Mostly just keeping up with the Torvalds & Tossati's.
I've not merged any of the mips/ia64/ppc etc changes from 2.4,
and I'm looking at dropping whats in my tree that touches those soon.
The rationale being that the arch maintainers can do (and will do)
a better job at pushing those to Linus than I can, and as I've no way of
testing a lot of these changes, fixing up rejects in foreign asm is nigh
on impossible.
It's likely I won't put out too many patches in the next week or so, with
both a pending housemove, and experimenting with bitkeeper to try and
get much of this huge diff pushed Linus' way.
I'll still be picking up the overspill/syncing with -current however.
As usual,..
Patch against 2.5.5 vanilla is available from:
ftp://ftp.kernel.org/pub/linux/kernel/people/davej/patches/2.5/
Merged patch archive: http://www.codemonkey.org.uk/patches/merged/
Also, please check http://www.codemonkey.org.uk/Linux-2.5.html before
reporting known bugs that are also in mainline.
-- Davej.
2.5.5-dj3
o Merge 2.5.6pre2
o Merge 2.4.19pre2
| Drop various arch bits. (See above)
| Drop LVM bits.
| Drop some SIS bits that rejected -- James please take a look.
| Fix some obvious sillies as per 2.4.19pre2ac1
o Add missing struct initialiser to nls_cp850 (OGAWA Hirofumi)
o natsemi rx/tx ring buffer confusion fixup. (Jeff Garzik)
o Backout bogus isofs makefile changes. (Me)
o Fix sys_shmdt return code. (Andreas Schwab)
o Numerous net driver last_rx = jiffies fixes. (Me, Celso Gonzalez)
o msync/mprotect POSIX fixes. (Thorsten Kukuk)
o Small SCSI generic updates. (Douglas Gilbert)
o Allow short commands in SCSI debug driver. (Douglas Gilbert)
o Update helptext for Machine Check. (Paul Gortmaker)
o Updated email address. (Thomas Molina)
o Fix potential oops in swapfile code. (Andries Brouwer)
o Reset/Reservation handling fixes. (James Bottomley)
o Various console fixes. (James Simmons)
o Update numerous fb drivers to new api. (James Simmons)
o Twin inquiry mode for SCSI LUN scanning. (Patrick Mansfield,
Douglas Gilbert)
o Update DMI / local APIC fixes. (Mikael Pettersson)
o Emu10k1 OSS highmem dma fixes. (Daniel Bertrand)
o Export dparent_lock (Petr Vandrovec)
o Clean up struct page shrinkage. (Rik van Riel, Dave Miller)
o More Config.help updates. (Steven Cole)
o Make highmem pte a boolean. (Steven Cole)
o opl3sa2 modular compile fix. (Zwane Mwaikambo)
o es18xx compile fix. (Zwane Mwaikambo)
o nfsd modular compile fix. (Stelian Pop)
o Use list heads for task list. (Brian Gerst)
o Fix up various compilation warnings. (Roberto Nibali)
o Power management for 3c509. (Zwane Mwaikambo)
o Small kbuild cleanup. (John Levon)
2.5.5-dj2
o Merge 2.5.6pre1
o Merge up to 2.4.19pre1
o Merge IDE changes from 2.5.5
o Drop S390 changes.
| These were incomplete & old, and would be better
| merged by the maintainers anyway.
o Fix decvte console graphic mode. (Nick Pasich)
o Chop non x86-64 bits out of x86-64 mtrr.c (Me)
o Eliminate annoying warning in serial.c (Andrey Panin)
o Use named initialisers in various filesystems. (Todor Todorov)
o More devexit fixes. (Andrew Morton)
o Fix null pointer checks in sd.c (Peter Wong)
o Further multi-desktop console work. (James Simmons)
o Docbook compile fixes. (Erlend Aasland)
o Additional Config.help texts. (Dan Carpenter,
Steven Cole)
o Convert neofb to new framebuffer api. (James Simmons)
o Sanitise arguments to mempool_create() (Balbir Singh)
o Fix binfmt_elf modular compile. (Paul Gortmaker)
o Only fsync() blockdev on last close/umount. (Miquel van Smoorenburg)
o Promise IDE driver update. (Peter Denison)
o irq router recognition for Intel ICH chipsets. (Wim Van Sebroeck)
o IPv6 initialisation workaround. (Ben Clifford)
| Works, but may not be correct fix.
