opensuse:kernel.git
15 years ago[PATCH] i2c-proc wasn't checking kmalloc result
Dave Jones [Wed, 3 Apr 2002 03:56:58 +0000 (19:56 -0800)]
[PATCH] i2c-proc wasn't checking kmalloc result

15 years ago[PATCH] strtok->strsep in shmem
Dave Jones [Wed, 3 Apr 2002 03:56:53 +0000 (19:56 -0800)]
[PATCH] strtok->strsep in shmem

15 years ago[PATCH] strtok->strsep in autofs
Dave Jones [Wed, 3 Apr 2002 03:56:48 +0000 (19:56 -0800)]
[PATCH] strtok->strsep in autofs

15 years ago[PATCH] typo in pci_set_mwi header
Dave Jones [Wed, 3 Apr 2002 03:56:43 +0000 (19:56 -0800)]
[PATCH] typo in pci_set_mwi header

15 years ago[PATCH] strtok->strsep in hfs
Dave Jones [Wed, 3 Apr 2002 03:56:37 +0000 (19:56 -0800)]
[PATCH] strtok->strsep in hfs

15 years ago[PATCH] strtok->strsep in hpfs
Dave Jones [Wed, 3 Apr 2002 03:56:32 +0000 (19:56 -0800)]
[PATCH] strtok->strsep in hpfs

15 years ago[PATCH] remove bogus return from mtrr driver.
Dave Jones [Wed, 3 Apr 2002 03:56:27 +0000 (19:56 -0800)]
[PATCH] remove bogus return from mtrr driver.

It's amazing what you spot when you Lindent things.

15 years ago[PATCH] remove workaround for old binutils.
Dave Jones [Wed, 3 Apr 2002 03:56:22 +0000 (19:56 -0800)]
[PATCH] remove workaround for old binutils.

Modern binutils spits out a warning, fixing this broke older binutils.
I guess we can do this in 2.5 ?

15 years ago[PATCH] strtok -> strsep in adfs
Dave Jones [Wed, 3 Apr 2002 03:56:17 +0000 (19:56 -0800)]
[PATCH] strtok -> strsep in adfs

15 years ago[PATCH] nbd compile fix.
Dave Jones [Wed, 3 Apr 2002 03:56:12 +0000 (19:56 -0800)]
[PATCH] nbd compile fix.

15 years ago[PATCH] khttpd logs wrong debug message on leaving function.
Dave Jones [Wed, 3 Apr 2002 03:56:07 +0000 (19:56 -0800)]
[PATCH] khttpd logs wrong debug message on leaving function.

15 years ago[PATCH] extra sanity checks for mempool
Dave Jones [Wed, 3 Apr 2002 03:56:02 +0000 (19:56 -0800)]
[PATCH] extra sanity checks for mempool

15 years ago[PATCH] updated documentation for w9966 driver.
Dave Jones [Wed, 3 Apr 2002 03:24:31 +0000 (19:24 -0800)]
[PATCH] updated documentation for w9966 driver.

From 2.4

15 years ago[PATCH] watchdog nowayout for advantechwdt
Dave Jones [Wed, 3 Apr 2002 03:24:26 +0000 (19:24 -0800)]
[PATCH] watchdog nowayout for advantechwdt

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

15 years ago[PATCH] More verbosity in VIA tweak
Dave Jones [Wed, 3 Apr 2002 03:24:19 +0000 (19:24 -0800)]
[PATCH] More verbosity in VIA tweak

From 2.4

15 years ago[PATCH] Remove guess from bttv docs.
Dave Jones [Wed, 3 Apr 2002 03:24:14 +0000 (19:24 -0800)]
[PATCH] Remove guess from bttv docs.

15 years ago[PATCH] reiserfs tools update.
Dave Jones [Wed, 3 Apr 2002 03:24:08 +0000 (19:24 -0800)]
[PATCH] reiserfs tools update.

Recommended version bump.

