]> git.hungrycats.org Git - linux/log
linux
20 years agoLinux 2.6.8-rc2 v2.6.8-rc2
Linus Torvalds [Sat, 17 Jul 2004 14:52:45 +0000 (07:52 -0700)]
Linux 2.6.8-rc2

Ready for the kernel summit in Ottawa...

20 years agoClean up ptrace child exit case.
Linus Torvalds [Sat, 17 Jul 2004 14:51:57 +0000 (07:51 -0700)]
Clean up ptrace child exit case.

This also fixes it for when the real parent is ignoring
SIGCHLD - noted by David Mosberger.

20 years ago[PATCH] Fix i386 bootup with HIGHMEM+SLAB_DEBUG+NUMA and no real
Andi Kleen [Sat, 17 Jul 2004 13:49:01 +0000 (06:49 -0700)]
[PATCH] Fix i386 bootup with HIGHMEM+SLAB_DEBUG+NUMA and no real

For some reason I booted a NUMA and SLAB_DEBUG i386 kernel on a non
NUMA 512MB machine.  This caused an oops at bootup in change_page_attr.
The reason was that highmem_start_start page ended up zero and
that triggered the highmem check in change_page_attr when the
slab debug code would unmap a kernel mapping.

Fix is straightforward: if there is no highmem set highmem_start_page
to max_low_pfn+1

20 years ago[PATCH] Fix memory corruption at x86-64 SMP bootup
Andi Kleen [Sat, 17 Jul 2004 13:48:50 +0000 (06:48 -0700)]
[PATCH] Fix memory corruption at x86-64 SMP bootup

This fixes a long standing corruption bug in the x86-64 code. The SMP
trampoline would corrupt page 2, which was sometimes used for other
data.  This corrupted the ioport/iomem list in some cases and causes
oopses while read /proc/iomem.

Fix is to check the correct location and don't zero it afterwards
because it gets reinitialized for the next CPU anyways.

Thanks to Alexander Nyberg for tracking it down.

20 years ago[PATCH] fix inode state incoherency
Miklos Szeredi [Fri, 16 Jul 2004 11:44:49 +0000 (04:44 -0700)]
[PATCH] fix inode state incoherency

This patch fixes a hard-to-trigger condition, where the inode is on the
inode_in_use list while it's state is dirty.  In this state dirty pages are
not written back in sync() or from kupdate, only from direct page reclaim.
And this causes a livelock in balance_dirty_pages after a while.

The actual sequence of events required to get into this state is:

thread   function                             inode state         inode list
----------------------------------------------------------------------------
1 __sync_single_inode (background)            I_DIRTY             sb->s_io
1 do_writepages ...                           I_LOCKED
2 __writeback_single_inode (sync) sleeps      I_LOCKED
1 __sync_single_inode (background) finish     0                   inode_in_use
2 __writeback_single_inode (sync) wakeup      0
2 __sync_single_inode (sync)                  0
2 do_writepages ...                           I_LOCKED
3 __mark_inode_dirty                          I_LOCKED | I_DIRTY
2 __sync_single_inode (sync) finish           I_DIRTY             left on
                                                                  inode_in_use

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] Altix serial driver updates
Pat Gefre [Fri, 16 Jul 2004 11:44:38 +0000 (04:44 -0700)]
[PATCH] Altix serial driver updates

Patch for our console driver.  We converted the driver to use the serial
core functions.  Also some changes to use sysfs/udev and a new major
number.

Cc: Jesse Barnes <jbarnes@engr.sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] pointer-to-number cast in binfmt_elf.c done right
Alexander Viro [Fri, 16 Jul 2004 05:52:14 +0000 (22:52 -0700)]
[PATCH] pointer-to-number cast in binfmt_elf.c done right

20 years ago[PATCH] more annotations in binfmt_aout.c
Alexander Viro [Fri, 16 Jul 2004 05:52:02 +0000 (22:52 -0700)]
[PATCH] more annotations in binfmt_aout.c

20 years ago[PATCH] hfs and hfsplus switched to use of ffs(3) instead of homegrown versions
Alexander Viro [Fri, 16 Jul 2004 05:51:51 +0000 (22:51 -0700)]
[PATCH] hfs and hfsplus switched to use of ffs(3) instead of homegrown versions

20 years ago[PATCH] fbmem.c partially annotated
Alexander Viro [Fri, 16 Jul 2004 05:51:40 +0000 (22:51 -0700)]
[PATCH] fbmem.c partially annotated

partially annotated, fixed dereferencing of userland pointer (trivial,
since we'd just copied the entire structure).

20 years ago[PATCH] 3w-9xxx.c annotated
Alexander Viro [Fri, 16 Jul 2004 05:51:29 +0000 (22:51 -0700)]
[PATCH] 3w-9xxx.c annotated

20 years agoppc64: More NULL/0 confusion in prom.c
Linus Torvalds [Fri, 16 Jul 2004 05:48:54 +0000 (22:48 -0700)]
ppc64: More NULL/0 confusion in prom.c

20 years agoIt's a pointer, dummy. Use NULL, not 0.
Linus Torvalds [Fri, 16 Jul 2004 05:17:10 +0000 (22:17 -0700)]
It's a pointer, dummy. Use NULL, not 0.

20 years ago[PATCH] annotations and NULL noise removal in drivers/char/drm
Alexander Viro [Fri, 16 Jul 2004 05:05:32 +0000 (22:05 -0700)]
[PATCH] annotations and NULL noise removal in drivers/char/drm

Massive, but trivial - drm annotated and got 0->NULL where needed.  That
patch kills ~2200 lines of warnings - out of 5400 that remained at that
point.

20 years ago[PATCH] sparse: iovec cleanups - the rest
Alexander Viro [Fri, 16 Jul 2004 05:05:19 +0000 (22:05 -0700)]
[PATCH] sparse: iovec cleanups - the rest

the rest of iovec cleanups: nbd, dvb-net, sock.c::sock_no_sendpage(),
econet over udp and ip_vs switched to use of kvec and kernel_...msg().