o Dump x86 MCE MISC register in right order. (Me)
o NFSv3->NFSv2 READDIRPLUS fix. (Trond Myklebust)
o More /proc/net signedness fixes. (Arnaud Giersch)
o Fix JFFS2 duplicate slabcache name. (Me)
o mtdblock bio work. (Me)
| Quick hack to test JFFS2 changes, WorksForMe(tm).
o Various other small MTD/JFFS2 fixes. (Me)
o sstfb support for interlace / doublescan modes. (Urs Ganse)
o Use correct timings for PIIX secondary slave. (Daniel Quinlan)
o Fix up some borken comments. (John Kim)
2.5.5-dj1
o Merge 2.5.5 final.
| Backout broken IDE changes.
o Implement proper locking in ALSA lseek methods. (Robert Love)
o Document lseek locking. (Robert Love)
o ALSA + YMFPCI compile fixes. (Stelian Pop)
o Further console reentrancy work. (James Simmons)
o NFS compile fix. (Neil Brown)
o Fix up some strsep changes from last time. (Ren? Scharfe)
o tmpfs link-count on dir rename fixes. (Christoph Rohland)
o USB vicam driver build fixes. (Greg KH)
o Split up terminal emulation. (James Simmons)
o Fix scsi_merge crash-on-boot problem. (Jens Axboe)
2.5.4-dj3
o Merge up to 2.4.18rc2
o Change <linux/malloc.h> -> <linux/slab.h> (Me)
o Fix borken locking in nfs ->lookup. (Jarno Paananen)
o Fix ext2 freeing blocks not in datazone. (Randy Hron, Chris Wright)
o Fix ext2/ext3 revision level checks. (Andreas Dilger)
o Fix ramdisk compilation failure. (Me, Rudmer van Dijk)
o More include dependancy tweaks. (Me)
o BSS janitor work. (Craig Christophel)
o Replace all strtok users with strsep. (Matthew Hawkins, Jason Thomas)
o scsi_debug ->address & other fixes. (Douglas Gilbert)
o Silence isapnp debug messages. (Andrey Panin)
o Clear passcred in sock_alloc() (OGAWA Hirofumi)
| Fixes slow sunrpc/portmap, and various
| gnome-terminal weirdness.
o Console reentrancy work. (James Simmons)
o ALSA Config.in fixes. (Ren? Scharfe)
o Fix Oxford Semiconductor PCI id. (Ed Vance)
o Power Management for es18xx. (Zwane Mwaikambo)
o Remove duplicate PCI ids. (Wim Van Sebroeck)
o Change Olympic driver to use spinlocks. (Mike Phillips)
o Fix pcilynx locking. (Manfred Spraul)
o Fix cris eeprom driver locking. (Robert Love)
o PPP/BSD Compression vfree in interrupt fix. (Paul Mackerras,
Dominik Brodowski)
o cli->spinlocks for aha1542 driver. (Douglas Gilbert)
o ALSA ISAPNP fixes. (Andrey Panin)
o /proc/net/udp signedness fix. (Arnaud Giersch)
o fcntl_[gs]etlk* cleanup. (Chris Wright)
--
Dave Jones. http://www.codemonkey.org.uk
SuSE Labs.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
in drivers/ide/ide_modes.h,
typedef ... ide_pio_timings_t;
is only defined #ifdef CONFIG_BLK_DEV_IDE_MODES.
But it is used in ide.c without any ifdefs around it, resulting in a
compile error.
In 2.5.5-dj2, this block was in ide_modes.h within the same #ifdef as the
typedef, but was moved by the -dj3 patch.
