2003-01-10 20:19:39

by Linus Torvalds

[permalink] [raw]
Subject: Linux v2.5.56


Trying to make releases slightly more often and slightly smaller.

ACPI, USB, networking (mainly netfilter) updates. Some syscall path
updates and a thread bug in mm_release() that would miss updating the TID
and cause a few extra traps at exec time.

And a watchdog forward port from 2.4.x by DaveJ.

Linus

----
Summary of changes from v2.5.55 to v2.5.56
============================================

<[email protected]>:
o [IPV4 ROUTE]: Fix some sysctl documentation

<[email protected]>:
o [NETFILTER]: Fix a locking bug in ip_conntrack_proto_tcp

<[email protected]>:
o [NETFILTER]: Fix excess logging of reused FTP expectations

<[email protected]>:
o [NETFILTER]: ipt_multiport invert fix

<[email protected]>:
o fix "assignment from incompatible pointer type"

<[email protected]>:
o USB Serial patch for old pl2303 devices

<[email protected]>:
o [NETFILTER]: UDP nat helper support

<[email protected]>:
o 2.5.54 dev_*(&<dev>,...): drivers/usb/input/pid.c

Alan Cox <[email protected]>:
o more unusual USB storage devices

Anders Gustafsson <[email protected]>:
o [IPV6]: cleanup_ipv6_mibs cannot be __exit, since it is called on
the ipv6_init error path

Andrew Morton <[email protected]>:
o [NET]: Uninline skb_headerinit
o [AF_UNIX]: Uninline unix_get_socket/maybe_unmark_and_push, mark
{pop,empty}_stack static
o [UNIX]: Uninline unix_peer_get
o [IPSEC]: Uninline _decode_session
o [IPV4 ROUTE]: Uninline rt_hash_code and rt_may_expire
o [IPV4 OUTPUT]: Uninline ip_finish_output and skb_fill_page_desc
o [IPV4 FRAG]: Uninline ipq_kill
o [IPV4 FIBHASH]: extern inline --> static inline
o [IPV4 TCP]: Dont export or inline __tcp_put_port, but do inline
tcp_put_port
o [IPV4 TCP]: Uninline tcp_rtt_estimator and tcp_urg. extern inline
--> static inline

Andy Grover <[email protected]>:
o ACPI: Use printk instead of pr_debug (Randy Dunlap)
o ACPI: Remove typedefs in favor of using "struct" and "union"
explicitly
o ACPI: Expose lid state to userspace (Zdenek OGAR Skalak)
o ACPI: Make button functions static (Pavel Machek)
o ACPI: Express state of lid in words, not a number
o cpufreq-ACPI: no longer use CPUFREQ_ALL_CPUS (Dominik Brodowski)
o ACPI: Eliminate spawning of thread from timer callback. Use
schedule_work for all cases. Thanks to Ingo Oeser, Andrew Morton,
and Pavel Machek for their wisdom.
o ACPI: Update version string to 20030109

Dave Jones <[email protected]>:
o [WATCHDOG] wdt_pci nowayout fixes from 2.4
o [WATCHDOG] eurotech indentation fixes
o [WATCHDOG] eurotech nowayout fixes from 2.4
o [WATCHDOG] wdt nowayout changes from 2.4
o [WATCHDOG] wdt977 nowayout fixes from 2.4
o c99 initialisers
o [WATCHDOG] Add several new watchdog drivers from 2.4
o [WATCHDOG] pcwd driver update from 2.4
o [WATCHDOG] acquirewdt nowayout fixes from 2.4 (plus some
CodingStyle reformatting)
o [WATCHDOG] Acquirewdt C99 struct initialisers
o [WATCHDOG] Advantech fixes from 2.4
o [WATCHDOG] simplify advwdt_open, and add C99 struct initialisers
o [WATCHDOG] Fix up incorrect C99 struct conversion
o [WATCHDOG] acquirewdt compile fixes
o [WATCHDOG] advantech compile fixes
o [WATCHDOG] ALIM7101 fixes from 2.4 + C99 structs
o [WATCHDOG] More alim7101 cleanups
o [WATCHDOG] i810-tco fix from 2.4
o [WATCHDOG] ib700wdt fixes from 2.4
o [WATCHDOG] ib700wdt c99 structs
o [WATCHDOG] indydog nowayout fixes from 2.4
o [WATCHDOG] machzwd nowayout fixes from 2.4
o [WATCHDOG] mixcomwd nowayout fixes from 2.4
o [WATCHDOG] pcwd nowayout fixes from 2.4
o [WATCHDOG] sbc60xxwdt nowayout fixes from 2.4
o [WATCHDOG] SC1200WDT nowayout fixes from 2.4
o [WATCHDOG] SC520 nowayout fixes from 2.4
o [WATCHDOG] C99 struct initialisers for shwdt
o [WATCHDOG] softdog nowayout fixes from 2.4
o [WATCHDOG] w83877f nowayout fixes from 2.4
o [WATCHDOG] nowayout fixes for wafer5823