20 years ago[PATCH] sparse: iovec cleanups - sunrpc, nfs and nfsd
Alexander Viro [Fri, 16 Jul 2004 05:05:10 +0000 (22:05 -0700)]
[PATCH] sparse: iovec cleanups - sunrpc, nfs and nfsd

sunrpc, nfs and nfsd switched to use of kvec and kernel_...msg()

20 years ago[PATCH] sparse: iovec cleanups - rxrpc
Alexander Viro [Fri, 16 Jul 2004 05:04:58 +0000 (22:04 -0700)]
[PATCH] sparse: iovec cleanups - rxrpc

rxrpc (low-level part of afs) switched to kernel_...msg(); it already was
using kvec instead of iovec.

20 years ago[PATCH] sparse: iovec cleanups - cifs
Alexander Viro [Fri, 16 Jul 2004 05:04:47 +0000 (22:04 -0700)]
[PATCH] sparse: iovec cleanups - cifs

cifs switched to kvec and kernel_...msg()

20 years ago[PATCH] sparse: iovec cleanups - ncpfs
Alexander Viro [Fri, 16 Jul 2004 05:04:35 +0000 (22:04 -0700)]
[PATCH] sparse: iovec cleanups - ncpfs

ncpfs switched to kvec and kernel_...msg()

20 years ago[PATCH] sparse: iovec cleanups - smbfs
Alexander Viro [Fri, 16 Jul 2004 05:04:24 +0000 (22:04 -0700)]
[PATCH] sparse: iovec cleanups - smbfs

smbfs switched to kvec and kernel_...msg()

20 years ago[PATCH] sparse: beginning of iovec cleanups - infrastructure
Alexander Viro [Fri, 16 Jul 2004 05:04:13 +0000 (22:04 -0700)]
[PATCH] sparse: beginning of iovec cleanups - infrastructure

Beginning of iovec cleanups - added two helpers (kernel_{send,recv}msg)
that do sock_sendmsg/sock_recvmsg with kvec instead of iovec; basically,
they were abstracted from earlier afs patch.  They take kvec/length of
kvec as separate arguments, do set_fs(), stick kvec into msghdr and call
sock_...msg().

The next group of patches will switch network filesystems to use of kvec
for kernel data + use of these helpers.  Basically, the same thing we'd
done for afs.

20 years ago[PATCH] switch sys32_timer_create() to compat_alloc_user_space()
Alexander Viro [Fri, 16 Jul 2004 05:04:02 +0000 (22:04 -0700)]
[PATCH] switch sys32_timer_create() to compat_alloc_user_space()

switched to compat_alloc_user_space(), cleaned up.

20 years ago[PATCH] misc sparse cleanups
Alexander Viro [Fri, 16 Jul 2004 05:03:51 +0000 (22:03 -0700)]
[PATCH] misc sparse cleanups

 - missing ; between default: and } in sun4setup.c
 - cast of pointer to unsigned long long instead of unsigned long in
   x86_64 signal.c
 - missed annotations for ioctl structure in sparc64 openpromio.h
   (should've been in the same patch as the rest of drivers/sbus/*
   annotations)
 - 0->NULL in list.h and pmdisk.c

20 years ago[PATCH] mark broken stuff as such in Kconfig
Alexander Viro [Fri, 16 Jul 2004 05:03:39 +0000 (22:03 -0700)]
[PATCH] mark broken stuff as such in Kconfig

A bunch of drivers are broken on sparc64; mark them as such in Kconfig.
arlan and smctr are not 64bit-clean; again, markes as such.

20 years ago[PATCH] NULL noise removal in sound/usb/*
Alexander Viro [Fri, 16 Jul 2004 05:03:28 +0000 (22:03 -0700)]
[PATCH] NULL noise removal in sound/usb/*

20 years ago[PATCH] annotated sound/pci/nm256/nm256.c
Alexander Viro [Fri, 16 Jul 2004 05:03:17 +0000 (22:03 -0700)]
[PATCH] annotated sound/pci/nm256/nm256.c

20 years ago[PATCH] ic31712: when storing a bitmask in pointer field, use unsigned long
Alexander Viro [Fri, 16 Jul 2004 05:03:05 +0000 (22:03 -0700)]
[PATCH] ic31712: when storing a bitmask in pointer field, use unsigned long

20 years ago[PATCH] au88x0: use proper field of snd_kcontrol_t and don't try to store pointer...
Alexander Viro [Fri, 16 Jul 2004 05:02:54 +0000 (22:02 -0700)]
[PATCH] au88x0: use proper field of snd_kcontrol_t and don't try to store pointer in int

The au88x0 driver tries to use the wrong private field - there are two,
one of them int and another void *.

While it's not a problem on 32bot boxen, it breaks on 64bit ones.

Fixed.

20 years ago[PATCH] remove bogus casts of pointers to unsigned int in sound/*
Alexander Viro [Fri, 16 Jul 2004 05:02:43 +0000 (22:02 -0700)]
[PATCH] remove bogus casts of pointers to unsigned int in sound/*

20 years ago[PATCH] mmap PROT_NONE fix for NX patch
Daniel McNeil [Fri, 16 Jul 2004 04:52:43 +0000 (21:52 -0700)]
[PATCH] mmap PROT_NONE fix for NX patch

This works around the current PROT_NONE problem from elf binaries that
do not have the PT_GNU_STACK so that the do not have execute permission.

The problem was that setting "def_flags" to include the VM_EXEC bit for
compatibility reasons would also make PROT_NONE pages executable, which
is obviously not correct.