15 years ago[PATCH] __init/__exit does nothing in prototypes
Dave Jones [Wed, 3 Apr 2002 03:24:03 +0000 (19:24 -0800)]
[PATCH] __init/__exit does nothing in prototypes

15 years ago[PATCH] remove dead comment
Dave Jones [Wed, 3 Apr 2002 03:23:57 +0000 (19:23 -0800)]
[PATCH] remove dead comment

Original from William Lee Irwin III <wli@holomorphy.com>

15 years ago[PATCH] more then enough typos.
Dave Jones [Wed, 3 Apr 2002 03:23:52 +0000 (19:23 -0800)]
[PATCH] more then enough typos.

s/more then/more than/

15 years ago[PATCH] watchdog nowayout for ib700wdt
Dave Jones [Wed, 3 Apr 2002 03:23:46 +0000 (19:23 -0800)]
[PATCH] watchdog nowayout for ib700wdt

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

15 years ago[PATCH] watchdog nowayout for eurotechwdt
Dave Jones [Wed, 3 Apr 2002 03:23:41 +0000 (19:23 -0800)]
[PATCH] watchdog nowayout for eurotechwdt

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

15 years ago[PATCH] printk levels for vme_scc driver
Dave Jones [Wed, 3 Apr 2002 03:23:34 +0000 (19:23 -0800)]
[PATCH] printk levels for vme_scc driver

15 years ago[PATCH] watchdog nowayout for acquirewdt
Dave Jones [Wed, 3 Apr 2002 03:23:28 +0000 (19:23 -0800)]
[PATCH] watchdog nowayout for acquirewdt

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

15 years ago[PATCH] videodev fixups / generic usercopy helper
Dave Jones [Wed, 3 Apr 2002 03:23:23 +0000 (19:23 -0800)]
[PATCH] videodev fixups / generic usercopy helper

Originally from Gerd...

I've just noticed that a hole left in the recent changes which should
allow the usb v4l drivers to unregister with open file handles.  The
drivers itself handle it just fine, but video_generic_ioctl() will barf
when called on unregistered devices.  Oops.

One way to fix this is to expect drivers call the helper function and
pass a pointer for the function doing the actual work, i.e. handle it
this way:

driver_ioctl(inode,file,cmd,userptr)
-> video_usercopy(inode,file,cmd,userptr,func)
copy_from_user(...)
-> func(inode,file,cmd,kernelptr);
copy_to_user(...)

Patch against 2.5.7-pre2 below. It updates videodev.[ch] and adapts
usbvideo.c to show how the driver changes will look like.

Note that this change makes the usercopy helper function a very generic
one, it probably could be used for other drivers to (as long as the API
has sane magic numbers based on _IO*(...) defines) as there is no
video4linux-related stuff in there any more.  So we might think of
renaming it an moving it to some more central place (fs/ioctl.c maybe).

15 years ago[PATCH] watchdog nowayout for i810-tco
Dave Jones [Wed, 3 Apr 2002 03:23:17 +0000 (19:23 -0800)]
[PATCH] watchdog nowayout for i810-tco

Originally from Matt Domsch.
Adds a nowayout module option to override CONFIG_WATCHDOG_NOWAYOUT
From 2.4

15 years ago[PATCH] Update bigphysarea URL
Dave Jones [Wed, 3 Apr 2002 03:23:12 +0000 (19:23 -0800)]
[PATCH] Update bigphysarea URL

15 years ago[PATCH] PCI IDS update.
Dave Jones [Wed, 3 Apr 2002 03:23:07 +0000 (19:23 -0800)]
[PATCH] PCI IDS update.

Mostly from 2.4. Adds some new ones, and removes some dupes.

15 years ago[PATCH] Update file list in INDEX
Dave Jones [Wed, 3 Apr 2002 03:23:01 +0000 (19:23 -0800)]
[PATCH] Update file list in INDEX

15 years ago[PATCH] unnecessary includes.
Dave Jones [Wed, 3 Apr 2002 03:22:56 +0000 (19:22 -0800)]
[PATCH] unnecessary includes.