David Brownell <[email protected]>:
o ehci, remove potential hangs
o zaurus B500 (sl-5600?) & usbnet
o 2.5.54 -- ohci-dbg.c: 358: In function `show_list': `data1'
o usbtest, covers control queueing and fault cleanup

David S. Miller <[email protected]>:
o [SUNZILOG]: Adapt sun4u get_zs for Peters new scanning scheme
o [CRYPTO]: Fix typo in aes.o rule
o [NET]: Kill __tcp_put_port module export
o [TCP]: Fix tcp_put_port declaration
o [IPSEC]: Dont check algorithm availability unless CONFIG_CRYPTO
o [IPSEC]: Kill warning in xfrm_algo.c
o [CRYPTO]: Use appropriate defaults if AH/ESP is enabled
o [SUNZILOG]: Fix uart_get_baud_rate args
o [AIC7XXX]: Include asm/io.h, necessary to get at inb/outb/etc
o [SPARC64]: Update defconfig

Duncan Sands <[email protected]>:
o USB: atmsar is not a module
o USB: speedtouch missing __init and __exit
o USB: speedtouch: add GPL notices

Duncan Sands <[email protected]>:
o USB: speedtouch: remove version string duplication

Erich Focht <[email protected]>:
o small migration thread fix

Filip Sneppe <[email protected]>:
o [NETFILTER]: ip_conntrack_ftp.c, fixes a typo in a DEBUG statement

Gabriel Paubert <[email protected]>:
o 'iret' segment fixup

Greg Kroah-Hartman <[email protected]>:
o USB: revert davem's compile time fix, now that it's fixed properly
o USB brlvger: Forward port 2.4 fix for misuse of types
o USB: removed MOD_INC_USE_COUNT and MOD_DEC_USE_COUNT from driver
that do not need it
o USB printer driver: forward port 2.4 fix for misuse of types
o USB mdc800: forward port 2.4 fix for misuse of types
o DEV: change dev_printk() to take a pointer to dev instead of the
structure itself
o USB: drivers/usb/core/ fixups due to dev_printk change
o USB: drivers/usb/host/ fixups due to dev_printk change
o USB: drivers/usb/serial/ fixups due to dev_printk change
o USB serial: pass the usb_device_id to the probe() function
o USB serial: fixup for probe function paramaters changing
o USB: fix ehci build for older versions of gcc

Harald Welte <[email protected]>:
o [NETFILTER]: This patch fixes the ULOG target when logging packets
without any ethernet header (mac address).

Henning Meier-Geinitz <[email protected]>:
o scanner.c: fix race in ioctl_scanner()
o USB scanner driver: updated documentation
o USB scanner driver: updated Kconfig
o scanner.c, scanner.h: Added vendor/product ids
o scanner.c: print user-supplied ids only on start-up
o scanner.c, scanner.h: Remove PV8630 ioctls
o [PATCH 2.5.54] scanner.c: endpoint detection cleanup
o scanner.c, scanner.h: Use symbolic name for interface class

James Morris <[email protected]>:
o [IPSEC]: Clean up key manager algorithm handling
o [CRYPTO]: Add AES algorithm
o [SUNSAB]: Comment out powering down of chip for now
o [CRYPTO]: More credits for AES

Linus Torvalds <[email protected]>:
o Fix exec_mmap() to release the MM while we still have it active, to
properly de-activate it and make the child_tid logic work
correctly.
o Fix kallsyms symbol lookup code. Let's do this trivial
one-character version before looking at more complicated changes.
o Make psmouse driver _much_ more lenient about packet data timeouts