Signed-off-by: Daniel McNeil <daniel@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] Remove all uses of '#ifdef MODULE_PARM' from kernel
Mika Kukkonen [Fri, 16 Jul 2004 04:18:21 +0000 (21:18 -0700)]
[PATCH] Remove all uses of '#ifdef MODULE_PARM' from kernel

I got sidetracked (I do that a lot ;-) and noticed that there are only
eight files in 2.6 that have code like:

#ifdef MODULE_PARAM
MODULE_PARAM(foo, "i");
#endif

which I think is old cruft from 2.1.x times.  Following patch removes
all those eight (six in netfilter, one net driver and one SCSI driver).

20 years ago[PATCH] fix cdrom mt rainier probe
Pat LaVarre [Fri, 16 Jul 2004 04:17:27 +0000 (21:17 -0700)]
[PATCH] fix cdrom mt rainier probe

Writing to RAM capable media was broken by the moving of write flag
detection from probe to drive open time.  The media would stay
terminally RO due to the set_disk_ro() call.

This fixes the code to do the right thing - the detection stays at open
time (which it must, since it depends on the media), and we fail with
-EROFS if the media can't be written.

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years agoMake obsolete NOR flash chip drivers depend on BROKEN.
David Woodhouse [Fri, 16 Jul 2004 17:43:19 +0000 (18:43 +0100)]
Make obsolete NOR flash chip drivers depend on BROKEN.

If nobody shouts soon about why the new, shiny, generic chip drivers
don't work where the old ones do, they can be removed entirely.

20 years agoM-Systems DiskOnChip driver: fix DiskOnChip Millennium ECC support
David Woodhouse [Fri, 16 Jul 2004 17:40:47 +0000 (18:40 +0100)]
M-Systems DiskOnChip driver: fix DiskOnChip Millennium ECC support
and fix a few compiler warnings while we're at it.

20 years agoRemove /proc/fs/jffs2 support.
David Woodhouse [Fri, 16 Jul 2004 17:38:54 +0000 (18:38 +0100)]
Remove /proc/fs/jffs2 support.

It wants reimplementing sanely, preferably in sysfs instead.

20 years ago[PATCH] ppc32: openpic driver cpumask_t changes
David Howells [Thu, 15 Jul 2004 13:09:32 +0000 (06:09 -0700)]
[PATCH] ppc32: openpic driver cpumask_t changes

This fixes the PPC openpic driver to use cpumask_t where appropriate.

Signed-Off-By: David Howells <dhowells@redhat.com>
Signed-Off-By: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ppc32: Fix IRQ setup on WindRiver SBC8560
David Woodhouse [Thu, 15 Jul 2004 13:08:00 +0000 (06:08 -0700)]
[PATCH] ppc32: Fix IRQ setup on WindRiver SBC8560

The internal interrupts are hardcoded to be level-triggered, but we need
to make sure the OpenPIC code knows that...

20 years ago[PATCH] ppc32: Fix UART initialisation on WindRiver SBC8560.
David Woodhouse [Thu, 15 Jul 2004 13:07:49 +0000 (06:07 -0700)]
[PATCH] ppc32: Fix UART initialisation on WindRiver SBC8560.

The UART tristates its IRQ output when the OUT2 bit in the MCR isn't
set. For reasons not known to civilised man, the hardware engineers
stopped it from floating by pulling it _high_ instead of low, so if
anybody requests that IRQ before OUT2 is turned on, we die in an IRQ
storm. The serial8250 driver does precisely that thing, with precisely
that effect.

ALPHA_KLUDGE_MCR was designed to fix exactly this kind of problem, so we
use it here as intended. Ideally ALPHA_KLUDGE_MCR would go away and be
replaced with something like UPF_FORCE_OUT2 and UPF_FORCE_OUT1 flags,
but now is not the time to get me started on a cleanup of everything
about serial8250 that offends me.

20 years ago[PATCH] ppc32: Fix UART detection on WindRiver SBC8560
David Woodhouse [Thu, 15 Jul 2004 13:07:40 +0000 (06:07 -0700)]
[PATCH] ppc32: Fix UART detection on WindRiver SBC8560

The UART is detected as a ST16650V2 but its FIFO doesn't behave nicely.
Disable autoprobe and hard-code it to be a ST16650 instead.

20 years ago[PATCH] tmpfs preempt count panic
Hugh Dickins [Thu, 15 Jul 2004 06:08:02 +0000 (23:08 -0700)]
[PATCH] tmpfs preempt count panic

Just unearthed another of my warcrimes: reading a 17-page sparse file,
I mean holey file, hits the in_interrupt panic in do_exit on a current
highmem kernel (but 2.6.7 is okay).  Fix mismatched preempt count from
shmem_swp_alloc's swapindex hole case by mapping an empty_zero_page.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years agoMerge bk://are.twiddle.net/axp-2.6
Linus Torvalds [Thu, 15 Jul 2004 06:04:46 +0000 (23:04 -0700)]
Merge bk://are.twiddle.net/axp-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

20 years agoMerge linux-mtd@bkbits.net:mtd-2.6
David Woodhouse [Thu, 15 Jul 2004 19:36:34 +0000 (20:36 +0100)]
Merge linux-mtd@bkbits.net:mtd-2.6
into shinybook.infradead.org:/home/dwmw2/bk/mtd-2.6

20 years agoNOR flash drivers update
David Woodhouse [Thu, 15 Jul 2004 19:32:25 +0000 (20:32 +0100)]
NOR flash drivers update

 - Handle cached access to flash chips on supporting platforms
 - Handle arrangements of chips larger than a single bus width
 - Clean up the AMD/Fujitsu chip driver
 - Update board 'mapping' drivers to match
 - New mapping drivers for new platforms.

