rm(). However, if you hold a reference to the host as
you remove the module, the host template (which proc uses) will be freed
and the system will panic when the host device is finally released.

Fix this by moving scsi_proc_hostdir_rm() to where it should be: in
scsi_remove_host().

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: remove aac_insert_entry
Mark Haverkamp [Tue, 20 Sep 2005 19:57:11 +0000 (12:57 -0700)]
[SCSI] aacraid: remove aac_insert_entry

Received from Mark Salyzyn from Adaptec.

High Priority Queues have *never* been used in the entire history of the
aac based adapters. Associated with this, aac_insert_entry can be
removed, SavedIrql can be removed & padding variable can be removed.
With the movement of SavedIrql out & replaced with an automatic variable
qflags, the locking can be refined somewhat. The sparse warnings did not
catch the need for byte swapping in the 'dprintk' debugging print
macros, so fixed this up when this code was moved outside of the now
refined locking.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: fib size math fix
Mark Haverkamp [Tue, 20 Sep 2005 19:57:04 +0000 (12:57 -0700)]
[SCSI] aacraid: fib size math fix

Received from Mark Salyzyn from Adaptec.

The size of the command packet's scatter gather list maximum size was
miscalculated in the low range leading to the driver initialization
limiting the maximum i/o size that could go to the Adapter. There were
no negative operational side effects resulting from this bad math, only
a subtle limit in performance of the Adapter at the top end of the
range.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: initialization timeout
Mark Haverkamp [Tue, 20 Sep 2005 19:56:50 +0000 (12:56 -0700)]
[SCSI] aacraid: initialization timeout

Received from Mark Salyzyn from Adaptec.

In the rare instances where the adapter, or the motherboard, is
misbehaving; driver initialization or shutdown becomes problematic. By
introducing a 3 minute timeout on the first interrupt driven command
during initialization, or the issuance of the adapter shutdown command
during driver unload, we can resolve the lockup problems induced by
common (but rare) hardware misbehaviors.

The timeout during initialization, should it occur, is accompanied by a
message presented to the console and the logs indicating that the user
should inspect and resolve problems with interrupt routing.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: error return checking
Mark Haverkamp [Tue, 20 Sep 2005 19:56:36 +0000 (12:56 -0700)]
[SCSI] aacraid: error return checking

This patch adds some additional error return checking and error return
value propagation during initialization. Also, the deprecation of
pci_module_init with pci_register_driver along with the change in return
values.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: handle AIF hotplug events (update)
Mark Haverkamp [Mon, 26 Sep 2005 20:04:56 +0000 (13:04 -0700)]
[SCSI] aacraid: handle AIF hotplug events (update)

Received from Mark Salyzyn from Adaptec.

Hotplug sniffs the AIFs (events) from the adapter and if a container
change resulting in the device going offline (container zero), online
(container zero completed) or changing capacity (morph) it will take
actions by calling the appropriate API.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: aacraid: AIF preallocation (update)
Mark Haverkamp [Mon, 26 Sep 2005 20:02:15 +0000 (13:02 -0700)]
[SCSI] aacraid: aacraid: AIF preallocation (update)

Recevied from Mark Salyzyn from Adaptec.

Aif pre-allocation is used to pull the kmalloc outside of the locks.

Applies to the scsi-misc-2.6 git tree.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] aacraid: Greater than 2TB capacity support
Mark Haverkamp [Thu, 22 Sep 2005 16:15:24 +0000 (09:15 -0700)]
[SCSI] aacraid: Greater than 2TB capacity support

Received from Mark Salyzyn from Adaptec.

There are a few adapters that are capable of creating devices with this large
of a capacity, but now that we have the large fib support in, the management
applications will be capable of generating them.  The problem is, once they are
created, the driver will not be able to access the devices correctly without
this patch.

Signed-off-by: Mark Haverkamp <markh@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] MegaRAID SAS RAID: new driver
Bagalkote, Sreenivas [Tue, 20 Sep 2005 21:46:58 +0000 (17:46 -0400)]
[SCSI] MegaRAID SAS RAID: new driver

Signed-off-by: Sreenivas Bagalkote <Sreenivas.Bagalkote@lsil.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] sas: fix remote phy removal
Christoph Hellwig [Sun, 25 Sep 2005 21:10:33 +0000 (23:10 +0200)]
[SCSI] sas: fix remote phy removal

Brown paperbag bug:  sas_rphy_delete was ordered completely
wrong.  Fix it up to be the same order as sas_phy_delete or
fc_rport_terminate and fix rphy objects that leaked after module
removal.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] qla2xxx: fix remote port timeout with qla2xxx driver
Andrew Vasquez [Tue, 20 Sep 2005 20:32:11 +0000 (13:32 -0700)]
[SCSI] qla2xxx: fix remote port timeout with qla2xxx driver

This patch fixes a hole in the rport unblock handling when processing
fabric events via the ADISC/PLOGI device state machine.  Original code
would not properly 'unblock' the port upon the port reloging into the
fabric.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] allow REPORT LUN scanning even for LUN 0 PQ of 3
James Bottomley [Fri, 23 Sep 2005 01:33:28 +0000 (20:33 -0500)]
[SCSI] allow REPORT LUN scanning even for LUN 0 PQ of 3

Currently we just ignore the device, which means there are a few
arrays out there that we don't find.

This patch updates the scsi_report_lun_scan() to take a target instead
of a device so it can be called on a return of
SCSI_SCAN_TARGET_PRESENT, which is what a PQ 3 device returns.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] 3ware 9000: Add support for 9550SX controllers
adam radford [Thu, 22 Sep 2005 00:20:14 +0000 (17:20 -0700)]
[SCSI] 3ware 9000: Add support for 9550SX controllers

Signed-off-by: Adam Radford <linuxraid@amcc.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] lpfc build fix
Andrew Morton [Wed, 21 Sep 2005 16:46:54 +0000 (09:46 -0700)]
[SCSI] lpfc build fix

gcc-2.95.x doesn't do anonymous unions.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Wed, 21 Sep 2005 20:21:35 +0000 (13:21 -0700)]
Merge /pub/scm/linux/kernel/git/roland/infiniband

19 years ago[IB] Fix RMPP receive length calculation
Hal Rosenstock [Wed, 21 Sep 2005 19:33:09 +0000 (12:33 -0700)]
[IB] Fix RMPP receive length calculation

Based on simplification idea from Sean Hefty <sean.hefty@intel.com>

Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[IB] Add MAD data field size definitions
Sean Hefty [Wed, 21 Sep 2005 19:31:26 +0000 (12:31 -0700)]
[IB] Add MAD data field size definitions

Clean up code by using enums instead of hard-coded magic numbers.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] fat: fix adate
Stephane Kardas [Wed, 21 Sep 2005 16:55:45 +0000 (09:55 -0700)]
[PATCH] fat: fix adate

During a forensic analysis on the fat file system, I found than the result for
the last access date on this file system was different between the stat
command and the istat command (package tct-utils).

The istat command display a true date (the right windows date) but the stat
primitive (so stat, find, ls command) displays a wrong date.

Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] update URL for HPET spec.
Randy Dunlap [Wed, 21 Sep 2005 16:55:44 +0000 (09:55 -0700)]
[PATCH] update URL for HPET spec.

Correct URL for HPET spec.

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Adds sys_set_mempolicy() in include/linux/syscalls.h
Eric Dumazet [Wed, 21 Sep 2005 16:55:43 +0000 (09:55 -0700)]
[PATCH] Adds sys_set_mempolicy() in include/linux/syscalls.h

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Cc: Andi Kleen <ak@suse.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] skge: add maintainer
Stephen Hemminger [Wed, 21 Sep 2005 16:55:41 +0000 (09:55 -0700)]
[PATCH] skge: add maintainer

Add MAINTAINER record for Andrew ;-)

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Cc: Jeff Garzik <jgarzik@pobox.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] fixup Documentation/DocBook/kernel-hacking.tmpl
Christoph Hellwig [Wed, 21 Sep 2005 16:55:40 +0000 (09:55 -0700)]
[PATCH] fixup Documentation/DocBook/kernel-hacking.tmpl

__FUNCTION__ is the prefered kernel idiom, __func__ is not supported by gcc
2.95 (we actually map __FUNCTION__ to __func__ for more recent compilers,
but it should never be used directly)

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: add a note about partially hardcoded VM_* flags
Paolo 'Blaisorblade' Giarrusso [Wed, 21 Sep 2005 16:55:39 +0000 (09:55 -0700)]
[PATCH] mm: add a note about partially hardcoded VM_* flags

Hugh made me note this line for permission checking in mprotect():

if ((newflags & ~(newflags >> 4)) & 0xf) {

after figuring out what's that about, I decided it's nasty enough.  Btw
Hugh itself didn't like the 0xf.

We can safely change it to VM_READ|VM_WRITE|VM_EXEC because we never change
VM_SHARED, so no need to check that.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] mm: update stale comment for removal of page->list
Paolo 'Blaisorblade' Giarrusso [Wed, 21 Sep 2005 16:55:38 +0000 (09:55 -0700)]
[PATCH] mm: update stale comment for removal of page->list

Update comment for the 2.6.6-rc1 conversion from page->list and
address_space->{clean,dirty,locked}_pages to radix tree tagging and ->lru.

I've mostly avoided to mention page lists (at least I've shortened the
comment).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] README update from the stone age
Paolo 'Blaisorblade' Giarrusso [Wed, 21 Sep 2005 16:55:37 +0000 (09:55 -0700)]
[PATCH] README update from the stone age

We have no options which the user can set in the Makefile.  Only the
EXTRAVERSION, which is also useful in place of the "backup modules"
suggestion.

We don't have configuration options in the top Makefile.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] fix locking comment in unmap_region()
Paolo 'Blaisorblade' Giarrusso [Wed, 21 Sep 2005 16:55:37 +0000 (09:55 -0700)]
[PATCH] fix locking comment in unmap_region()

That comment is plain wrong (we even take the pagetable lock inside
unmap_region()).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Fix I2O config-osm init to return proper error
Deepak Saxena [Wed, 21 Sep 2005 16:55:35 +0000 (09:55 -0700)]
[PATCH] Fix I2O config-osm init to return proper error

We currently unregister the config-osm driver if initialization of the
legacy ioctl() handlers failed but still return success.  We should be
returning -EBUSY in this case.

Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Cc: Markus Lidel <Markus.Lidel@shadowconnect.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] ppc64: Build zImage.vmode for G5
Benjamin Herrenschmidt [Wed, 21 Sep 2005 16:55:34 +0000 (09:55 -0700)]
[PATCH] ppc64: Build zImage.vmode for G5

zImage.vmode was recently added.  It's a version of zImage in which the ELF
note section used by open firmware indicates that it requires a virtual
mode instance of OF instead of real mode.  This allows it to work with
Apple OF, and thus is directly bootable (or netbootable) from OF command
line.  (Unfortunately, pSeries OF sort-of requires real mode and Apple OF
sort-of requires virtual mode, and both tend to be unhappy if no notes
section specifies the mode at all).

However, we forgot to add zImage.vmode to the default G5 build. This
fixes it.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] remove blkdev_scsi_issue_flush_fn again
Christoph Hellwig [Wed, 21 Sep 2005 16:55:33 +0000 (09:55 -0700)]
[PATCH] remove blkdev_scsi_issue_flush_fn again

This function was removed a while ago, but crept in again via a recent
scsi merge.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] ppc64: Fix issue with non zero boot cpu
Anton Blanchard [Wed, 21 Sep 2005 16:55:33 +0000 (09:55 -0700)]
[PATCH] ppc64: Fix issue with non zero boot cpu

The new version of the flattened device tree passes the boot cpuid in the
header instead of via a linux,boot-cpu property.

We need to update the in kernel OF parsing code to do this, otherwise
machines with a non zero boot cpuid fail to come up.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] cciss: busy_initializing bug fix
Mike Miller [Wed, 21 Sep 2005 16:55:32 +0000 (09:55 -0700)]
[PATCH] cciss: busy_initializing bug fix

This patch fixes the problem Bjorn reported.  The busy_initializing flag
should have cleared before going into the for loop.

Signed-off-by: Mike Miller <mike.miller@hp.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] PPC64: Fix boot for some pre-POWER4 systems
Olof Johansson [Wed, 21 Sep 2005 16:55:31 +0000 (09:55 -0700)]
[PATCH] PPC64: Fix boot for some pre-POWER4 systems

Some RS64 systems (such as F80) have non-python host bridges with EADS.
However, they have two EADS with 4 buses each under them, so the old logic
that assumed no more than 7 busses per PHB failed miserably.

Big thanks to Olaf Hering for helping me test this, he's got one of the few
machines that broke from the previous logic.

Also, to be a bit smarter at detecting the need for a PHB-level IOMMU table
by checking for the presence of an ISA bus.  Only PHBs with ISA bridges
should need the PHB-level table.

Signed-off-by: Olof Johansson <olof@lixom.net>
Cc: Anton Blanchard <anton@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Add printk_clock()
Andrew Morton [Wed, 21 Sep 2005 16:55:30 +0000 (09:55 -0700)]
[PATCH] Add printk_clock()

ia64's sched_clock() accesses per-cpu data which isn't set up at boot time.
Hence ia64 cannot use printk timestamping, because printk() will crash in
sched_clock().

So make printk() use printk_clock(), defaulting to sched_clock(), overrideable
by the architecture via attribute(weak).

Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] Fix invisible threads problem
Sripathi Kodi [Mon, 19 Sep 2005 23:26:12 +0000 (18:26 -0500)]
[PATCH] Fix invisible threads problem

When the main thread of a thread group has done pthread_exit() and died,
the other threads are still happily running, but will not be visible
under /proc because their leader is no longer accessible.

This fixes the access control so that we can see the sub-threads again.

Signed-off-by: Sripathi Kodi <sripathik@in.ibm.com>
Acked-by: Al Viro <viro@ftp.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] ppc64: Fix PCI flags when using OF device tree
Paul Mackerras [Wed, 21 Sep 2005 04:14:22 +0000 (14:14 +1000)]
[PATCH] ppc64: Fix PCI flags when using OF device tree

My code to set up the PCI tree from the Open Firmware device tree was
setting IORESOURCE_* flags on the resources for the devices, but not
the PCI_BASE_ADDRESS_* flags.  This meant that some drivers
misbehaved, and /proc/pci showed the wrong types for the resources.
This fixes it.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Wed, 21 Sep 2005 14:53:38 +0000 (07:53 -0700)]
Merge /pub/scm/linux/kernel/git/davem/sparc-2.6

19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Wed, 21 Sep 2005 14:53:05 +0000 (07:53 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

19 years ago[SPARC] cs4231: Fix SBUS support in this driver.
Christopher Zimmermann [Wed, 21 Sep 2005 07:41:22 +0000 (00:41 -0700)]
[SPARC] cs4231: Fix SBUS support in this driver.

From: Christopher Zimmermann <madroach@zakweb.de>

This patch enables SBus support for the cs4231 sound driver.
It is tested on an Ultra2. Capture and playback both work.
I experienced lags and crashes using certain threaded
players like ogg123 and mp3blaster, while the former is
lagging far more. This behavior may be specific to SMP
systems. It is reproducable using the dummy sound card
driver. Sox works flawlessly.

Setting up the calculation of ptr in snd_cs4231_playback_pointer
was a bit strange. I got it to work by not incrementing the
[pc]_periods_sent counter when starting DMA the first time
in cs4231_dma_trigger. Therefore this dummy thing.

[ I did some minor cleanups -DaveM ]

Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[ATYFB]: Fix build with CONFIG_FB_ATY_GENERIC_LCD disabled.
Tom 'spot' Callaway [Wed, 21 Sep 2005 07:29:16 +0000 (00:29 -0700)]
[ATYFB]: Fix build with CONFIG_FB_ATY_GENERIC_LCD disabled.

Signed-off-by: Tom 'spot' Callaway <tcallawa@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[TCP]: Set default congestion control correctly for incoming connections.
Stephen Hemminger [Wed, 21 Sep 2005 07:19:46 +0000 (00:19 -0700)]
[TCP]: Set default congestion control correctly for incoming connections.

Patch from Joel Sing to fix the default congestion control algorithm
for incoming connections. If a new congestion control handler is added
(via module), it should become the default for new
connections. Instead, the incoming connections use reno. The cause is
incorrect initialisation causes the tcp_init_congestion_control()
function to return after the initial if test fails.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Acked-by: Ian McDonald <imcdnzl@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[FIB_TRIE]: message cleanup
Stephen Hemminger [Wed, 21 Sep 2005 07:15:39 +0000 (00:15 -0700)]
[FIB_TRIE]: message cleanup

Cleanup the printk's in fib_trie:
* Convert a couple of places in the dump code to BUG_ON
* Put log level's on each message
The version message really needed the message since it leaks out
on the pretty Fedora bootup.

Signed-off-by: Stephen Hemminger <shemminger@osdl.org>
Acked-by: Robert Olsson <Robert.Olsson@data.slu.se>,
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[AF_PACKET]: Allow for > 8 byte hardware addresses.
Eric W. Biederman [Wed, 21 Sep 2005 07:11:37 +0000 (00:11 -0700)]
[AF_PACKET]: Allow for > 8 byte hardware addresses.

The convention is that longer addresses will simply extend
the hardeware address byte arrays at the end of sockaddr_ll and
packet_mreq.

In making this change a small information leak was also closed.
The code only initializes the hardware address bytes that are
used, but all of struct sockaddr_ll was copied to userspace.
Now we just copy sockaddr_ll to the last byte of the hardware
address used.

For error checking larger structures than our internal
maximums continue to be allowed but an error is signaled if we can
not fit the hardware address into our internal structure.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/roland/infiniband
Linus Torvalds [Wed, 21 Sep 2005 00:38:51 +0000 (17:38 -0700)]
Merge /pub/scm/linux/kernel/git/roland/infiniband

19 years ago[PATCH] fbdev: Fix reversed back and front porches
Antonino A. Daplas [Tue, 20 Sep 2005 23:32:49 +0000 (07:32 +0800)]
[PATCH] fbdev: Fix reversed back and front porches

In fbdev perspective, the frontporch is the lower/right margin and the
backporch is the upper/left margin.

Correct.

Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] nvidiafb: Fix absence of cursor in nvidiafb
Antonino A. Daplas [Tue, 20 Sep 2005 23:30:21 +0000 (07:30 +0800)]
[PATCH] nvidiafb: Fix absence of cursor in nvidiafb

A recent change in nvidiafb caused nvidiafb_cursor to always return -ENXIO
instead of using the soft_cursor.  This will happen if the parameter "hwcur"
is not set, which happens to be the default.

Signed-off-by: Antonino Daplas <adaplas@pol.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[SPARC64]: Verify vmalloc TLB misses more strictly.
David S. Miller [Tue, 20 Sep 2005 19:18:38 +0000 (12:18 -0700)]
[SPARC64]: Verify vmalloc TLB misses more strictly.

Arrange the modules, OBP, and vmalloc areas such that a range
verification can be done quite minimally.

Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[PATCH] IB/mthca: Fix device removal memory leak
Michael S. Tsirkin [Tue, 20 Sep 2005 17:54:48 +0000 (10:54 -0700)]
[PATCH] IB/mthca: Fix device removal memory leak

Clean up QP table array on device removal.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] IPoIB: Don't flush workqueue from within workqueue
Roland Dreier [Tue, 20 Sep 2005 17:52:04 +0000 (10:52 -0700)]
[PATCH] IPoIB: Don't flush workqueue from within workqueue

ipoib_mcast_restart_task() is always called from within the
single-threaded IPoIB workqueue, so flushing the workqueue from within
the function can lead to a recursion overflow.  But since we're
running in a single-threaded workqueue, we're already synchronized
against other items in the workqueue, so just get rid of the flush in
ipoib_mcast_restart_task().

Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/aia21/ntfs-2.6
Linus Torvalds [Tue, 20 Sep 2005 15:54:33 +0000 (08:54 -0700)]
Merge /pub/scm/linux/kernel/git/aia21/ntfs-2.6

19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-for-linus-2.6
Linus Torvalds [Tue, 20 Sep 2005 15:50:49 +0000 (08:50 -0700)]
Merge /linux/kernel/git/jejb/scsi-for-linus-2.6

19 years agoNTFS: More runlist handling fixes from Richard Russon and myself.
Anton Altaparmakov [Tue, 20 Sep 2005 13:19:30 +0000 (14:19 +0100)]
NTFS: More runlist handling fixes from Richard Russon and myself.

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>
19 years ago[SPARC64]: Move DCACHE_ALIASING_POSSIBLE define to asm/page.h
David S. Miller [Tue, 20 Sep 2005 03:11:57 +0000 (20:11 -0700)]
[SPARC64]: Move DCACHE_ALIASING_POSSIBLE define to asm/page.h

This showed that arch/sparc64/kernel/ptrace.c was not getting
the define properly, and thus the code protected by this ifdef
was never actually compiled before.  So fix that too.

Signed-off-by: David S. Miller <davem@davemloft.net>
19 years agoLinux v2.6.14-rc2
Linus Torvalds [Tue, 20 Sep 2005 03:00:41 +0000 (20:00 -0700)]
Linux v2.6.14-rc2

Avast, ye scurvy land-lubbers! Time to try out a new release.

Arrr!

19 years ago[BYTEORDER]: Document alignment and byteorder macros
Ed L. Cashin [Tue, 20 Sep 2005 02:57:36 +0000 (19:57 -0700)]
[BYTEORDER]: Document alignment and byteorder macros

This patch comments the fact that although passing le64_to_cpup et
al. is within the intended use of the byteorder macros, using
get_unaligned is the recommended way to go.

Signed-off-by: Ed L. Cashin <ecashin@coraid.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[SPARC64]: Handle little-endian unaligned loads/stores correctly.
David S. Miller [Tue, 20 Sep 2005 02:56:06 +0000 (19:56 -0700)]
[SPARC64]: Handle little-endian unaligned loads/stores correctly.

Because we use byte loads/stores to cons up the value
in and out of registers, we can't expect the ASI endianness
setting to take care of this for us.  So do it by hand.

This case is triggered by drivers/block/aoe/aoecmd.c in the
ataid_complete() function where it goes:

/* word 100: number lba48 sectors */
ssize = le64_to_cpup((__le64 *) &id[100<<1]);

This &id[100<<1] address is 4 byte, rather than 8 byte aligned,
thus triggering the unaligned exception.

Signed-off-by: David S. Miller <davem@davemloft.net>
19 years agoMake fsnotify possibly work better for the inode removal case
Linus Torvalds [Tue, 20 Sep 2005 02:54:29 +0000 (19:54 -0700)]
Make fsnotify possibly work better for the inode removal case

Checking i_nlink is dubious, but the alternatives look even
less appetizing.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Tue, 20 Sep 2005 01:46:11 +0000 (18:46 -0700)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

19 years ago[PATCH] raw_sendmsg DoS on 2.6
Mark J Cox [Tue, 20 Sep 2005 00:55:30 +0000 (17:55 -0700)]
[PATCH] raw_sendmsg DoS on 2.6

Fix unchecked __get_user that could be tricked into generating a
memory read on an arbitrary address.  The result of the read is not
returned directly but you may be able to divine some information about
it, or use the read to cause a crash on some architectures by reading
hardware state.  CAN-2004-2492.

Fix from Al Viro, ack from Dave Miller.

Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[TCP]: Handle SACK'd packets properly in tcp_fragment().
Herbert Xu [Tue, 20 Sep 2005 01:18:38 +0000 (18:18 -0700)]
[TCP]: Handle SACK'd packets properly in tcp_fragment().

The problem is that we're now calling tcp_fragment() in a context
where the packets might be marked as SACKED_ACKED or SACKED_RETRANS.
This was not possible before as you never retransmitted packets that
are so marked.

Because of this, we need to adjust sacked_out and retrans_out in
tcp_fragment().  This is exactly what the following patch does.

We also need to preserve the SACKED_ACKED/SACKED_RETRANS marking
if they exist.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[8021Q]: Add endian annotations.
Alexey Dobriyan [Mon, 19 Sep 2005 22:41:28 +0000 (15:41 -0700)]
[8021Q]: Add endian annotations.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[WAN] hdlc_cisco: Fix regression introduced by skb->tail changes.
Krzysztof Halasa [Mon, 19 Sep 2005 22:39:32 +0000 (15:39 -0700)]
[WAN] hdlc_cisco: Fix regression introduced by skb->tail changes.

The following commit breaks cisco mode with my WAN drivers:
author David S. Miller <davem@davemloft.net>
Tue, 28 Jun 2005 22:25:31 +0000 (15:25 -0700)
commit 689be43945e9ca7dd704522e55af1b8a73a994d3

"[NET]: Remove gratuitous use of skb->tail in network drivers."

The following patch fixes it - please apply (cisco_hard_header does
skb_push(4 bytes)).

Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[NETFILTER]: Export ip_nat_port_{nfattr_to_range,range_to_nfattr}
Harald Welte [Mon, 19 Sep 2005 22:35:57 +0000 (15:35 -0700)]
[NETFILTER]: Export ip_nat_port_{nfattr_to_range,range_to_nfattr}

Those exports are needed by the PPTP helper following in the next
couple of changes.

Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[NETFILTER]: Rename misnamed function
Patrick McHardy [Mon, 19 Sep 2005 22:35:31 +0000 (15:35 -0700)]
[NETFILTER]: Rename misnamed function

Both __ip_conntrack_expect_find and ip_conntrack_expect_find_get take
a reference to the expectation, the difference is that callers of
__ip_conntrack_expect_find must hold ip_conntrack_lock.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[NETFILTER] ip6tables: remove duplicate code
Yasuyuki Kozakai [Mon, 19 Sep 2005 22:34:40 +0000 (15:34 -0700)]
[NETFILTER] ip6tables: remove duplicate code

Some IPv6 matches have very similar loops to find IPv6 extension header
and we can unify them. This patch introduces ipv6_find_hdr() to do it.
I just checked that it can find the target headers in the packet which has
dst,hbh,rt,frag,ah,esp headers.

Signed-off-by: Yasuyuki Kozakai <yasuyuki.kozakai@toshiba.co.jp>
Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[NETFILTER]: Add new PPTP conntrack and NAT helper
Harald Welte [Mon, 19 Sep 2005 22:33:08 +0000 (15:33 -0700)]
[NETFILTER]: Add new PPTP conntrack and NAT helper

This new "version 3" PPTP conntrack/nat helper is finally ready for
mainline inclusion.  Special thanks to lots of last-minute bugfixing
by Patric McHardy.

Signed-off-by: Harald Welte <laforge@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[IPV4]: fib_trie RCU refinements
Robert Olsson [Mon, 19 Sep 2005 22:31:18 +0000 (15:31 -0700)]
[IPV4]: fib_trie RCU refinements

* This patch is from Paul McKenney's RCU reviewing.

Signed-off-by: Robert Olsson <robert.olsson@its.uu.se>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[IPV4]: fib_trie tnode stats refinements
Robert Olsson [Mon, 19 Sep 2005 22:29:52 +0000 (15:29 -0700)]
[IPV4]: fib_trie tnode stats refinements

* Prints the route tnode and set the stats level deepth as before.

Signed-off-by: Robert Olsson <robert.olsson@its.uu.se>
Signed-off-by: David S. Miller <davem@davemloft.net>
19 years ago[PATCH] IB: Fix data length for RMPP SA sends
Hal Rosenstock [Mon, 19 Sep 2005 20:51:01 +0000 (13:51 -0700)]
[PATCH] IB: Fix data length for RMPP SA sends

We need to subtract off the header length from our payload
length when sending multi-packet SA messages.

Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[SCSI] zfcp: add additional fc_host attributes
Andreas Herrmann [Mon, 19 Sep 2005 14:56:17 +0000 (16:56 +0200)]
[SCSI] zfcp: add additional fc_host attributes

this patch adds some fc host attributes and removes its equivalents
from the zfcp_adapter structure and zfcp specific sysfs subtree.

Furthermore it removes superfluous calls to fc_remort_port_delete when
an adapter is set offline because rports will be removed by
fc_remove_host anyway.

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] zfcp: provide support for NPIV
Maxim Shchetynin [Tue, 13 Sep 2005 19:51:16 +0000 (21:51 +0200)]
[SCSI] zfcp: provide support for NPIV

N_Port ID Virtualization (NPIV) allows a single FCP port to appear as
multiple, distinct ports providing separate port identification. NPIV
is supported by FC HBAs on System z9. zfcp was adapted to support this
new feature.

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] zfcp: enhancement of zfcp debug features
Maxim Shchetynin [Tue, 13 Sep 2005 19:50:38 +0000 (21:50 +0200)]
[SCSI] zfcp: enhancement of zfcp debug features

Debug features (DBFs) els_dbf, cmd_dbf and abt_dbf were removed and
san_dbf, hba_dbf and scsi_dbf were introduced. The erp_dbf did not
change.
The new traces improve debugging of problems with zfcp, scsi-stack,
multipath and hardware in the SAN. san_dbf traces things like ELS and
CT commands, hba_dbf saves HBA specific information of requests, and
scsi_dbf saves FCP and SCSI specific information of requests. Common
to all new DBFs is that they provide a so called structured view. This
significantly improves readability of the traces.

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] zfcp: shorten eh_bus_reset and eh_host_reset handlers
Andreas Herrmann [Tue, 13 Sep 2005 19:49:52 +0000 (21:49 +0200)]
[SCSI] zfcp: shorten eh_bus_reset and eh_host_reset handlers

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] zfcp: remove function zfcp_fsf_req_wait_and_cleanup
Andreas Herrmann [Tue, 13 Sep 2005 19:48:33 +0000 (21:48 +0200)]
[SCSI] zfcp: remove function zfcp_fsf_req_wait_and_cleanup

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] zfcp: remove union zfcp_req_data, use unit refcount for FCP commands
Andreas Herrmann [Tue, 13 Sep 2005 19:47:52 +0000 (21:47 +0200)]
[SCSI] zfcp: remove union zfcp_req_data, use unit refcount for FCP commands

o union zfcp_req_data removed
o increment unit refcount when processing FCP commands
 (This fixes a theoretical race: When all scsi commands of a unit
  are aborted and the scsi_device is removed then the unit could be
  removed before all fsf_requests of that unit are completely processed.)

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] zfcp: fix race conditions when accessing erp_action lists
Andreas Herrmann [Tue, 13 Sep 2005 19:47:11 +0000 (21:47 +0200)]
[SCSI] zfcp: fix race conditions when accessing erp_action lists

o always use locking when changing erp_action lists,
o avoid escalation to ERP_ACTION_REOPEN_PORT_FORCED if erp_action is
  still in use for ERP_ACTION_REOPEN_PORT

Signed-off-by: Andreas Herrmann <aherrman@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] fusion SAS support (mptsas driver) minor fix
Moore, Eric Dean [Thu, 15 Sep 2005 19:17:14 +0000 (13:17 -0600)]
[SCSI] fusion SAS support (mptsas driver) minor fix

On Thursday, September 15, 2005 6:22 AM, Christoph Hellwig wrote:
> Looks good to me, except for the spurious scsi_print_command prototype
> in mptscsih.h.

The attached patch addresses that concern.

Signed-off-by: Eric Moore <Eric.Moore@lsil.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] fusion SAS support (mptsas driver) updates
Moore, Eric Dean [Thu, 15 Sep 2005 00:09:10 +0000 (18:09 -0600)]
[SCSI] fusion SAS support (mptsas driver) updates

Summary of Changes:
* splitting mpt_interrupt per Christophs suggestion
about a month ago
* rename ScsiCfgData to SpiCfgData structure,
then move all the raid related info into
new structure called RaidCfgData.  This is
done because SAS supports RAID, as well as SPI,
so the raid stuff should be seperate.
* incorrect timeout calculation for cntdn
inside WaitForDoorbellAck and WaitForDoortbellInt
* add support for interpreting SAS Log Info
* Increase Event Log Size from 0xA to 0x32
* Fix bug in mptsas/mptfc/mptspi - when controller
has Initiator Mode Disabled, and only running in
TargetMode, the mptctl would panic when loading.
The fix is to return 0, instead of -ENODEV, in
SCSI LLD respective probe routines
* Fix bug in mptlan.c - driver will panic if
there is host reset, due to dev being set to
zero in mpt_lan_ioc_reset
* Fix's for SPI - Echo Buffer
* Several fix's in mptscsih_io_done - FCP Response
info, RESIDUAL_MISMATCH, Data Underrun, etc.
* Cleanup Error Handling - EH handlers,
mptscsih_flush_cmds, and zeroing out ScsiLookup
from mptscsih_qcmd
* Cleanup asyn event handling from
mptscsih -> mptscsih_event_process.  Also
added support for SAS Persistent Table Full,
an asyn event

Signed-off-by: Eric Moore <Eric.Moore@lsil.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] fusion SAS support (mptsas driver)
Christoph Hellwig [Fri, 9 Sep 2005 14:27:19 +0000 (16:27 +0200)]
[SCSI] fusion SAS support (mptsas driver)

Adds the actual mptsas driver, based upon the LSI driver with new work
for SAS transport class integration from Eric Moore and me.

This obviously depends on the SAS transport class.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] fusion core changes for SAS support
Christoph Hellwig [Fri, 9 Sep 2005 14:25:54 +0000 (16:25 +0200)]
[SCSI] fusion core changes for SAS support

 - various bits for SAS support from the LSI driver.
 - use the device private data for the fusion target private data.
   this should be using the midlayer target data framework, but we
   can't move over to that until fusion has been switched to the
   generic DV code
 - use target ID and channel from the fusion target private data,
   because those in scsi_device will be different for mptsas

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[PATCH] Alpha: ISA IRQs fixup for dp264
Ivan Kokshaysky [Mon, 19 Sep 2005 14:55:51 +0000 (18:55 +0400)]
[PATCH] Alpha: ISA IRQs fixup for dp264

Basically, this extends original dp264 fixup to all dp264 variations.
Here is one minor change: mask out bits 4-7 of a value assigned by SRM,
because
- newer consoles report ISA IRQs with offset 0xe0;
- even if console IRQ value is bogus, we'll have a value < 16
  so it should be harmless as it won't clash with native IRQs.

Particularly this fixes USB interrupt problem on xp1000 and es40.

Signed-off-by: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Acked-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
19 years ago[PATCH] IB/mthca: Don't try to set srq->last for userspace SRQs
Roland Dreier [Mon, 19 Sep 2005 16:17:56 +0000 (09:17 -0700)]
[PATCH] IB/mthca: Don't try to set srq->last for userspace SRQs

Userspace SRQs don't have a buffer allocated for them in the kernel, so
it doesn't make sense to set srq->last during initialization.  In fact,
this can crash trying to follow a nonexistent buffer pointer.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[SCSI] Fix thread termination for the SCSI error handle
James Bottomley [Mon, 19 Sep 2005 14:50:04 +0000 (09:50 -0500)]
[SCSI] Fix thread termination for the SCSI error handle

From:  Alan Stern <stern@rowland.harvard.edu>

This patch (as561) fixes the error handler's thread-exit code.  The
kthread_stop call won't wake the thread from a down_interruptible, so
the patch gets rid of the semaphore and simply does

        set_current_state(TASK_INTERRUPTIBLE);

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Modified to simplify the termination loop and correct the sleep condition.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[SCSI] fix oops on usb storage device disconnect
James Bottomley [Sun, 18 Sep 2005 20:05:20 +0000 (15:05 -0500)]
[SCSI] fix oops on usb storage device disconnect

We fix the oops by enforcing the host state model.  There have also
been two extra states added: SHOST_CANCEL_RECOVERY and
SHOST_DEL_RECOVERY so we can take the model through host removal while
the recovery thread is active.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years agoMerge branch 'master' of /home/src/linux-2.6/
Anton Altaparmakov [Mon, 19 Sep 2005 08:47:49 +0000 (09:47 +0100)]
Merge branch 'master' of /home/src/linux-2.6/

19 years agoNTFS: Fix ntfs_{read,write}page() to cope with concurrent truncates better.
Anton Altaparmakov [Mon, 19 Sep 2005 08:41:39 +0000 (09:41 +0100)]
NTFS: Fix ntfs_{read,write}page() to cope with concurrent truncates better.

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>
19 years agoNTFS: Fix handling of compressed directories that I broke in earlier changeset.
Anton Altaparmakov [Mon, 19 Sep 2005 08:38:41 +0000 (09:38 +0100)]
NTFS: Fix handling of compressed directories that I broke in earlier changeset.

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>
19 years agoNTFS: Fix various bugs in the runlist merging code. (Based on libntfs
Anton Altaparmakov [Mon, 19 Sep 2005 08:33:40 +0000 (09:33 +0100)]
NTFS: Fix various bugs in the runlist merging code.  (Based on libntfs
      changes by Richard Russon.)

Signed-off-by: Anton Altaparmakov <aia21@cantab.net>
19 years ago[PATCH] IB/mthca: Fix posting work requests to shared receive queues
Roland Dreier [Sun, 18 Sep 2005 21:00:17 +0000 (14:00 -0700)]
[PATCH] IB/mthca: Fix posting work requests to shared receive queues

The error handling paths in mthca_tavor_post_srq_recv() and
mthca_arbel_post_srq_recv() are quite bogus, the result of a
screwed up merge.  Fix them so they work as intended.

Pointed out by Michael S. Tsirkin <mst@mellanox.co.il>

Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] IB/mthca: Initialize eq->nent before we use it
Roland Dreier [Sun, 18 Sep 2005 20:52:06 +0000 (13:52 -0700)]
[PATCH] IB/mthca: Initialize eq->nent before we use it

In mthca_create_eq(), we call get_eqe() before setting eq->nent.  This
is wrong, because get_eqe() uses eq->nent.  Fix this, and clean up the
code a little while we're at it.  (We got lucky with the current code,
because eq->nent was cleared to 0, which get_eqe() made happen to do
the right thing)

Pointed out by Michael S. Tsirkin <mst@mellanox.co.il>

Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] IPoIB: Fix SA client retransmission strategy
Hal Rosenstock [Sun, 18 Sep 2005 20:47:53 +0000 (13:47 -0700)]
[PATCH] IPoIB: Fix SA client retransmission strategy

We got a little mixed up with what the backoff member holds in the
IPoIB multicast group structure: sometimes it was used as a number of
seconds, and sometimes it was used as a number of jiffies.  Fix the
code so that backoff is always in seconds.

Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] IB/mthca: fix posting of first work request
Roland Dreier [Tue, 13 Sep 2005 17:41:03 +0000 (10:41 -0700)]
[PATCH] IB/mthca: fix posting of first work request

Fix posting first WQE for mem-free HCAs: we need to link to previous
WQE even in that case.  While we're at it, simplify code for
Tavor-mode HCAs.  We don't really need the conditional test there
either; we can similarly always link to the previous WQE.

Based on Michael S. Tsirkin's analogous fix for userspace libmthca.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] IB/mthca: assign ACK timeout field correctly
Roland Dreier [Mon, 12 Sep 2005 21:08:51 +0000 (14:08 -0700)]
[PATCH] IB/mthca: assign ACK timeout field correctly

The hardware reads the ACK timeout field from the most significant 5
bits of struct mthca_qp_path's ackto field, not the least significant
bits.  This fix has the driver put the timeout in the right place.
Without this, we get a timeout that is 2^8 times too small.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years ago[PATCH] IPoIB: fix module removal race
Michael S. Tsirkin [Mon, 12 Sep 2005 16:52:28 +0000 (09:52 -0700)]
[PATCH] IPoIB: fix module removal race

Since ipoib uses queue_delayed_work to run flush task on port state events,
it must flush scheduled work after unregistering the event handler.

Signed-off-by: Michael S. Tsirkin <mst@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
19 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Sun, 18 Sep 2005 22:31:20 +0000 (15:31 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

19 years ago[ARM] Fix warning in arch/arm/mach-sa1100/generic.c
Russell King [Sun, 18 Sep 2005 21:16:41 +0000 (22:16 +0100)]
[ARM] Fix warning in arch/arm/mach-sa1100/generic.c

Fix:

arch/arm/mach-sa1100/generic.c:224: warning: 'struct mcp_plat_data' declared inside parameter list

caused by mussing structure and function declaration.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
19 years ago[ARM] Fix warning in arch/arm/kernel/semaphore.c
Russell King [Sun, 18 Sep 2005 20:49:29 +0000 (21:49 +0100)]
[ARM] Fix warning in arch/arm/kernel/semaphore.c

Newer binutils complains:
/tmp/cc07pbI9.s:146: Warning: ignoring changed section type for .sched.text

Fix this warning by adding %progbits to the .section.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
19 years ago[SCSI] SCSI scanning and removal fixes
Alan Stern [Tue, 26 Jul 2005 14:27:10 +0000 (10:27 -0400)]
[SCSI] SCSI scanning and removal fixes

This patch (as545) fixes the list traversals in __scsi_remove_target and
scsi_forget_host.  In each case the existing code list_for_each_entry_safe
in an _unsafe_ manner, because the list was not protected from outside
modification while the iteration was running.

The new scsi_forget_host routine takes the moderately controversial step
of iterating over devices for removal rather than iterating over targets.
This makes more sense to me because the current scheme treats targets as
second-class citizens, created and removed on demand, rather than as
objects corresponding to actual hardware.  (Also I couldn't figure out any
safe way to iterate over the target list, since it's not so easy to tell
when a target has already been removed.)

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
19 years ago[ARM] 2917/1: Make IXP4xx mach_desc's depend on config options
Deepak Saxena [Sun, 18 Sep 2005 20:11:56 +0000 (21:11 +0100)]
[ARM] 2917/1: Make IXP4xx mach_desc's depend on config options

Patch from Deepak Saxena

Building a kernel for IXDP425 currently includes the machine descriptors
for IXDP465 and PRPMC1100 even if those machines are not configured.
This means we can build a kernel that boots on those machines even
though the machine_is_xxx() macro will always return 0 and other bits
such as PCI won't be compiled in. This can lead to many wasted hours
wondering what you have done to your kernel to make it randomly crash
thus requireing large quantities of beer to be consumed. While I am
all for consumption of large quantities of beer, there are better
reasons to do so then stupid kernel bugs.

Signed-off-by: Deepak Saxena <dsaxena@plexity.net>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>