Luca Barbieri <[email protected]>:
o Use %ebp rather than %ebx for thread_info pointer
o Remove all register pops before sysexit

Matthew Dharm <[email protected]>:
o USB storage: remove usb_stor_tranfer_length()

Oliver Neukum <[email protected]>:
o USB: kaweth freeing skbs

Patrick McHardy <[email protected]>:
o [NETFILTER]: Fix ipt_REJECT udp checksums
o [NETFILTER]: Fix incremental TCP checksum in ECN module

Paul Mackerras <[email protected]>:
o PPC32: Add support for the IBM PPC 405GPR-based "Sycamore" board
o PPC32: Handle machine checks on 4xx processors better

Petko Manolov <[email protected]>:
o USB pegasus: small patch for 2.5
o again rtl8150

Rob Radez <[email protected]>:
o [SPARC32]: Copy over sparc64 exception table changes

Robert Olsson <[email protected]>:
o [NAPI]: Discuss some more issues in driver HOWTO

Russell King <[email protected]>:
o [SERIAL] Add prototypes and rename UPF_FLAGS
o [SERIAL] Remove unused info->event
o [SERIAL] Convert change_speed() to settermios()
o [SERIAL] Change settermios to set_termios
o [SERIAL] Bug fix: remove infinite loop in sa1100 serial driver
o [SERIAL] Restrict the baud rates returnable from
uart_get_baud_rate()
o [SERIAL] Fix build errors caused in previous cset

Stephen Rothwell <[email protected]>:
o [COMPAT]: Sparc64 use get/put compat_timespec

Tom Rini <[email protected]>:
o PPC32: Default to 768MB of lowmem



2003-01-11 05:38:54

by Udo A. Steinberg

[permalink] [raw]
Subject: Re: Linux v2.5.56

On Fri, 10 Jan 2003 12:26:56 -0800 (PST) Linus Torvalds (LT) wrote:

LT> Summary of changes from v2.5.55 to v2.5.56
LT> ============================================

Hello,

I just got the following bug with 2.5.56 pretty much out of the blue:

VFS: brelse: Trying to free free buffer
buffer layer error at fs/buffer.c:1182
Call Trace:
[<c0147105>] __brelse+0x35/0x40
[<c01472c9>] bh_lru_install+0xa9/0xe0
[<c0147371>] __find_get_block+0x71/0x80
[<c0146fa3>] __getblk_slow+0x23/0x100
[<c01473cf>] __getblk+0x4f/0x60
[<c01473ff>] __bread+0x1f/0x40
[<c0180265>] ext3_get_inode_loc+0xe5/0x190
[<c018033f>] ext3_read_inode+0x2f/0x390
[<c015cd5e>] iget_locked+0x6e/0x70
[<c0182908>] ext3_lookup+0xe8/0x100
[<c0151330>] real_lookup+0xc0/0x100
[<c015159a>] do_lookup+0x11a/0x170
[<c01519e2>] link_path_walk+0x3f2/0x790
[<c0152239>] __user_walk+0x49/0x60
[<c014d8fc>] vfs_lstat+0x1c/0x60
[<c014df3b>] sys_lstat64+0x1b/0x40
[<c010933b>] syscall_call+0x7/0xb

Regards,
-Udo.


Attachments:
(No filename) (189.00 B)

2003-01-11 06:12:12

by Andrew Morton

[permalink] [raw]
Subject: Re: Linux v2.5.56

On Fri January 10 2003 21:47, Udo A. Steinberg wrote:
>
> On Fri, 10 Jan 2003 12:26:56 -0800 (PST) Linus Torvalds (LT) wrote:
>
> LT> Summary of changes from v2.5.55 to v2.5.56
> LT> ============================================
>
> Hello,
>
> I just got the following bug with 2.5.56 pretty much out of the blue:
>
> VFS: brelse: Trying to free free buffer
> buffer layer error at fs/buffer.c:1182
> Call Trace:
> [<c0147105>] __brelse+0x35/0x40

Presumably one of your filesystems is using htree (indexed directories).
It's a bug - nobody has found it yet.

The modestly good news is that I have found a workload which triggers it in
ten minutes, but I have not looked into it at all.

The modestly bad news is that e2fsprogs-1.32's mke2fs enables htree by
default. It really shouldn't be doing that - htree isn't ready yet.