20 years agoMTD core include and device code cleanup
David Woodhouse [Thu, 15 Jul 2004 19:27:17 +0000 (20:27 +0100)]
MTD core include and device code cleanup

 - Move user-visible bits from headers to include/mtd/ directory.
 - Update old DiskOnChip drivers for newer hardware.
 - Switch NFTL and INFTL support to work with new DiskOnChip/NAND code.
 - New phram driver, reimplenting the ugly slram driver.
 - Bug fixes in partitioning code

20 years agoJFFS2 file system update
David Woodhouse [Thu, 15 Jul 2004 19:19:18 +0000 (20:19 +0100)]
JFFS2 file system update

 - Improved support for NAND flash
 - More generic compression support, allowing for extra compressors
 - Fix potential deadlock with kupdated

20 years agoNAND flash driver updates.
David Woodhouse [Thu, 15 Jul 2004 19:11:45 +0000 (20:11 +0100)]
NAND flash driver updates.

Update the core NAND code:
 - support multiple chips
 - support bad block tables
 - improved generic ECC support and 'spare area' usage.
 - 16-bit NAND
 - Large-block NAND devices
 - Renesas AG-AND devices
 - M-Systems DiskOnChip devices
 - Other new board support wrappers

Most of the work was done by Thomas Gleixner.

Signed-Off-By: David Woodhouse <dwmw2@infradead.org>
20 years agoMerge kanga.twiddle.home:/home/rth/work/linux/linus-2.6
Richard Henderson [Thu, 15 Jul 2004 03:56:19 +0000 (20:56 -0700)]
Merge kanga.twiddle.home:/home/rth/work/linux/linus-2.6
into kanga.twiddle.home:/home/rth/work/linux/axp-2.6

20 years ago[ALPHA] Pass pt_regs as pointer to execve and sigprocmask syscalls.
Richard Henderson [Thu, 15 Jul 2004 03:48:49 +0000 (20:48 -0700)]
[ALPHA] Pass pt_regs as pointer to execve and sigprocmask syscalls.

20 years ago[PATCH] CFQ: allocation under lock, missing memset on allocation
Jens Axboe [Thu, 15 Jul 2004 02:23:57 +0000 (19:23 -0700)]
[PATCH] CFQ: allocation under lock, missing memset on allocation

This fixes two issues in the CFQ IO-scheduler:

o Still a bad allocation under the queue lock
o We need to clear crq after mempool_alloc(), otherwise the rbtree
  pointers can contain garbage if slab poisoning is enabled. This causes
  crashes on front/back merges because rb_next() returns crap.

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years agoMerge bk://kernel.bkbits.net/gregkh/linux/driver-2.6
Linus Torvalds [Wed, 14 Jul 2004 11:03:28 +0000 (04:03 -0700)]
Merge bk://kernel.bkbits.net/gregkh/linux/driver-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

20 years agoMerge bk://kernel.bkbits.net/gregkh/linux/i2c-2.6
Linus Torvalds [Wed, 14 Jul 2004 11:02:30 +0000 (04:02 -0700)]
Merge bk://kernel.bkbits.net/gregkh/linux/i2c-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

20 years agoMerge bk://kernel.bkbits.net/gregkh/linux/usb-2.6
Linus Torvalds [Wed, 14 Jul 2004 11:01:31 +0000 (04:01 -0700)]
Merge bk://kernel.bkbits.net/gregkh/linux/usb-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

20 years ago[PATCH] ide: use "normal" handlers for "flagged" taskfiles (ide-taskfile.c)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:42:30 +0000 (03:42 -0700)]
[PATCH] ide: use "normal" handlers for "flagged" taskfiles (ide-taskfile.c)

This fixes following issues for PIO-in:
 - shared PCI IRQs handling
 - fail request if the last status is bad
and PIO-out:
 - set hwgroup->handler/timer in prehandlers
 - handle drive->unmask in prehandlers
 - check for !rq->nr_sectors and DRQ_STAT bit set
 - use drive->bad_wstat instead of BAD_W_STAT

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: merge CONFIG_IDE_TASKFILE_IO=y|n PIO handlers together
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:42:19 +0000 (03:42 -0700)]
[PATCH] ide: merge CONFIG_IDE_TASKFILE_IO=y|n PIO handlers together

This fixes a couple of CONFIG_IDE_TASKFILE_IO=n issues:
 - check status after last sector for PIO-in transfers
 - handle drive->unmask properly in PIO-out prehandlers
 - use rq->[hard]_nr_sectors where appropriate

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: no partial completions for PIO (CONFIG_IDE_TASKFILE_IO=y)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:42:07 +0000 (03:42 -0700)]
[PATCH] ide: no partial completions for PIO (CONFIG_IDE_TASKFILE_IO=y)

Don't do partial completions but instead acknowledge already transferred
sectors with verified good status on error.  This allows us to complete
"good" sectors to block layer even if bio they belong to wasn't finished
and simplifies code.

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: pre_task_mulout_intr() cleanup (CONFIG_IDE_TASKFILE_IO=n)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:41:55 +0000 (03:41 -0700)]
[PATCH] ide: pre_task_mulout_intr() cleanup (CONFIG_IDE_TASKFILE_IO=n)

drive_is_ready() reads STATUS register if CONFIG_IDEPCI_SHARE_IRQ is not
defined and ALTSTATUS register if it is defined.  Therefore drive_is_ready()
in pre_task_mulout_intr() only makes sense if we can't trust STATUS register
(because we call ide_wait_stat() which reads STATUS register earlier).

Remove this "workaround" for now as it is not present in ide-disk.c
and whole multi PIO-out code (CONFIG_IDE_TASKFILE_IO=n) was buggy before.

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] remove pre_task_out_intr() comment (CONFIG_IDE_TASKFILE_IO=n)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:41:44 +0000 (03:41 -0700)]
[PATCH] remove pre_task_out_intr() comment (CONFIG_IDE_TASKFILE_IO=n)

disable_irq_nosync() in ide-io.c:ide_do_request() protects
pre_task_out_intr() from racing with the task_out_intr().

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: remove BUSY check from task_in_intr() (CONFIG_IDE_TASKFILE_IO=n)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:41:32 +0000 (03:41 -0700)]
[PATCH] ide: remove BUSY check from task_in_intr() (CONFIG_IDE_TASKFILE_IO=n)

We shouldn't ever get there if drive is busy and we can't start transfer
in this case.  ide-disk.c:read_intr() also doesn't check for BUSY_STAT bit.

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: PIO-out error handling fixes (CONFIG_IDE_TASKFILE_IO=y)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:41:21 +0000 (03:41 -0700)]
[PATCH] ide: PIO-out error handling fixes (CONFIG_IDE_TASKFILE_IO=y)

We shouldn't ever get into ->handler() if drive is busy so
just call ->error() unconditionally if status check fails.

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: PIO-out ->prehandler() fixes (CONFIG_IDE_TASKFILE_IO=y)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:41:10 +0000 (03:41 -0700)]
[PATCH] ide: PIO-out ->prehandler() fixes (CONFIG_IDE_TASKFILE_IO=y)

Setup handler and output first data block directly from ->prehandler()
instead of calling ->handler().  The only change in functionality is that
we no longer check DRIVE_READY status bits (there is no need to do it).

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years ago[PATCH] ide: PIO-out fixes for ide-taskfile.c (CONFIG_IDE_TASKFILE_IO=n)
Bartlomiej Zolnierkiewicz [Wed, 14 Jul 2004 10:40:58 +0000 (03:40 -0700)]
[PATCH] ide: PIO-out fixes for ide-taskfile.c (CONFIG_IDE_TASKFILE_IO=n)

 - in task_out_intr() fix off-by-1 bug and (stat & DRQ_STAT) check,
 Â  previously "if" was always true for rq->current_nr_sectors == 1
 - fail request if DRQ_STAT is not set and rq->current_nr_sectors != 0
   (instead of setting handler and waiting for the next IRQ) or if DRQ_STAT
   is set but !rq->current_nr_sectors (in task_mulout_intr() this was OK)
 - in task_mulout_intr() check also DRIVE_READY and WRERR_STAT status bits

Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@elka.pw.edu.pl>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
20 years agoMerge kroah.com:/home/greg/linux/BK/bleed-2.6
Greg Kroah-Hartman [Wed, 14 Jul 2004 10:00:00 +0000 (03:00 -0700)]
Merge kroah.com:/home/greg/linux/BK/bleed-2.6
into kroah.com:/home/greg/linux/BK/driver-2.6

20 years agoMerge kroah.com:/home/greg/linux/BK/bleed-2.6
Greg Kroah-Hartman [Wed, 14 Jul 2004 09:34:20 +0000 (02:34 -0700)]
Merge kroah.com:/home/greg/linux/BK/bleed-2.6
into kroah.com:/home/greg/linux/BK/i2c-2.6

20 years ago[PATCH] add removeable sysfs block device attribute
Olaf Hering [Wed, 14 Jul 2004 09:20:29 +0000 (02:20 -0700)]
[PATCH] add removeable sysfs block device attribute

This patch adds a /block/*/removeable sysfs attribute. A value of 1
indicates the media can change anytime. This is a hint for userland
to poll such devices for possible media changes, and leave all others alone.
There is currently no way to see if a connected usb-storage device is a
disk or a card reader. It will also show 1 for CD and ZIP drives.

It was done by Patrick Mansfield a while ago. I can probably not
sigh-off his work. ;)

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] Upgrade security/root_plug.c to new module parameter syntax
Mika Kukkonen [Wed, 14 Jul 2004 09:19:51 +0000 (02:19 -0700)]
[PATCH] Upgrade security/root_plug.c to new module parameter syntax

Hi again,

Still doing my compile, and got this:

 CC [M]  security/root_plug.o
security/root_plug.c:39: warning: missing initializer
security/root_plug.c:39: warning: (near initialization for `__parm_vendor_id.addr')
security/root_plug.c:42: warning: missing initializer
security/root_plug.c:42: warning: (near initialization for `__parm_product_id.addr')
security/root_plug.c:48: warning: missing initializer
security/root_plug.c:48: warning: (near initialization for `__parm_debug.addr')

Simply upgrading root_plug to use the new module parameter syntax seemed to do
the trick. I made the debug writable, the others just readable to root.

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] Driver core: Fix OOPS in device_platform_unregister
Dmitry Torokhov [Wed, 14 Jul 2004 09:09:44 +0000 (02:09 -0700)]
[PATCH] Driver core: Fix OOPS in device_platform_unregister

Driver core: platform_device_unregister should release resources first
             and only then call device_unregister, otherwise if there
             are no more references to the device it will be freed and
             the fucntion will try to access freed memory.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] I2C: patch quirks.c - SMBus hidden on hp laptop
Örjan Persson [Wed, 14 Jul 2004 09:04:27 +0000 (02:04 -0700)]
[PATCH] I2C: patch quirks.c - SMBus hidden on hp laptop

This patch unhides the SMBus on the hp nc8000 and nc6000 laptops. The
patch has been co-written by Jean Delvare and Rudolf Marek. I've only
tested this on nc8000, but it should work for the nc6000 too.

Unfortunatley, we had to little information to fix the problem described
in the reported bug below, as is probably the same problem. But if we're
very lucky it might solve it too.
http://bugzilla.kernel.org/show_bug.cgi?id=2976

Signed-off-by: Örjan Persson <orange@fobie.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years agomerge fixups due to sparse cleanups being done by al and greg at the same time.
Greg Kroah-Hartman [Wed, 14 Jul 2004 08:44:39 +0000 (01:44 -0700)]
merge fixups due to sparse cleanups being done by al and greg at the same time.

20 years agoMerge bk://kernel.bkbits.net/jgarzik/libata-upstream-2.6
Linus Torvalds [Wed, 14 Jul 2004 08:10:06 +0000 (01:10 -0700)]
Merge bk://kernel.bkbits.net/jgarzik/libata-upstream-2.6
into ppc970.osdl.org:/home/torvalds/v2.6/linux

20 years ago[PCI, libata] Fix "combined mode" PCI quirk for ICH6
Jeff Garzik [Wed, 14 Jul 2004 12:20:05 +0000 (08:20 -0400)]
[PCI, libata] Fix "combined mode" PCI quirk for ICH6

The hardware vendor, in their infinite wisdom, make the combined
mode configuration register different between ICH5 and ICH6.

Take this into account.

20 years ago[PATCH] libata: fix kunmap() of incorrect page, in PIO data xfer
Sergio Gelato [Wed, 14 Jul 2004 12:11:41 +0000 (08:11 -0400)]
[PATCH] libata: fix kunmap() of incorrect page, in PIO data xfer

Obvious bug.  Fixes highmem oops.

20 years ago[PATCH] USB: USB w9968cf compile error
Adrian Bunk [Wed, 14 Jul 2004 08:08:37 +0000 (01:08 -0700)]
[PATCH] USB:  USB w9968cf compile error

On Tue, Jul 13, 2004 at 06:25:59PM -0700, Andrew Morton wrote:
>...
> All 252 patches:
>...
> bk-usb.patch
>...

This patch marks w9968cf_valid_depth as inline, although it's used
before it's defined.

gcc 3.4 therefore correctly fails with:

<--  snip  -->

...
  CC      drivers/usb/media/w9968cf.o
drivers/usb/media/w9968cf.c: In function `w9968cf_set_picture':
drivers/usb/media/w9968cf.c:487: sorry, unimplemented: inlining failed
in call to 'w9968cf_valid_depth': function body not available
drivers/usb/media/w9968cf.c:1722: sorry, unimplemented: called from here
make[3]: *** [drivers/usb/media/w9968cf.o] Error 1

<--  snip  -->

This patch moves w9968cf_valid_depth above it's first user (it also uses
two other functions to keep the ordering of functions a bit more
consistent).

Signed-off-by: Adrian Bunk <bunk@fs.tum.de>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb hub, don't check speed before reset
David Brownell [Wed, 14 Jul 2004 08:08:21 +0000 (01:08 -0700)]
[PATCH] USB: usb hub, don't check speed before reset

Signaling (even D- vs D+ pullup) may not be stable
for a while.

At least one OTG root hub won't reliably (<10%) report device speed until
after reset (and the OTG state transitions have quiesced for a while).

This patch makes that not matter.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: W99[87]CF fix
Luca Risolia [Wed, 14 Jul 2004 08:08:05 +0000 (01:08 -0700)]
[PATCH] USB: W99[87]CF fix

Signed-off-by: Luca Risolia <luca.risolia@studio.unibo.it>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb host side updates, mostly for suspend
David Brownell [Wed, 14 Jul 2004 08:07:48 +0000 (01:07 -0700)]
[PATCH] USB: usb host side updates, mostly for suspend

This adds some of the infrastructure needed to support some more
USB capabilities:

 -  CONFIG_USB_SUSPEND, so Linux can put individual devices
    into the USB "suspend" state.  They can (sometimes) use
    "remote wakeup" to resume the host; or they can each be
    resumed by the host.

      + New usbcore device selective suspend/resume APIs
* Define them, as stubs for now
* Call them on the paths sysfs uses (renamed functions)
      + HCD support
* Define root hub suspend calls; delegate them to HCDs.
* OHCI and EHCI can suspend/resume root hubs that way.
* Not called yet, until suspend/resume calls exist

 -  CONFIG_USB_OTG, which depends on the selective suspend APIs
    to allow devices to switch roles (host to peripheral, etc).
    This patch just adds a few key flags in usb_bus, needed by
    usbcore (during enumeration) and by HCD and OTG controllers
    on OTG-capable boards.

 -  Related bugfix:  power budgeting is supposed to place a
    100mA per port (non-OTG) for bus-powered devices.

This patch changes no behavior; later patches will do that,
and they'll be smaller because of this.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: ftdi_sio VID/PID updates
Ian Abbott [Wed, 14 Jul 2004 08:07:29 +0000 (01:07 -0700)]
[PATCH] USB: ftdi_sio VID/PID updates

Greg,

This patch adds various VID/PIDs to the ftdi_sio driver that I've
accumulated from the ftdi-usb-sio-devel mailing list, and one from
the 2.4 kernel sources:

* Interbiometrics USB I/O boards VID/PIDs from 2.4 kernel (Rudolf
  Gugler).
* Intrepid Control Systems ValueCAN and NeoVI VID/PIDs from Scott
  Wolchok.
* Falcom Twist USB GPRS modem VID/PID from Justin Schoeman.
* Suunto Sports instrument PID from Panu Kekalainen.

Also a few minor clean-ups:

* Removed a stray PID macro constant FTDI_GUDEADS_889_PID.
* Copied the Inside Accesso VID/PID into the FT232BM ID table so
  that the "combined" table is a proper union of the others (unless I've
  missed something else).
* Added a comment that the OCT US101 is also rebadged as a SIIG Inc.
  US2308.

Signed-off-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: ftdi_sio debug trace for TIOCMSET
Ian Abbott [Wed, 14 Jul 2004 08:07:09 +0000 (01:07 -0700)]
[PATCH] USB: ftdi_sio debug trace for TIOCMSET

Someone spotted that ftdi_sio outputs a debug trace for TIOCMGET,
but not for TIOCMSET, so this patch adds a debug trace for TIOCMSET
too.

Signed off by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb/core/hcd.c::usb_init() missing audit.
Luiz Capitulino [Wed, 14 Jul 2004 08:06:51 +0000 (01:06 -0700)]
[PATCH] USB: usb/core/hcd.c::usb_init() missing audit.

Signed-off-by: Luiz Capitulino <lcapitulino@prefeitura.sp.gov.br>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usbserial/ipaq update
Andreas John [Wed, 14 Jul 2004 08:06:33 +0000 (01:06 -0700)]
[PATCH] USB: usbserial/ipaq update

However, attached I provide you a patch against 2.6.7 which should add
the ASUS A620 PDA to the ipaq Kernel module.

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: unusual_devs.h update
Frank Engel [Wed, 14 Jul 2004 08:06:15 +0000 (01:06 -0700)]
[PATCH] USB: unusual_devs.h update

added another device to unusual_devs.h. Its a Sony Clie PEG-675C. Now I
can mount the memory stick.

cat /proc/bus/usb/devices

T:  Bus=03 Lev=01 Prnt=01 Port=01 Cnt=01 Dev#= 18 Spd=12  MxCh= 0
D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=16 #Cfgs=  1
P:  Vendor=054c ProdID=0099 Rev= 1.00
S:  Manufacturer=Sony
S:  Product=Sony PEG Mass Storage
C:* #Ifs= 1 Cfg#= 1 Atr=c0 MxPwr=  6mA
I:  If#= 0 Alt= 0 #EPs= 3 Cls=08(stor.) Sub=05 Prot=00 Driver=usb-storage
E:  Ad=81(I) Atr=03(Int.) MxPS=  64 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: fix SN9C10[12] driver for ia64
Olaf Hering [Wed, 14 Jul 2004 08:02:24 +0000 (01:02 -0700)]
[PATCH] USB: fix SN9C10[12] driver for ia64

 On Sun, Jul 11, Luca Risolia wrote:

> This single patch contains some updates and cleanups for
> the W996[87]CF driver and a new experimental V4L2 driver
> for SONiX SN9C10[12] PC Camera Controllers connected to various
> image sensors. I have not divided the patch in two logical
> sub-patches becouse of two independent changes in one common
> file, KConfigure. Since this mailing list refused the patch,
> due to its size, I have uploaded it here, as
> Documentation/SubmittingPatches suggests:

I need this patch for x86_64.

  CC [M]  drivers/usb/media/sn9c102_core.o
In file included from drivers/usb/media/sn9c102.h:34,
                 from drivers/usb/media/sn9c102_core.c:43:
include/asm/rwsem.h:55: error: redefinition of `struct rw_semaphore'
In file included from drivers/usb/media/sn9c102.h:34,
                 from drivers/usb/media/sn9c102_core.c:43:
include/asm/rwsem.h:79:1: warning: "__RWSEM_INITIALIZER" redefined
In file included from include/linux/rwsem.h:25,
                 from include/asm/semaphore.h:43,
                 from include/linux/sched.h:18,
                 from include/linux/module.h:10,
                 from drivers/usb/media/sn9c102_core.c:21:
include/linux/rwsem-spinlock.h:49:1: warning: this is the location of the previous definition
include/asm/rwsem.h:87: warning: static declaration for `init_rwsem' follows non-static
include/asm/rwsem.h:100: warning: `__down_read' declared inline after being called
include/asm/rwsem.h:100: warning: static declaration for `__down_read' follows non-static
include/asm/rwsem.h:122: warning: `__down_read_trylock' declared inline after being called
include/asm/rwsem.h:122: warning: static declaration for `__down_read_trylock' follows non-static
include/asm/rwsem.h:146: warning: `__down_write' declared inline after being called
include/asm/rwsem.h:146: warning: static declaration for `__down_write' follows non-static
include/asm/rwsem.h:171: warning: `__down_write_trylock' declared inline after being called
include/asm/rwsem.h:171: warning: static declaration for `__down_write_trylock' follows non-static
include/asm/rwsem.h:184: warning: `__up_read' declared inline after being called
include/asm/rwsem.h:184: warning: static declaration for `__up_read' follows non-static
include/asm/rwsem.h:208: warning: `__up_write' declared inline after being called
include/asm/rwsem.h:208: warning: static declaration for `__up_write' follows non-static
include/asm/rwsem.h:233: warning: `__downgrade_write' declared inline after being called
include/asm/rwsem.h:233: warning: static declaration for `__downgrade_write' follows non-static
make[3]: *** [drivers/usb/media/sn9c102_core.o] Error 1

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: sort the order in which the usb-serial drivers get built
Greg Kroah-Hartman [Wed, 14 Jul 2004 08:02:08 +0000 (01:02 -0700)]
[PATCH] USB: sort the order in which the usb-serial drivers get built

all other usb drivers get built in alphabetical order, why not these?

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: remove CONFIG_USB_SERIAL_DEBUG
Greg Kroah-Hartman [Wed, 14 Jul 2004 08:01:49 +0000 (01:01 -0700)]
[PATCH] USB: remove CONFIG_USB_SERIAL_DEBUG

This involved reworking the usb_serial_debug_data() function too.
Based on a request from SuSE, and numerous user confusions with how
to enable this option.  Now that all usb-serial drivers are using
module_param() it can be set from the boot command line if the drivers
are built into the kernel.

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: change all usbserial drivers to use module_param()
Greg Kroah-Hartman [Wed, 14 Jul 2004 08:01:31 +0000 (01:01 -0700)]
[PATCH] USB: change all usbserial drivers to use module_param()

Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] Updates for W99[87]CF and new SN9C10[12] driver
Luca Risolia [Wed, 14 Jul 2004 08:01:10 +0000 (01:01 -0700)]
[PATCH] Updates for W99[87]CF and new SN9C10[12] driver

This single patch contains some updates and cleanups for
the W996[87]CF driver and a new experimental V4L2 driver
for SONiX SN9C10[12] PC Camera Controllers connected to various
image sensors. I have not divided the patch in two logical
sub-patches becouse of two independent changes in one common
file, KConfigure.