A few follow ups to the cleanup done circa 2.5.4

15 years ago[PATCH] UDF write support problem in 2.5.7
Dave Jones [Wed, 3 Apr 2002 03:22:50 +0000 (19:22 -0800)]
[PATCH] UDF write support problem in 2.5.7

 From Peter Osterlund <petero2@telia.com>

I can't get UDF write support to work in kernel 2.5.7 or 2.5.7-pre2.
The problem is that linux/config.h is not included, so CONFIG_UDF_RW
is undefined and the driver is compiled without write support. This
patch fixes my problem:

15 years ago[PATCH] region handling cleanups for tpqic02
Dave Jones [Wed, 3 Apr 2002 03:22:45 +0000 (19:22 -0800)]
[PATCH] region handling cleanups for tpqic02

15 years ago[PATCH] Various typo fixes.
Dave Jones [Wed, 3 Apr 2002 03:22:39 +0000 (19:22 -0800)]
[PATCH] Various typo fixes.

Mostly harmless but 1-2 break compilation.

15 years ago[PATCH] CREDITS updates
Dave Jones [Wed, 3 Apr 2002 03:22:33 +0000 (19:22 -0800)]
[PATCH] CREDITS updates

15 years ago[PATCH] Gameport patch for drivers/sound/mad16.c
Dave Jones [Wed, 3 Apr 2002 03:22:28 +0000 (19:22 -0800)]
[PATCH] Gameport patch for drivers/sound/mad16.c

From: Michael Haardt <michael@moria.de> for 2.4, munged by me.

the MAD16 driver is able to enable/disable its gameport, but it does
not register it in the input subsystem.  The appended patch against
2.4.19-pre4 fixes that and also allows to enable to gameport via kernel
command line, not only when loaded as module.

15 years ago[PATCH] group #include's together in x86 ioremap.c
Dave Jones [Wed, 3 Apr 2002 03:22:23 +0000 (19:22 -0800)]
[PATCH] group #include's together in x86 ioremap.c

15 years ago[PATCH] sonypi driver update from 2.4
Dave Jones [Wed, 3 Apr 2002 03:22:16 +0000 (19:22 -0800)]
[PATCH] sonypi driver update from 2.4

15 years ago[PATCH] Simple boot flag specification support.
Dave Jones [Wed, 3 Apr 2002 03:22:09 +0000 (19:22 -0800)]
[PATCH] Simple boot flag specification support.

Originally by Alan. It was in the 2.4-ac tree for ages, much tweaked
by myself and Thomas Hood. Its main purpose is for skipping the full
memory test and other time-wasting diagnostics on reboot.

15 years ago[PATCH] silence DVD_INVALIDATE_AGID output.
Dave Jones [Wed, 3 Apr 2002 03:22:04 +0000 (19:22 -0800)]
[PATCH] silence DVD_INVALIDATE_AGID output.

Ancient patch from Jens to hush certain drives which generate a lot of
printk noise when playing DVDs.

15 years ago[PATCH] proc race on task_struct->sig
Dave Jones [Wed, 3 Apr 2002 03:21:59 +0000 (19:21 -0800)]
[PATCH] proc race on task_struct->sig

Originally from Chris Mason <mason@suse.com>..

collect_sigign_sigcatch can race against exit_sighand.  I haven't been
able to reproduce it, but I think it causes the oops reported in the
'Kernel Hangs 2.4.16 on heavy io Oracle Tivolie TSM' thread.

15 years ago[PATCH] fix up broken comment delimiters.
Dave Jones [Wed, 3 Apr 2002 03:21:54 +0000 (19:21 -0800)]
[PATCH] fix up broken comment delimiters.

15 years ago[PATCH] document new address space operations.
Dave Jones [Wed, 3 Apr 2002 03:21:48 +0000 (19:21 -0800)]
[PATCH] document new address space operations.

15 years ago[PATCH] Fix up broken do while macros.
Dave Jones [Wed, 3 Apr 2002 03:21:43 +0000 (19:21 -0800)]
[PATCH] Fix up broken do while macros.

15 years ago[PATCH] faster kiobuf init.
Dave Jones [Wed, 3 Apr 2002 03:21:31 +0000 (19:21 -0800)]
[PATCH] faster kiobuf init.