Use

dumpe2fs -h /dev/hda1 | grep dir_index

to see if htree is enabled.

Use

tune2fs -O '^dir_index' /dev/hda1

to disable it.


There are no trmendously bad bugs in htree. Mainly this one (which _could_
cause crashes and scribbled directories) and a small memory leak on each
unlink.


2003-01-11 09:35:04

by Adrian Bunk

[permalink] [raw]
Subject: Re: Linux v2.5.56

On Fri, Jan 10, 2003 at 12:26:56PM -0800, Linus Torvalds wrote:

>...
> Summary of changes from v2.5.55 to v2.5.56
> ============================================
>...
> Dave Jones <[email protected]>:
>...
> o [WATCHDOG] Add several new watchdog drivers from 2.4
>...

FYI:

drivers/char/watchdog/sc1200wdt.c doesn't compile:

<-- snip -->

...
gcc -Wp,-MD,drivers/char/watchdog/.sc1200wdt.o.d -D__KERNEL__
-Iinclude -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common
-pipe -mpreferred-stack-boundary=2 -march=k6 -Iinclude/asm-i386/mach-default
-nostdinc -iwithprefix include -DKBUILD_BASENAME=sc1200wdt
-DKBUILD_MODNAME=sc1200wdt -c -o drivers/char/watchdog/sc1200wdt.o
drivers/char/watchdog/sc1200wdt.c
drivers/char/watchdog/sc1200wdt.c: In function `sc1200wdt_isapnp_probe':
drivers/char/watchdog/sc1200wdt.c:339: warning: implicit declaration of function `isapnp_find_dev'
drivers/char/watchdog/sc1200wdt.c:339: warning: assignment makes pointer from integer without a cast
drivers/char/watchdog/sc1200wdt.c:343: structure has no member named `prepare'
drivers/char/watchdog/sc1200wdt.c:353: structure has no member named `activate'
drivers/char/watchdog/sc1200wdt.c: In function `sc1200wdt_init':
drivers/char/watchdog/sc1200wdt.c:427: structure has no member named `deactivate'
drivers/char/watchdog/sc1200wdt.c: In function `sc1200wdt_exit':
drivers/char/watchdog/sc1200wdt.c:440: structure has no member named `deactivate'
make[3]: *** [drivers/char/watchdog/sc1200wdt.o] Error 1

<-- snip -->

cu
Adrian

--

"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed

2003-01-11 09:48:07

by Zwane Mwaikambo

[permalink] [raw]
Subject: Re: Linux v2.5.56

On Sat, 11 Jan 2003, Adrian Bunk wrote:

> On Fri, Jan 10, 2003 at 12:26:56PM -0800, Linus Torvalds wrote:
>
> >...
> > Summary of changes from v2.5.55 to v2.5.56
> > ============================================
> >...
> > Dave Jones <[email protected]>:
> >...
> > o [WATCHDOG] Add several new watchdog drivers from 2.4
> >...
>
> FYI:
>
> drivers/char/watchdog/sc1200wdt.c doesn't compile:

I can take care of that, although i think this driver could be removed due
scx200_wdt.c supporting the same hardware. I'll look into it.

Zwane
--
function.linuxpower.ca

2003-01-11 23:51:04

by Dave Jones

[permalink] [raw]
Subject: Re: Linux v2.5.56

On Sat, Jan 11, 2003 at 10:43:45AM +0100, Adrian Bunk wrote:

> drivers/char/watchdog/sc1200wdt.c doesn't compile:

Yup, it didn't before I poked it too, it needs updating to the new PNP-fu

Dave

--
| Dave Jones. http://www.codemonkey.org.uk

2003-01-11 23:56:09

by Dave Jones

[permalink] [raw]
Subject: Re: Linux v2.5.56

On Fri, Jan 10, 2003 at 12:26:56PM -0800, Linus Torvalds wrote:
>
> Trying to make releases slightly more often and slightly smaller.
>
> ACPI, USB, networking (mainly netfilter) updates. Some syscall path
> updates and a thread bug in mm_release() that would miss updating the TID
> and cause a few extra traps at exec time.
>
> And a watchdog forward port from 2.4.x by DaveJ.

just to stem the number of 'this still isnt finished' reports I'm
getting, I'm working through the 2.4 diffs incrementally. I'm not
done yet, so please, be patient..

Dave

--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs