]> git.neil.brown.name Git - history.git/log
history.git
21 years ago[PCMCIA] add EnE specific initialization to fix HDSP
Daniel Ritz [Tue, 4 May 2004 22:31:38 +0000 (23:31 +0100)]
[PCMCIA] add EnE specific initialization to fix HDSP

Patch from Daniel Ritz.

This patch clears an almost undocumented EnE specific test register
that makes sound on RME Hammerfall DSP Carbus work...should even work
after suspend.

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/net-2.6
David S. Miller [Sun, 2 May 2004 15:34:46 +0000 (08:34 -0700)]
Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6
into kernel.bkbits.net:/home/davem/net-2.6

21 years ago[NET]: Fix MODULE_PARM_DESC typo in dummy driver.
Rusty Russell [Sun, 2 May 2004 15:31:59 +0000 (08:31 -0700)]
[NET]: Fix MODULE_PARM_DESC typo in dummy driver.

21 years agoMerge nuts.davemloft.net:/disk1/BK/sparcwork-2.6
David S. Miller [Sun, 2 May 2004 11:24:01 +0000 (04:24 -0700)]
Merge nuts.davemloft.net:/disk1/BK/sparcwork-2.6
into nuts.davemloft.net:/disk1/BK/sparc-2.6

21 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-serial
Linus Torvalds [Sun, 2 May 2004 07:36:29 +0000 (00:36 -0700)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-serial
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years ago[SERIAL] Fix the calculation of the number of UARTs
Russell King [Sun, 2 May 2004 21:47:52 +0000 (22:47 +0100)]
[SERIAL] Fix the calculation of the number of UARTs

The calculation ended up believing we had one less UART than we
really had.  Fix it.

21 years agoBe more careful about semaphore contention memory ordering.
Linus Torvalds [Sun, 2 May 2004 06:06:12 +0000 (23:06 -0700)]
Be more careful about semaphore contention memory ordering.

Don't touch the wakee stack after marking it runnable.

21 years agoMake types of big integers in bitops.h explicit.
Linus Torvalds [Sun, 2 May 2004 06:04:40 +0000 (23:04 -0700)]
Make types of big integers in bitops.h explicit.

"sparse" warns about implicit type conversions that may cause
surprising results. Did you know that large decimal types have
different type conversions from large hexadecimals?

21 years ago[PATCH] FAT: remove symbols exports from msdosfs/vfat
Hirofumi Ogawa [Sun, 2 May 2004 04:26:25 +0000 (21:26 -0700)]
[PATCH] FAT: remove symbols exports from msdosfs/vfat

From Christoph Hellwig <hch@lst.de>

If we're ever going to ressurect umsdos it should be a stackable
filesystem..

21 years ago[PATCH] FAT: small cleanup
Hirofumi Ogawa [Sun, 2 May 2004 04:26:13 +0000 (21:26 -0700)]
[PATCH] FAT: small cleanup

21 years ago[PATCH] FAT: simple error handling cleanup
Hirofumi Ogawa [Sun, 2 May 2004 04:26:03 +0000 (21:26 -0700)]
[PATCH] FAT: simple error handling cleanup

From: RenĂ© Scharfe <l.s.r@web.de>

the following patch converts the error handling paths in VFAT fs to use
goto, making it more consistent with other filesystem code. Shrinks the
resulting binary by 144 bytes in my build.

21 years ago[PATCH] FAT: Fix nfsv2 support
Hirofumi Ogawa [Sun, 2 May 2004 04:25:52 +0000 (21:25 -0700)]
[PATCH] FAT: Fix nfsv2 support

The ->dentry_to_fh() can use the 20 bytes in the case of NFSv2, but
fat_dentry_to_fh() requires 24 bytes by my patch.

So nfsd reply the EOPNOTSUPP to nfs client, then nfs client convert
the unknown error to -EIO.

This patch fixes the problem by pushing the handle data into 20 bytes.

21 years agoMerge bk://cifs.bkbits.net/linux-2.5cifs
Linus Torvalds [Sun, 2 May 2004 04:16:13 +0000 (21:16 -0700)]
Merge bk://cifs.bkbits.net/linux-2.5cifs
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years agodo not refresh mode (e.g. in revalidate) to windows servers
Steve French [Sun, 2 May 2004 06:13:48 +0000 (01:13 -0500)]
do not refresh mode (e.g. in revalidate) to windows servers

21 years agoreduce excessive stack space usage in smb password hashing
Steve French [Sun, 2 May 2004 03:53:52 +0000 (22:53 -0500)]
reduce excessive stack space usage in smb password hashing

21 years agoeven if O_CREAT specified do not reset mode when file not actually created
Steve French [Sat, 1 May 2004 17:12:12 +0000 (12:12 -0500)]
even if O_CREAT specified do not reset mode when file not actually created