Originally from Intel, has been around various vendor trees for a while.
Aparently worth a noticable speed up in some applications.

15 years ago[PATCH] Extra cards support for MOXA driver
Dave Jones [Wed, 3 Apr 2002 03:21:26 +0000 (19:21 -0800)]
[PATCH] Extra cards support for MOXA driver

15 years ago[PATCH] add AMD Elan resources.
Dave Jones [Wed, 3 Apr 2002 03:21:21 +0000 (19:21 -0800)]
[PATCH] add AMD Elan resources.

15 years ago[PATCH] Detect get_block() errors in block_read_full_page()
Dave Jones [Wed, 3 Apr 2002 03:21:15 +0000 (19:21 -0800)]
[PATCH] Detect get_block() errors in block_read_full_page()

Originally from Anton Altaparmakov..

This causes errors from get_block() in block_read_full_page() to be
detected and handled properly (by setting page error flag).  Without the
patch the page (or parts of the page) will contain random data on
get_block() failing without any form of error being signalled which can
be catastrophic for filesystems using block_read_full_page() for
accessing their metadata.  And for normal data it would mean the user
would see random data instead of what they expected.

15 years ago[PATCH] export rbtree routines for modules.
Dave Jones [Wed, 3 Apr 2002 03:21:10 +0000 (19:21 -0800)]
[PATCH] export rbtree routines for modules.

15 years ago[PATCH] AMD Elan uses slightly different clock freq
Dave Jones [Wed, 3 Apr 2002 03:21:05 +0000 (19:21 -0800)]
[PATCH] AMD Elan uses slightly different clock freq

15 years ago[PATCH] DMI scanner update.
Dave Jones [Wed, 3 Apr 2002 03:21:00 +0000 (19:21 -0800)]
[PATCH] DMI scanner update.

Various quirks from 2.4 and other sources.

15 years ago[PATCH] faster update_atime.
Dave Jones [Wed, 3 Apr 2002 03:20:53 +0000 (19:20 -0800)]
[PATCH] faster update_atime.

Another 2.4 forward port, original from Andrew Morton.

15 years ago[PATCH] Document an errata workaround in apic code.
Dave Jones [Wed, 3 Apr 2002 03:20:48 +0000 (19:20 -0800)]
[PATCH] Document an errata workaround in apic code.

15 years ago[PATCH] Various completion users.
Dave Jones [Wed, 3 Apr 2002 03:20:43 +0000 (19:20 -0800)]
[PATCH] Various completion users.

These files use completion, but don't include header.

15 years ago[PATCH] Cyclades driver region cleanup
Dave Jones [Wed, 3 Apr 2002 03:20:38 +0000 (19:20 -0800)]
[PATCH] Cyclades driver region cleanup

15 years ago[PATCH] Cyrix irq router tweak
Dave Jones [Wed, 3 Apr 2002 03:20:32 +0000 (19:20 -0800)]
[PATCH] Cyrix irq router tweak

Another small change from 2.4

15 years ago[PATCH] AMD ELAN support.
Dave Jones [Wed, 3 Apr 2002 03:20:27 +0000 (19:20 -0800)]
[PATCH] AMD ELAN support.

Add support for AMD Elan.
(More ELAN patches to follow which rely on CONFIG_MELAN)

15 years ago[PATCH] extra codepage support.
Dave Jones [Wed, 3 Apr 2002 03:20:22 +0000 (19:20 -0800)]
[PATCH] extra codepage support.

Another forward port from 2.4

15 years ago[PATCH] eicon driver was sleeping with lock held.
Dave Jones [Wed, 3 Apr 2002 03:20:17 +0000 (19:20 -0800)]
[PATCH] eicon driver was sleeping with lock held.

15 years ago[PATCH] watchdog API documentation.
Dave Jones [Wed, 3 Apr 2002 03:20:11 +0000 (19:20 -0800)]
[PATCH] watchdog API documentation.

15 years ago[PATCH] 3ware driver update for 2.5.8-pre1
Adam Radford [Wed, 3 Apr 2002 03:20:06 +0000 (19:20 -0800)]
[PATCH] 3ware driver update for 2.5.8-pre1

Self explanatory driver update from vendor.