- --
Ben Clifford [email protected] GPG: 30F06950
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE8hb0LsYXoezDwaVARApQzAJ43FGatrKZU/Dht5bEgsRPwCYqNagCfUsMu
mjD6zffn1bgeJtyYjn6O3ng=
=mEOw
-----END PGP SIGNATURE-----
On Tue, Mar 05, 2002 at 10:53:58PM -0800, Ben Clifford wrote:
> in drivers/ide/ide_modes.h,
> typedef ... ide_pio_timings_t;
> is only defined #ifdef CONFIG_BLK_DEV_IDE_MODES.
> But it is used in ide.c without any ifdefs around it, resulting in a
> compile error.
> In 2.5.5-dj2, this block was in ide_modes.h within the same #ifdef as the
> typedef, but was moved by the -dj3 patch.
It came from the 2.5.6pre2 merge. Hopefully the next round of Martins
patches will fix that up.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
Dave Jones wrote:
> On Tue, Mar 05, 2002 at 10:53:58PM -0800, Ben Clifford wrote:
> > in drivers/ide/ide_modes.h,
> > typedef ... ide_pio_timings_t;
> > is only defined #ifdef CONFIG_BLK_DEV_IDE_MODES.
> > But it is used in ide.c without any ifdefs around it, resulting in a
> > compile error.
> > In 2.5.5-dj2, this block was in ide_modes.h within the same #ifdef as the
> > typedef, but was moved by the -dj3 patch.
>
> It came from the 2.5.6pre2 merge. Hopefully the next round of Martins
> patches will fix that up.
Yes I see the fault.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Here's another one.
2.5.5-dj3 removes EXPORT_SYMBOL(ide_fops)
but doesn't remove ide_fops from the code.
Hence modprobe ide-cd doesn't work.
- --
Ben Clifford [email protected] GPG: 30F06950
Live Ben-cam: http://barbarella.hawaga.org.uk/benc-cgi/watchers.cgi
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE8hrkjsYXoezDwaVARAq20AJ484wxymank0VdYnX83JPIlMqwJBwCfYWiR
M9S2II7fpkGjPi4D7/yqCvw=
=EbrH
-----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Another one...
modprobe psmouse
doesn't trigger a modprobe of i8042
I don't know if you think this should happen in kernel code, or if it
should be in modules.conf...
- --
Ben Clifford [email protected] GPG: 30F06950
Live Ben-cam: http://barbarella.hawaga.org.uk/benc-cgi/watchers.cgi
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org
iD8DBQE8hrqasYXoezDwaVARAmyHAJ0Q5RExa2EGRWTsHo7mt2ZjMVOGqgCcCfwW
B+sskq3bq9/Bmp8FhzlvWm4=
=JAV+
-----END PGP SIGNATURE-----
Ben Clifford wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> Here's another one.
>
> 2.5.5-dj3 removes EXPORT_SYMBOL(ide_fops)
>
> but doesn't remove ide_fops from the code.
>
> Hence modprobe ide-cd doesn't work.
>
Wow! I have found the following in ide-cd.c
devinfo->de = devfs_register(drive->de, "cd", DEVFS_FL_DEFAULT,
HWIF(drive)->major, minor,
S_IFBLK | S_IRUGO | S_IWUGO,
ide_fops, NULL);
So in fact we are devfs_registering the ide-cd rom driver *twice*.
The proper resolution will most propably be to remove this info
there and to use this only in ide.c.
On Wed, Mar 06, 2002 at 04:55:49PM -0800, Ben Clifford wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> Another one...
>
> modprobe psmouse
>
> doesn't trigger a modprobe of i8042
>
> I don't know if you think this should happen in kernel code, or if it
> should be in modules.conf...
They don't depend on each other symbol-wise, and psmouse can work with
other i8042-like interfaces (like ct82c710.o or sun8042.o), so it can't
modprobe that automatically. Add the needed lines to your modules.conf.
--
Vojtech Pavlik
SuSE Labs