21 years agofix ppc64 build problem due to missing header
Steve French [Sat, 1 May 2004 22:33:19 +0000 (17:33 -0500)]
fix ppc64 build problem due to missing header

21 years ago[PATCH] mcdx.c insanity removal
Alexander Viro [Sat, 1 May 2004 12:53:15 +0000 (05:53 -0700)]
[PATCH] mcdx.c insanity removal

The mcdx.c author had pulled off something absolutely amazing - he had
declared several unsigned variables (ISA port numbers) as void *, using
explicit cast to unsigned in almost all places that used them.
Exception: printk.  There he proudly used them as pointers - with %3p in
format.  That cute trick allowed him to avoid using %03x, which
apparently scared him for some reason.

Switched to use of unsigned, killed casts, replaced %3p with %03x in
formats.  BTW, the code had been that way since the initial merge back
in 1.3.7...

21 years ago[PATCH] ppc64: fix incorrect signal handler argument
Paul Mackerras [Sat, 1 May 2004 12:10:20 +0000 (05:10 -0700)]
[PATCH] ppc64: fix incorrect signal handler argument

This fixes a bug in the ppc64 signal delivery code where the signal
number argument to a signal handler can get corrupted before the handler
is called.  The specific scenario is that a process is in a blocking
system call when two signals get generated for it, both of which have
handlers.

The signal code will stack up two signal frames on the process stack
(assuming the mask for the first signal delivered doesn't block the
second signal) and return to userspace to run the handler for the second
signal.  On return from that handler the first handler gets run with an
incorrect signal number argument because we end up with regs->result
still having a negative value (left over from when the system call was
interrupted) when it should be zero.  This patch sets it to zero when we
set up the signal frame (in three places; for 64-bit processes, and for
32-bit processes for RT and non-RT signals).

The way we handle signal delivery and signal handler return using the
regs->result field in ppc64 is more complicated than it needs to be.  In
ppc32 I have already simplified it and eliminated use of the
regs->result field.  I am going to do the same in the ppc64 code, but I
think this patch should go in for now to fix the bug.

The patch also fixes a couple of places where we were unnecessarily and
incorrectly truncating the regs->result value to 32 bits
(sys32_sigreturn and sys32_rt_sigreturn return a long value, as all
syscalls do, and if regs->result is negative we need those syscalls to
return a negative value).

Thanks to Maneesh Soni for identifying the specific circumstances
under which this bug shows up.

21 years agoMerge bk://kernel.bkbits.net/davem/net-2.6
Linus Torvalds [Sat, 1 May 2004 12:01:49 +0000 (05:01 -0700)]
Merge bk://kernel.bkbits.net/davem/net-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/net-2.6
David S. Miller [Sat, 1 May 2004 10:42:39 +0000 (03:42 -0700)]
Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6
into kernel.bkbits.net:/home/davem/net-2.6

21 years ago[IPV4/IPV6]: Fix listing of listening sockets.
Herbert Xu [Sat, 1 May 2004 10:39:13 +0000 (03:39 -0700)]
[IPV4/IPV6]: Fix listing of listening sockets.

There is a bug in listening_get_first() which used by /proc/net/tcp*
where it wasn't looping through all the sockets in each hash chain.
This problem doesn't show up unless the first socket in a chain doesn't
match the family that is being looked up.

The following patch fixes this by getting rid of listening_get_first()
altogether.

21 years ago[PATCH] add missing #include
Denis Vlasenko [Sat, 1 May 2004 10:32:45 +0000 (03:32 -0700)]
[PATCH] add missing #include

There's a subtle problem with "inline" usage in <linux/string.h>:

 <linux/string.h>:
        this pulls in __constant_c_and_count_memset()

 <linux/mm.h>:
        this pulls <compiler.h>, re-defining
        inline == __inline__ __attribute__((always_inline)).

 But by now it is too late! The compiler has already seen the bare
 "inline" in string.h, and hasn't inlined it.

Result:

# grep __constant System.map
c0144670 t __constant_c_and_count_memset
c0145c60 t __constant_c_and_count_memset
... many more copies of this function ...

Fixed by including <compiler.h> early enough.

21 years agoMerge bk://linux.bkbits.net/linux-2.5
Steve French [Sat, 1 May 2004 08:01:22 +0000 (01:01 -0700)]
Merge bk://linux.bkbits.net/linux-2.5
into hostme.bitkeeper.com:/repos/c/cifs/linux-2.5cifs

21 years agoMerge bk://are.twiddle.net/axp-2.6/
Linus Torvalds [Sat, 1 May 2004 05:51:38 +0000 (22:51 -0700)]
Merge bk://are.twiddle.net/axp-2.6/
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years ago[ALPHA] Add message queue syscalls.
Richard Henderson [Sat, 1 May 2004 04:02:13 +0000 (21:02 -0700)]
[ALPHA] Add message queue syscalls.

21 years agoMerge nuts.davemloft.net:/disk1/BK/network-2.6
David S. Miller [Sat, 1 May 2004 02:23:45 +0000 (19:23 -0700)]
Merge nuts.davemloft.net:/disk1/BK/network-2.6
into nuts.davemloft.net:/disk1/BK/net-2.6

21 years ago[PATCH] serverworks.c: fix DMA for OSB4
Bartlomiej Zolnierkiewicz [Sat, 1 May 2004 02:11:44 +0000 (19:11 -0700)]
[PATCH] serverworks.c: fix DMA for OSB4

From: Patrick Wildi <patrick@wildi.com>

On OSB4 the hwif->ultra_mask is set to not support UDMA.
Unfortunately in that case svwks_config_drive_xfer_rate()
falls through to the end of the function, instead of trying
other DMA modes.

21 years agoFix fixed fadvice length handling
Linus Torvalds [Fri, 30 Apr 2004 14:28:54 +0000 (07:28 -0700)]
Fix fixed fadvice length handling

 - Correctly handle wraparound on offset+len
 - fix FADV_WILLNEED handling of non-page-aligned (offset+len)

Let's hope we don't need to fix the fixed fix.

21 years ago[PATCH] usb linkage fix
Andrew Morton [Fri, 30 Apr 2004 14:09:43 +0000 (07:09 -0700)]
[PATCH] usb linkage fix

On sparc64 toolchain:

drivers/built-in.o(.init.text+0xaf8c): In function `usb_init':
: undefined reference to `usbfs_cleanup'

usb_init() is __init and usbfs_cleanup() is __exit.  No can do.

21 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-rmk
Linus Torvalds [Fri, 30 Apr 2004 08:46:47 +0000 (01:46 -0700)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-rmk
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years ago[ARM PATCH] 1841/1: Lubbock defconfig update
Nicolas Pitre [Fri, 30 Apr 2004 21:59:26 +0000 (22:59 +0100)]
[ARM PATCH] 1841/1: Lubbock defconfig update

Patch from Nicolas Pitre

21 years ago[ARM PATCH] 1840/1: recognize more XScale CPU variants
Nicolas Pitre [Fri, 30 Apr 2004 21:55:03 +0000 (22:55 +0100)]
[ARM PATCH] 1840/1: recognize more XScale CPU variants

Patch from Nicolas Pitre

... including the new PXA270 aka Bulverde.

21 years ago[ARM PATCH] 1839/1: fix lubbock_flash.c which used a bogus reg name
Nicolas Pitre [Fri, 30 Apr 2004 21:51:18 +0000 (22:51 +0100)]
[ARM PATCH] 1839/1: fix lubbock_flash.c which used a bogus reg name

Patch from Nicolas Pitre

Before previous patch this driver compiled OK but was buggy.
Now it doesn't compile anymore as the bogus macro has been
deleted.  Fix that in any case.

The same fix has been committed to the MTD CVS already, but please forward
this to Linus otherwise Lubbock won't compile from kernel.org tree anymore
(waiting for dwmw2 to update this might prove ... hrm ... long)

21 years ago[ARM PATCH] 1838/1: Lubbock leds and macro namespace cleanup
Nicolas Pitre [Fri, 30 Apr 2004 21:47:10 +0000 (22:47 +0100)]
[ARM PATCH] 1838/1: Lubbock leds and macro namespace cleanup

Patch from Nicolas Pitre

Too many macro with too generic names.  Let's remove unneeded code and
redundant/unused macros.  This also prevent namespace clash with upcoming
patches.

21 years ago[ARM PATCH] 1837/1: small Lubbock cleanup
Nicolas Pitre [Fri, 30 Apr 2004 21:43:05 +0000 (22:43 +0100)]
[ARM PATCH] 1837/1: small Lubbock cleanup

Patch from Nicolas Pitre

Minor cleanup of Lubbock specific code, like removal of
redundant mappings.
Also a prerequisite for some upcoming patches.

21 years ago[ARM PATCH] 1836/1: don't hardcode virtual addresses
Nicolas Pitre [Fri, 30 Apr 2004 21:39:17 +0000 (22:39 +0100)]
[ARM PATCH] 1836/1: don't hardcode virtual addresses

Patch from Nicolas Pitre

virtual address mapping can change.

21 years ago[PATCH] task_struct alignment fix
Andrew Morton [Fri, 30 Apr 2004 07:51:54 +0000 (00:51 -0700)]
[PATCH] task_struct alignment fix

The recent slab alignment changes broke an unknown number of architectures
(parisc and x86_64 for sure) by causing task_structs to be insufficiently
aligned.

We need good alignemnt because architectures do things like dumping FP state
into the task_struct with instructions which require particular alignment (I
think).

So change the default alignment to L1_CACHE_BYTES, which is what we used to
have, via SLAB_HW_CACHE_ALIGN.

21 years ago[PATCH] Update MTD concatenating driver
Russell King [Fri, 30 Apr 2004 05:37:45 +0000 (22:37 -0700)]
[PATCH] Update MTD concatenating driver

This patch updates the MTD concatenating driver from MTD CVS, which
fixes issues found with this driver which concatenates multiple MTD
devices into one MTD device.

From David Woodhouse, through CVS:

revision 1.8
date: 2003/06/30 11:01:26;  author: dwmw2;  state: Exp;  lines: +5 -5
I will not commit stuff whilst pissed
I will not commit stuff whilst pissed

revision 1.7
date: 2003/06/29 21:26:34;  author: dwmw2;  state: Exp;  lines: +9 -9
Fix ecc/oob subdev comparisions

revision 1.6
date: 2003/06/25 12:37:50;  author: dwmw2;  state: Exp;  lines: +14 -6
Don't pretend to have {read,write}_{oob,ecc} functions if subdevices don't

revision 1.5
date: 2003/06/25 12:21:16;  author: dwmw2;  state: Exp;  lines: +390 -397
coding style cleanup

21 years agoMerge http://lia64.bkbits.net/to-linus-2.5
Linus Torvalds [Fri, 30 Apr 2004 05:11:01 +0000 (22:11 -0700)]
Merge http://lia64.bkbits.net/to-linus-2.5
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years agoCset exclude: davidm@tiger.hpl.hp.com|ChangeSet|20040427053149|28511
David Mosberger [Fri, 30 Apr 2004 03:26:18 +0000 (20:26 -0700)]
Cset exclude: davidm@tiger.hpl.hp.com|ChangeSet|20040427053149|28511

21 years ago[PATCH] fadvise length handling fix
Andrew Morton [Fri, 30 Apr 2004 00:42:49 +0000 (17:42 -0700)]
[PATCH] fadvise length handling fix

POSIX sez: "If len is zero, all data following offset is specified."

21 years ago[PATCH] ppc64: shmget() translation bugfix
Andrew Morton [Fri, 30 Apr 2004 00:42:38 +0000 (17:42 -0700)]
[PATCH] ppc64: shmget() translation bugfix

From: David Gibson <david@gibson.dropbear.id.au>

The 32->64 bit syscall translation layer on ppc64 incorrectly sign-
extends rather than zero-extending the second parameter to shmget(),
which should be a size_t.  This means that it is impossible to shmget()
more 2GB or more from a 32-bit process.

21 years ago[PATCH] make ikconfig quiet
Andrew Morton [Fri, 30 Apr 2004 00:42:26 +0000 (17:42 -0700)]
[PATCH] make ikconfig quiet

From: "Randy.Dunlap" <rddunlap@osdl.org>

From: Pavel Machek <pavel@ucw.cz>

Kill uninformative boot-time message.

21 years ago[PATCH] s390: oprofile Kconfig fixes
Andrew Morton [Fri, 30 Apr 2004 00:42:14 +0000 (17:42 -0700)]
[PATCH] s390: oprofile Kconfig fixes

From: Arnd Bergmann <arnd@arndb.de>

Enable basic profiling code on s390 depending on CONFIG_PROFILING, not
CONFIG_OPROFILE.

CONFIG_PROFILING should enable the generic profiling code here, even if
CONFIG_OPROFILE is not set.  Note that the identical code on i386 is always
compiled in, regardless of CONFIG_PROFILING and CONFIG_OPROFILE.

21 years ago[PATCH] DVB:Fix adapter module removal bug
Andrew Morton [Fri, 30 Apr 2004 00:42:01 +0000 (17:42 -0700)]
[PATCH] DVB:Fix adapter module removal bug

From: Michael Hunold <hunold@convergence.de>

unfortunately it's possible to remove a DVB adapter module even if a DVB
network device has been set up using this adapter.

The attached patch fixes this problem.

21 years ago[PATCH] efivars sysfs fix
Andrew Morton [Fri, 30 Apr 2004 00:41:49 +0000 (17:41 -0700)]
[PATCH] efivars sysfs fix

From: Matt Tolentino <metolent@snoqualmie.dp.intel.com>

This trailing space in sysfs contents is not needed.

21 years ago[PATCH] nfs printk warning fix
Andrew Morton [Fri, 30 Apr 2004 00:41:36 +0000 (17:41 -0700)]
[PATCH] nfs printk warning fix

From: Trond Myklebust <trond.myklebust@fys.uio.no>

fs/nfs/direct.c: In function `nfs_direct_IO':
fs/nfs/direct.c:458: warning: int format, different type arg (arg 2)

21 years ago[PATCH] parport pnp detection fix
Andrew Morton [Thu, 29 Apr 2004 15:17:22 +0000 (08:17 -0700)]
[PATCH] parport pnp detection fix

From: Adam Belay <ambx1@neo.rr.com>

Fix http://bugme.osdl.org/show_bug.cgi?id=2540

It prevents PnP detection if devices were already detected by SuperIO.

21 years ago[PATCH] NFSv3: Fix SETATTR call after O_EXCL create
Trond Myklebust [Thu, 29 Apr 2004 15:17:01 +0000 (08:17 -0700)]
[PATCH] NFSv3: Fix SETATTR call after O_EXCL create

Ensure that when we send the SETATTR call after doing an O_EXCL create,
we always set the atime and ctime fields.

See RFC1813 for details on why the server is allowed to clobber these
two fields in order to cache a verifier that protects CREATE in case of
a timeout+resend from the client.

Patch is by Olaf Kirch.

21 years ago[NET]: More network layer static funcs and data.
Stephen Hemminger [Thu, 29 Apr 2004 09:26:49 +0000 (02:26 -0700)]
[NET]: More network layer static funcs and data.

21 years agoMerge http://lia64.bkbits.net/to-linus-2.5
Linus Torvalds [Thu, 29 Apr 2004 08:25:17 +0000 (01:25 -0700)]
Merge http://lia64.bkbits.net/to-linus-2.5
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years ago[PATCH] ia64: SN2 fix
Pat Gefre [Thu, 29 Apr 2004 08:15:15 +0000 (01:15 -0700)]
[PATCH] ia64: SN2 fix

Move HW interrupt register init. to the proper place
and don't force an interrupt if the IRQ is disabled
or in progress.

21 years agoia64: Fix Exec-Only stack patch so X can work again.
David Mosberger [Thu, 29 Apr 2004 08:14:05 +0000 (01:14 -0700)]
ia64: Fix Exec-Only stack patch so X can work again.

No pun intended, but X can't work if it can't turn on eXecute-permission
on the some data pages...

21 years ago[PATCH] mremap offset type
Hugh Dickins [Thu, 29 Apr 2004 07:13:28 +0000 (00:13 -0700)]
[PATCH] mremap offset type

Just found I never changed type of move_page_tables when I changed it to
return offset: einormous mremap moves would fail on 64-bit.

21 years agoMerge tiger.hpl.hp.com:/data1/bk/vanilla/linux-2.5
David Mosberger [Thu, 29 Apr 2004 06:57:12 +0000 (23:57 -0700)]
Merge tiger.hpl.hp.com:/data1/bk/vanilla/linux-2.5
into tiger.hpl.hp.com:/data1/bk/lia64/to-linus-2.5

21 years ago[PATCH] writeback livelock fix
Andrew Morton [Thu, 29 Apr 2004 00:21:10 +0000 (17:21 -0700)]
[PATCH] writeback livelock fix

To avoid various livelocks, the writeback code parks all the dirty inodes onto
sb->s_io and then works through that list until it is empty.  This assumes
that each inode will be moved to some other list as it is processed.

But there's a loophole: if the ->writepages() implementation does nothing at
all, the inode is not redirtied (which would move it to s_dirty).  This causes
s_io to not empty and pdflush goes nuts.

So when this happens, move the inode onto s_dirty within
__sync_single_inode().  Use list_move_tail() to attempt to preserve the
time-ordering of the s_dirty list.

21 years ago[PATCH] gcc-3.4.0 fixes
Andrew Morton [Thu, 29 Apr 2004 00:20:56 +0000 (17:20 -0700)]
[PATCH] gcc-3.4.0 fixes

From: Mikael Pettersson <mikpe@user.it.uu.se>

This patch fixes three warnings from gcc-3.4.0 in 2.6.6-rc3:

- arch/i386/pci/pcbios.c: use of "+m" constraint

- drivers/char/ftape/: use of cast-as-lvalue

- drivers/char/ftape/: __attribute__((packed)) on something containing only
  bytes

21 years ago[PATCH] Fix support for the Motorola PrPMC800
Andrew Morton [Thu, 29 Apr 2004 00:20:44 +0000 (17:20 -0700)]
[PATCH] Fix support for the Motorola PrPMC800

From: Tom Rini <trini@kernel.crashing.org>

Makes the Motorola PrPMC800 platform functional again.  This comes from Randy
Vinson <rvinson@mvista.com>.

21 years ago[PATCH] static functions in as-iosched.c
Andrew Morton [Thu, 29 Apr 2004 00:20:22 +0000 (17:20 -0700)]
[PATCH] static functions in as-iosched.c

From: Stephen Hemminger <shemminger@osdl.org>

21 years ago[PATCH] s390: oprofile for s390
Andrew Morton [Thu, 29 Apr 2004 00:20:11 +0000 (17:20 -0700)]
[PATCH] s390: oprofile for s390

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Add oprofile support for s/390.

21 years ago[PATCH] s390: zfcp host adapter
Andrew Morton [Thu, 29 Apr 2004 00:19:55 +0000 (17:19 -0700)]
[PATCH] s390: zfcp host adapter

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

zfcp host adapter change:
 - Fix addressing exception due to uninitialized host_scribble pointer.

21 years ago[PATCH] s390: 3270 console driver
Andrew Morton [Thu, 29 Apr 2004 00:19:43 +0000 (17:19 -0700)]
[PATCH] s390: 3270 console driver

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

3270 device driver change:
 - Add missing irb error checking.

21 years ago[PATCH] s390: network driver
Andrew Morton [Thu, 29 Apr 2004 00:19:30 +0000 (17:19 -0700)]
[PATCH] s390: network driver

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Network driver changes:
 - ctc: Add missing irb error checking.
 - iucv: Add name of net_device to iucvMagic to more than one
         connection between two guests.
 - qeth: Don't send IPA command if card is not in state SOFTSETUP or UP.
 - qeth: Fix number base in simple_strtoul call for buffer_count attribute.
 - qeth: Fix reallocating of buffers when buffer_count attribute is changed.
 - qeth: Correct handling of return codes in qeth_realloc_buffer_pool.
 - qeth: Don't call dev_close/dev_open on STOPLAN/STARTLAN commands.
         Use netif_carrier_off/netif_carrier_on instead.

21 years ago[PATCH] s390: common i/o layer
Andrew Morton [Thu, 29 Apr 2004 00:19:17 +0000 (17:19 -0700)]
[PATCH] s390: common i/o layer

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

Common i/o layer changes:
 - Don't use bus ids in crw debug feature.
 - Use cio_oper for oper notification to disconnected devices.
 - Remove __get_subchannel_by_stsch.
 - Make cio workqueue a single threaded workqueue.
 - Introduce addiotnal cio_notify workqueue for device driver notification.
 - Switch off path in vpm if cio_start returned -ENODEV.
 - Fix rescan for new subchannels after a logical vary on.

21 years ago[PATCH] s390: core s390
Andrew Morton [Thu, 29 Apr 2004 00:19:03 +0000 (17:19 -0700)]
[PATCH] s390: core s390

From: Martin Schwidefsky <schwidefsky@de.ibm.com>

s390 core changes:
 - Move setting/clearing of TIF_31BIT thread flag to SET_PERSONALITY.
 - Use TASK_UNMAPPED_BASE in elf_map32 for mmaps with address 0.
 - Define ARCH_KMALLOC_MINALIGN.
 - Define ARCH_MIN_TASKALIGN.

21 years ago[PATCH] cciss MAINTAINERS update
Andrew Morton [Thu, 29 Apr 2004 00:18:52 +0000 (17:18 -0700)]
[PATCH] cciss MAINTAINERS update

From: <mikem@beardog.cca.cpqcorp.net>

Here's an update for the MAINTAINERS file.

21 years ago[PATCH] cciss build fix
Andrew Morton [Thu, 29 Apr 2004 00:18:37 +0000 (17:18 -0700)]
[PATCH] cciss build fix

From: <mikem@beardog.cca.cpqcorp.net>

This patch fixes the linux/include/cciss_ioctl.h file.  When support for the
cciss big ioctl was added the stucture in the header was put in the wrong
place.  If an application includes the file it will fail to compile.

21 years ago[PATCH] Update kerneltraffic link in SubmittingDrivers and kernel-docs.txt
Andrew Morton [Thu, 29 Apr 2004 00:18:25 +0000 (17:18 -0700)]
[PATCH] Update kerneltraffic link in SubmittingDrivers and kernel-docs.txt

From: Coywolf Qi Hunt <coywolf@greatcn.org>

This updates the kerneltraffic url link found in
Documentation/SubmittingDrivers and Documentation/kernel-docs.txt.

21 years ago[PATCH] ppc32: compile error in signal.c
Andrew Morton [Thu, 29 Apr 2004 00:18:11 +0000 (17:18 -0700)]
[PATCH] ppc32: compile error in signal.c

From: Meelis Roos <mroos@linux.ee>

arch/ppc/kernel/signal.c: In function `handle_signal':
arch/ppc/kernel/signal.c:518: error: `newspp' undeclared (first use in this function)
arch/ppc/kernel/signal.c:518: error: (Each undeclared identifier is reported only once
arch/ppc/kernel/signal.c:518: error: for each function it appears in.)
arch/ppc/kernel/signal.c:518: warning: long unsigned int format, pointer arg (arg 3)

21 years ago[PATCH] cifssmb.c warning fix
Andrew Morton [Thu, 29 Apr 2004 00:17:59 +0000 (17:17 -0700)]
[PATCH] cifssmb.c warning fix

On ppc64, __u64 is `unsigned long', so:

fs/cifs/cifssmb.c: In function `CIFSSMBSetFileSize':
fs/cifs/cifssmb.c:2466: warning: long long int format, __u64 arg (arg 2)

21 years ago[PATCH] Fix might_sleep in /proc/swaps code
Alexander Viro [Thu, 29 Apr 2004 00:15:20 +0000 (17:15 -0700)]
[PATCH] Fix might_sleep in /proc/swaps code

This fixes a locking problem noted by Tim Hockin:
    * /proc/swaps uses seq_file code, calling seq_path() with swaplock held
    * seq_path() calls d_path()
    * d_path() calls mntput() which might_sleep()

We add a new semaphore protecting insertions/removals in the set of swap
components + switch of ->start()/->stop() to the same semaphore [fixes
deadlocks] + trivial cleanup of ->next().

21 years ago[PATCH] POWER5 erratum workaround
David Gibson [Thu, 29 Apr 2004 00:12:31 +0000 (17:12 -0700)]
[PATCH] POWER5 erratum workaround

Early POWER5 revisions (<DD2.1) have a problem requiring slbie
instructions to be repeated under some circumstances.  The patch below
adds a workaround (patch made by Anton Blanchard).

21 years ago[PATCH] Fix overeager stack-expansion on ppc64
David Gibson [Thu, 29 Apr 2004 00:11:59 +0000 (17:11 -0700)]
[PATCH] Fix overeager stack-expansion on ppc64

This fix is from Paul Mackerras and was applied in 2.4 sometime late
last year.

On ppc64, touching addresses between the highest other mapping and the
stack can cause the stack to be extended way, way down, rather than
causing a SEGV as you would expect.  This patch only allows the stack
mapping to be extended to cover addresses actually within the stack
(as determined by looking at the process's r1).  This fix is ported
from 2.4

This fixes failures on the LTP's shmdt01, munmap01 and munmap02 tests.

21 years ago[PATCH] fix default IDE interfaces initialization for PPC32
Bartlomiej Zolnierkiewicz [Thu, 29 Apr 2004 00:11:24 +0000 (17:11 -0700)]
[PATCH] fix default IDE interfaces initialization for PPC32

In ide_init_default_irq() patch I overlooked that ppc_ide_md.init_hwif_ports()
called from generic ide_init_hwif_ports() can set hwif->irq and it will
be overwritten by ide_init_default_irq() if CONFIG_PCI is defined.  Fix
it.

I will clean it up properly later after killing ide_init_hwif_ports() on
ARM{26}.  Doing it now is just wasted effort.

21 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-rmk
Linus Torvalds [Thu, 29 Apr 2004 00:10:01 +0000 (17:10 -0700)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-rmk
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years ago[ARM PATCH] 1835/1: Make ALTERA Excalibur work again in 2.6.5
Dirk Behme [Thu, 29 Apr 2004 15:08:42 +0000 (16:08 +0100)]
[ARM PATCH] 1835/1: Make ALTERA Excalibur work again in 2.6.5

Patch from Dirk Behme

Make ALTERA Excalibur work again in 2.6.4. Update serial driver uart00.c and defconfig.

This is an update of patch 1833/1.

Remove #include <config/pld/hotswap.h>, it isn't necessary at all.

21 years ago[ARM] Add Versatile default configuration
Russell King [Thu, 29 Apr 2004 13:50:28 +0000 (14:50 +0100)]
[ARM] Add Versatile default configuration

21 years ago[ARM] Move all page fault handling code to fault.c
Russell King [Thu, 29 Apr 2004 10:59:19 +0000 (11:59 +0100)]
[ARM] Move all page fault handling code to fault.c

21 years ago[ARM] Fix atomic bitops earlyclobber
Russell King [Wed, 28 Apr 2004 23:24:57 +0000 (00:24 +0100)]
[ARM] Fix atomic bitops earlyclobber

atomic_dec_and_test and atomic_add_negative didn't mark their
temporary variables as early-clobber.  Fix this.

21 years ago[ARM] Fix read_cpuid()
Russell King [Wed, 28 Apr 2004 21:32:24 +0000 (22:32 +0100)]
[ARM] Fix read_cpuid()

21 years ago[ARM] Fix shared mmap()ings for ARM VIPT caches.
Russell King [Wed, 28 Apr 2004 21:22:33 +0000 (22:22 +0100)]
[ARM] Fix shared mmap()ings for ARM VIPT caches.

This allows us to appropriately align shared mappings on VIPT caches
with aliasing issues.

21 years ago[ARM] Remove Anakin default configuration file.
Russell King [Wed, 28 Apr 2004 20:06:49 +0000 (21:06 +0100)]
[ARM] Remove Anakin default configuration file.

21 years ago[ARM] Fix monspecs in ARM-related framebuffer drivers.
Russell King [Wed, 28 Apr 2004 18:51:49 +0000 (19:51 +0100)]
[ARM] Fix monspecs in ARM-related framebuffer drivers.

Use named initialisers for monspecs; the format of the structure
changed a while back and it broke.

21 years ago[ARM] Oprofile should use asm/irq.h not asm/arch/irqs.h
Russell King [Wed, 28 Apr 2004 18:33:54 +0000 (19:33 +0100)]
[ARM] Oprofile should use asm/irq.h not asm/arch/irqs.h

21 years ago[ARM] Fix dependencies of SERIO_AMBAKMI and SERIO_RPCKBD
Russell King [Wed, 28 Apr 2004 18:24:24 +0000 (19:24 +0100)]
[ARM] Fix dependencies of SERIO_AMBAKMI and SERIO_RPCKBD

21 years ago[ARM] Update ioremap implementation.
Russell King [Wed, 28 Apr 2004 17:52:43 +0000 (18:52 +0100)]
[ARM] Update ioremap implementation.

Use flush_cache_vmap() after creating mappings.  Also use BUG_ON()
rather than if() BUG().

21 years ago[ARM] Update assabet_defconfig.
Russell King [Wed, 28 Apr 2004 17:16:11 +0000 (18:16 +0100)]
[ARM] Update assabet_defconfig.

21 years ago[ARM] Fix BE find_*_bit operations
Russell King [Wed, 28 Apr 2004 15:15:04 +0000 (16:15 +0100)]
[ARM] Fix BE find_*_bit operations

These broke when find_first_bit/find_next_bit was added.

21 years ago[ARM] Add read_cpuid() to aid reading CPU ID registers.
Russell King [Wed, 28 Apr 2004 15:02:29 +0000 (16:02 +0100)]
[ARM] Add read_cpuid() to aid reading CPU ID registers.

21 years ago[ARM PATCH] 1832/1: Typo in dma_unregister_dev printk
David Vrabel [Wed, 28 Apr 2004 13:58:49 +0000 (14:58 +0100)]
[ARM PATCH] 1832/1: Typo in dma_unregister_dev printk

Patch from David Vrabel

Fix a trivial typo in a dma_unregister_dev printk.

21 years agoMerge bk://kernel.bkbits.net/davem/net-2.6
Linus Torvalds [Wed, 28 Apr 2004 10:06:47 +0000 (03:06 -0700)]
Merge bk://kernel.bkbits.net/davem/net-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

21 years agoTell the sparse checker to use 64-bit mode when checking
Linus Torvalds [Wed, 28 Apr 2004 09:08:37 +0000 (02:08 -0700)]
Tell the sparse checker to use 64-bit mode when checking
a ppc64 tree.

21 years agoAdd __user annotations to ppc64 user access functions.
Linus Torvalds [Wed, 28 Apr 2004 09:06:55 +0000 (02:06 -0700)]
Add __user annotations to ppc64 user access functions.

21 years ago[NETLINK]: Mark some functions/data static.
Stephen Hemminger [Wed, 28 Apr 2004 09:05:29 +0000 (02:05 -0700)]
[NETLINK]: Mark some functions/data static.

21 years agoMerge davem@nuts.davemloft.net:/disk1/BK/net-2.6
David S. Miller [Wed, 28 Apr 2004 06:10:27 +0000 (23:10 -0700)]
Merge davem@nuts.davemloft.net:/disk1/BK/net-2.6
into kernel.bkbits.net:/home/davem/net-2.6

21 years ago[SPARC64]: Update defconfig.
David S. Miller [Wed, 28 Apr 2004 05:51:11 +0000 (22:51 -0700)]
[SPARC64]: Update defconfig.

21 years ago[CRYPTO]: Provide crc32c as a type of digest.
Clay Haapala [Wed, 28 Apr 2004 05:34:04 +0000 (22:34 -0700)]
[CRYPTO]: Provide crc32c as a type of digest.

21 years ago[LIB]: Use compiler.h's pure attribute macros in crc32.c.
Clay Haapala [Wed, 28 Apr 2004 05:30:17 +0000 (22:30 -0700)]
[LIB]: Use compiler.h's pure attribute macros in crc32.c.

21 years ago[LIB]: Add CRC32c (Castagnoli, et al Cyclic Redundancy-Check)
Clay Haapala [Wed, 28 Apr 2004 05:29:12 +0000 (22:29 -0700)]
[LIB]: Add CRC32c (Castagnoli, et al Cyclic Redundancy-Check)