15 years ago[PATCH] Support for ITE interrupt router
Dave Jones [Wed, 3 Apr 2002 03:20:00 +0000 (19:20 -0800)]
[PATCH] Support for ITE interrupt router

Support for an extra interrupt router.

15 years ago[PATCH] about locations of various sound files.
Dave Jones [Wed, 3 Apr 2002 03:19:55 +0000 (19:19 -0800)]
[PATCH] about locations of various sound files.

introduction of ALSA moved some files around.
This updates various references.

15 years ago[PATCH] EFI GUID partition table support.
Dave Jones [Wed, 3 Apr 2002 03:19:50 +0000 (19:19 -0800)]
[PATCH] EFI GUID partition table support.

EFI GUID partition table support from Matt Domsch

15 years ago[PATCH] AGPGART capability handling cleanup
Dave Jones [Wed, 3 Apr 2002 03:19:44 +0000 (19:19 -0800)]
[PATCH] AGPGART capability handling cleanup

Clean up capability handling in AGPGart.
This came forward from 2.4

15 years ago[PATCH] small fix for mpparse.c
Greg Kroah-Hartman [Wed, 3 Apr 2002 03:19:39 +0000 (19:19 -0800)]
[PATCH] small fix for mpparse.c

Here's a very tiny bugfix for arch/i386/kernel/mpparse.c in the
2.4.19-pre2 kernel.  It fixes the problem if there is an error in the
MP_processor_info() function where the mpc_apicid value is greater than
MAX_APICS, then we need to decrement the number of valid processors
before we return (the number was just incremented before the check.)

The patch was written by James Cleverdon.

15 years ago[PATCH] fix broken asm constraint
Dave Jones [Wed, 3 Apr 2002 03:19:34 +0000 (19:19 -0800)]
[PATCH] fix broken asm constraint

Simple compile fix.

15 years ago[PATCH] Hyperthreading binfmt.
Dave Jones [Wed, 3 Apr 2002 03:19:29 +0000 (19:19 -0800)]
[PATCH] Hyperthreading binfmt.

Another from 2.4, see comments for details

15 years ago[PATCH] BKL reduction in do_exit
Dave Hansen [Wed, 3 Apr 2002 02:58:29 +0000 (18:58 -0800)]
[PATCH] BKL reduction in do_exit

Push BKL down to the (few) routines that actually need it,
remove it from the do_exit() path.

15 years agoMerge master.kernel.org:/home/davem/BK/net-2.5
Linus Torvalds [Tue, 2 Apr 2002 10:14:20 +0000 (02:14 -0800)]
Merge master.kernel.org:/home/davem/BK/net-2.5
into home.transmeta.com:/home/torvalds/v2.5/linux

15 years agoIn tcp_v4_send_reset, use inet_sk to get at
David S. Miller [Tue, 2 Apr 2002 09:51:27 +0000 (01:51 -0800)]
In tcp_v4_send_reset, use inet_sk to get at
ttl of tcp_socket.

15 years agoupdate version and defconfig
Linus Torvalds [Tue, 2 Apr 2002 09:30:07 +0000 (01:30 -0800)]
update version and defconfig

15 years ago[PATCH] brw_kiovec() converted to struct block_device *
Alexander Viro [Tue, 2 Apr 2002 09:17:36 +0000 (01:17 -0800)]
[PATCH] brw_kiovec() converted to struct block_device *

brw_kiovec() and ll_rw_kiovec() switched to struct block_device *.

15 years ago[PATCH] fsync_bdev() conversion
Alexander Viro [Tue, 2 Apr 2002 09:17:31 +0000 (01:17 -0800)]
[PATCH] fsync_bdev() conversion

Bunch of places converted from fsync_dev/invalidate_buffers to
fsync_bdev/invalidate_bdev.

15 years ago[PATCH] restoring block size upon umount
Alexander Viro [Tue, 2 Apr 2002 09:17:27 +0000 (01:17 -0800)]
[PATCH] restoring block size upon umount

get_sb_bdev() stores original block size in ->s_old_blocksize and
kill_block_super() restores it.

This kills 99% of crap with "oh, I've mounted/umounted that device and
its behaviour had changed" (remaining 1% can be dealt in pretty similar
ways; ideally I'd like to see ioctls that get/set block size dead and
gone).

15 years ago[PATCH] hfs compile fix
Alexander Viro [Tue, 2 Apr 2002 09:17:23 +0000 (01:17 -0800)]
[PATCH] hfs compile fix

Fix typo

15 years ago[PATCH] set_blocksize() in JFS
Alexander Viro [Tue, 2 Apr 2002 09:17:19 +0000 (01:17 -0800)]
[PATCH] set_blocksize() in JFS

Use sb_set_blocksize() in JFS instead of trying to do it by hand.

15 years ago[PATCH] minixfs cleanups (4/4)
Alexander Viro [Tue, 2 Apr 2002 09:17:15 +0000 (01:17 -0800)]
[PATCH] minixfs cleanups (4/4)

Clean up minix/bitmap.c

15 years ago[PATCH] minixfs cleanups (3/4)
Alexander Viro [Tue, 2 Apr 2002 09:17:10 +0000 (01:17 -0800)]
[PATCH] minixfs cleanups (3/4)

Kill BKL in minix/itree* (similar to ext2 patches)

15 years ago[PATCH] minixfs cleanups (2/4)
Alexander Viro [Tue, 2 Apr 2002 09:17:06 +0000 (01:17 -0800)]
[PATCH] minixfs cleanups (2/4)

Clean up the write_super-related code in minixfs

15 years ago[PATCH] minixfs cleanups (1/4)
Alexander Viro [Tue, 2 Apr 2002 09:17:02 +0000 (01:17 -0800)]
[PATCH] minixfs cleanups (1/4)

Move minix-private stuff to fs/minix/minix.h.

15 years ago[PATCH] conditional system call cleanup
Alexander Viro [Tue, 2 Apr 2002 09:16:17 +0000 (01:16 -0800)]
[PATCH] conditional system call cleanup

This version of sys_nfsservctl() fix hadn't made DaveM complain.

15 years ago[PATCH] romfs inode allocation
Alexander Viro [Tue, 2 Apr 2002 09:16:13 +0000 (01:16 -0800)]
[PATCH] romfs inode allocation

Obvious romfs fix.

15 years agoMerge master.kernel.org:/home/davem/BK/net-2.5
Linus Torvalds [Tue, 2 Apr 2002 09:08:15 +0000 (01:08 -0800)]
Merge master.kernel.org:/home/davem/BK/net-2.5
into home.transmeta.com:/home/torvalds/v2.5/linux

15 years agoMerge master.kernel.org:/home/davem/BK/sparc-2.5
Linus Torvalds [Tue, 2 Apr 2002 09:04:13 +0000 (01:04 -0800)]
Merge master.kernel.org:/home/davem/BK/sparc-2.5
into home.transmeta.com:/home/torvalds/v2.5/linux

15 years ago[PATCH] ext2_fill_super breakage
Andrew Morton [Tue, 2 Apr 2002 03:44:16 +0000 (19:44 -0800)]
[PATCH] ext2_fill_super breakage

In 2.5.7 there is a thinko in the allocation and initialisation
of the fs-private superblock for ext2.  It's passing the wrong type
to the sizeof operator (which of course gives the wrong size)
when allocating and clearing the memory.

Lesson for the day: this is one of the reasons why this idiom:

some_type *p;

p = malloc(sizeof(*p));
...
memset(p, 0, sizeof(*p));

is preferable to

some_type *p;

p = malloc(sizeof(some_type));
...
memset(p, 0, sizeof(some_type));

I checked the other filesystems.  They're OK (but idiomatically
impure).  I've added a couple of defensive memsets where
they were missing.

15 years ago[PATCH] 2.5.7 IDE 28a
Martin Dalecki [Tue, 2 Apr 2002 03:41:07 +0000 (19:41 -0800)]
[PATCH] 2.5.7 IDE 28a

- Apply Pavel Macheks suspend resume double resume fix.

- Finally remove the busy field for ata_operations and replace it with
   MOD_INC_USE_COUNT and MOD_DEC_USE_COUNT.

- Fix ali15xx chipset support by removing initialization differences,
   apparently caused by mislead interpretation of the specs or a mismatch
   between the specification and reality.

- Guard calls to ide_set_handler with checks to see whatever the previously
   installed IRQ handler already served it's purpose.

- Convert timeout checks on poll_timeout to the time_before() interface.

- Consolidate the two different IRQ handlers for multi mode PIO writes into
   one. The problems remain the same but at least now we will only have to
   tangle one single problem.

15 years ago[PATCH] 2.5.7 IDE 27
Martin Dalecki [Tue, 2 Apr 2002 03:41:01 +0000 (19:41 -0800)]
[PATCH] 2.5.7 IDE 27

- Make for less terse error messages in ide-tape.c.

- Replaced all timecomparisions done by hand with all the proper timer_after()
   commands.

- Remove the drive niec1 mechanisms alltogether. There are several reasons for
   this:

   1. The code implementing it is nonintelliglible and therefore propably
   broken.

   2. If we have to invent somethig about sceduling drive IO, it should be done
   on the BIO level.

   3. We may in fact interleave with the IO sceduling on the upper layers and
   the results of two overlapping signal filters overlapped with each other can
   be disasterous to the overall throughput. (In fact they *are* most of the
   time.)

   4. It was not working if you had intermixed modes on different drives
   DMA versus PIO.

   5. Our goal is to have a driver which is able to share the badwidth
   properly and shouldn't needing this kind of "tuning".

- Remove unused nice2 from disk struct.

- Rename channel member of ata_channel to unit and device to dev to
   just prevent wrong interpretations. This prevents constructs like
   channel->channel...

15 years ago[PATCH] 2.5.7 IDE 26
Martin Dalecki [Tue, 2 Apr 2002 03:39:39 +0000 (19:39 -0800)]
[PATCH] 2.5.7 IDE 26

- Mark all members of structures, which get jiffies assigned or involved in
   ugly timeout calculations with the prefix PADAM_  for easy spotting. This is
   Polish for "I'm falling down" or "This brings me to the knees" or slag
   comment for "What a sh..". Please be assured that it doesn't sound vulgar.

   Please grep for it to see immediately why this nomenclature is justified.

- Rename hwifs_s to ata_channel and eliminate ide_hwifs_t as well as the HWIF
   macro. OK this step makes this patch rather big.

15 years ago[PATCH] 2.5.7 IDE 25
Martin Dalecki [Tue, 2 Apr 2002 03:37:33 +0000 (19:37 -0800)]
[PATCH] 2.5.7 IDE 25

- Replace the task_io_reg_t with the simple u8. There is no need to obfuscate
   the code more then necessary.

- kill some unnecessary type definitions out from hdreg.h.

- Add proper attributes to register files in hdreg.h.

- Don't use raw arrays for tfRegister and hobRegister in ide_task_s.  Use out
   nice global structures describing the fields in them.  This allows to kill
   the following defines:

     IDE_DATA_OFFSET
     IDE_FEATURE_OFFSET
     IDE_NSECTOR_OFFSET
     IDE_SECTOR_OFFSET
     IDE_LCYL_OFFSET
     IDE_HCYL_OFFSET
     IDE_SELECT_OFFSET
     IDE_COMMAND_OFFSET

   and many many others.

- Please have a look at the following in pdc4030.c. It couldn't have worked!
   This has been fixed in one go with the above change:

   memcpy(args.hobRegister, NULL, sizeof(struct hd_drive_hob_hdr));

- Kill the redundant *_REG_HOB definitions. They don't help readability in any
   way.

15 years ago[PATCH] 2.5.7 IDE 24
Martin Dalecki [Tue, 2 Apr 2002 03:37:27 +0000 (19:37 -0800)]
[PATCH] 2.5.7 IDE 24

- Push BAD_DMA_DRIVE and GOOD_DMA_DRIVE to the ide-pmac.c file, since this is
   the only place where those get used.

- Kill unused fields from the ide_task_s structure. In esp. we pass a task
   attached to a request and not the other way around!

- Rename hwif field to channel in struct ide_drive_s.

- Move the request queue to the level where proper serialization has to happen
   anyway - the channel structure.

15 years ago[PATCH] 2.5.7 IDE 23
Martin Dalecki [Tue, 2 Apr 2002 03:37:22 +0000 (19:37 -0800)]
[PATCH] 2.5.7 IDE 23

- Support for additional Promise controller id's (PDC20276).

- Remove code duplication between do_rw_taskfile and do_taskfile.
   This will evolve into a more reasonable ata_command() function
   finally. The ata_taskfile function has far too many arguments, but
   I favour this over having two different code paths for getting
   actual data to the drive.

15 years ago[PATCH] initrd issue
Alexander Viro [Tue, 2 Apr 2002 03:31:25 +0000 (19:31 -0800)]
[PATCH] initrd issue

Fix initrd problem that appeared back in 2.5.2-pre6 when kdev_t type
changed and comparison function was incorrectly converted..

15 years agonet/core/sock.c needs linux/tcp.h to get at TCP state macros.
David S. Miller [Sat, 30 Mar 2002 14:26:25 +0000 (06:26 -0800)]
net/core/sock.c needs linux/tcp.h to get at TCP state macros.

15 years agoSparc SBUS fix: Make for_all_sbusdev work with an empty SBUS.
David S. Miller [Thu, 28 Mar 2002 08:02:25 +0000 (00:02 -0800)]
Sparc SBUS fix: Make for_all_sbusdev work with an empty SBUS.

15 years agoTigon3 net driver bug fix:
David S. Miller [Wed, 27 Mar 2002 14:10:21 +0000 (06:10 -0800)]
Tigon3 net driver bug fix:
- ETHTOOL_GLINK handling forgets to return on success

15 years agoIn SBUS probing, handle empty SBUS correctly.
David S. Miller [Wed, 27 Mar 2002 08:57:22 +0000 (00:57 -0800)]
In SBUS probing, handle empty SBUS correctly.

15 years agoTigon3 net driver fixes:
David S. Miller [Wed, 27 Mar 2002 08:15:52 +0000 (00:15 -0800)]
Tigon3 net driver fixes:
- Make use of pci_unmap_xxx storage optimization.
- In tg3_get_invariants, fix cacheline_sz_reg shifts.
  On 5703 chips, if latency timer is less than 64, set
  it to 64.
- In tg3_test_dma, only run the actual test on 5700/5701
  chips.  Aparently 5702/5703 revs have some problems.

15 years agoSunHME driver updates:
David S. Miller [Tue, 26 Mar 2002 09:45:37 +0000 (01:45 -0800)]
SunHME driver updates:
- Fix SMP locking throughout
- Deal with hw bug involving lost writes to rxring base address
- Disable Never Give Up mode on transmitter, set TX attempts
  limit to 16 (the default).  This deals with a possible hang
  when the TX deadlocks with the PHY when a jabber occurs.
- Delete some dead code.

15 years agoBig netfilter newnat patch for 2.5.7:
Harald Welte [Tue, 26 Mar 2002 07:18:27 +0000 (23:18 -0800)]
Big netfilter newnat patch for 2.5.7:
- support for multiple expected connections
  (necessary for protocols like H.323, SIP, PPTP)
- helper-definable limit of unconfirmed expectations
- timeouts for expectations
- full graph of connection relations, even after expectation confirmed
- various changes in the API towards conntrack and NAT helper
- automatic conntrack helper loading when at helper is loaded
- NAT mangling of TCP SACK in case of sequence number alteration
  (no need to delete SACKPERM anymore, I hope Alexey is happy now)

15 years agoFix device list locking.
David S. Miller [Tue, 26 Mar 2002 07:11:11 +0000 (23:11 -0800)]
Fix device list locking.
Based upon a patch from Maxim Krasnyansky.