More informations about the SN9C10[12] can be found below in the
documentation. The driver is marked as "EXPERIMENTAL", meaning
that there are no known bugs, but further testing is necessary
before considering it stable. This the first driver using the new
SBGGR8 video format, which has been recently added to the mainline
kernel, so there are no available user application at the moment:
this is one more reason why it should be in the kernel now.

Changes in W996[87]CF:
- remove w9968cf_externaldef.h now that ovcamchip.h is in the kernel;
- mark user pointers with __user in a cleaner way to avoid sparse
  warnings;
- use appropriate exclusive wait macro during open();
- replace info(), err(), warn() with dev_info(), dev_err(), dev_warn(),
  pr_debug(), pr_info();
- replace usb_unlink_urb() + wait_for_completion() with usb_kill_urb();
- fix memory offsets for buffers in the chip to be used with generic
  image sensors;
- 'vppmod_load', 'debug', 'specific_debug' and 'simcams' module
  parameters are now writeable by default;
- fix possible race conditions between disconnect() and open();
- add automatic 'ovcamchip' module loading option with 'ovmod_load'
  module parameter;
- get rid of deprecated intermodule communication routines and use the
  correct module registration/unregistration approach;
- remove period at the end of kernel messages;
- fix several typos;
- use MODULE_VERSION() macro;
- other small internal cleanups;
- documentation updates.

Signed-off-by: Luca Risolia <luca.risolia@studio.unibo.it>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb ethernet gadget, minor fixes + basic OTG support
David Brownell [Wed, 14 Jul 2004 08:00:36 +0000 (01:00 -0700)]
[PATCH] USB: usb ethernet gadget, minor fixes + basic OTG support

Update CDC Ethernet/RNDIS gadget driver to the latest:

    - Basics of OTG support: providing the OTG descriptor
in each configuration (as needed).   No HNP yet.

    - Stop issuing partial-packet reads.  There's some hardware that
only counts reads in packets, not bytes, so let's not bother.
There are still software checks to catch framing gone wild.

    - Fix a small bug that crept in with a memory leak fix: after
        RNDIS requests, ep0 responses would use the RNDIS completion
handler even for non-RNDIS requests.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb gadget zero, basic OTG updates
David Brownell [Wed, 14 Jul 2004 08:00:19 +0000 (01:00 -0700)]
[PATCH] USB: usb gadget zero, basic OTG updates

This patch teaches "gadget zero" enough about OTG to pass simple USBCV
tests, mostly by including OTG descriptors in each configuration.  It
tests and reports OTG status, as reported by the USB controller driver.

It also adds an option to build gadget zero to act as the designated
OTG "HNP Test Device", which exists primarily to trigger HNP.  However,
it won't currently request HNP.

Includes other minor tweaks:  delete a timer on disconnect, reset
the req->zero flag, don't autoresume after disconnect.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb gadget API updates
David Brownell [Wed, 14 Jul 2004 08:00:02 +0000 (01:00 -0700)]
[PATCH] USB: usb gadget API updates

Gadget API updates, including new features:

    - Adds several new optional calls that can be made to the
USB peripheral controller:

    * VBUS session reporting, for use mostly by external
      transcievers (such as isp1301).  Detection of VBUS
      power is the first step in enumeration, and usually
      corresponds to a device being plugged into a hub.

    * "Soft Connect" feature, wherein the D+ pullup is under
      gadget driver control.  This is a second step during
      enumeration, which lets the hub see the new device.
      (Based on a patch by Alex Sanks <alex@netchip.com>.)

    * Control over VBUS current draw.  So for example this is
      what a gadget driver uses during SET_CONFIGURATION to
      say "it's OK to draw 300 mA from VBUS to recharge".

    - Basic interfaces to support device-side USB OTG.  Feature
flags and descriptors are in <linux/usb_ch9.h> already.

    * Reports whether the device has a Mini-AB port, so
      that OTG support (desriptors etc) is required.

    * Reports OTG device feature flags for HNP.  (The OTG
      spec requires user interfaces to report this stuff.)
      Say if this is a B-Peripheral or an A-Peripheral.

    * Says that usb_gadget_wakeup() is how to access SRP,
      and usb_gadget_disconnect() is how to access HNP.

    - Minor updates/cleanups to comments (Linux 2.5-->2.6 etc)

Currently there's no support for SRP-only OTG devices.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb gadgetfs, handle omap_udc
David Brownell [Wed, 14 Jul 2004 07:59:45 +0000 (00:59 -0700)]
[PATCH] USB: usb gadgetfs, handle omap_udc

Tell gadgetfs about omap_udc.
Add some missing __user annotations.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years ago[PATCH] USB: usb serial gadget, add omap_udc
David Brownell [Wed, 14 Jul 2004 07:59:25 +0000 (00:59 -0700)]
[PATCH] USB: usb serial gadget, add omap_udc

This lets the serial gadget work with another controller.

Tell serial about omap_udc.  This driver still needs updating to
use the endpoint autoconfig suppport; like Gadget Zero, it
really shouldn't need _any_ hardware-specific #ifdeffery.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>
20 years agoMerge http://jfs.bkbits.net/linux-2.5
Linus Torvalds [Wed, 14 Jul 2004 07:53:40 +0000 (00:53 -0700)]
Merge http://jfs.bkbits.net/linux-2.5
into ppc970.osdl.org:/home/torvalds/v2.6/linux

20 years agoMerge pobox.com:/spare/repo/linux-2.6
Jeff Garzik [Wed, 14 Jul 2004 11:24:08 +0000 (07:24 -0400)]
Merge pobox.com:/spare/repo/linux-2.6
into pobox.com:/spare/repo/net-drivers-2.6

20 years agoMerge pobox.com:/spare/repo/netdev-2.6/misc
Jeff Garzik [Wed, 14 Jul 2004 11:19:30 +0000 (07:19 -0400)]
Merge pobox.com:/spare/repo/netdev-2.6/misc
into pobox.com:/spare/repo/net-drivers-2.6

20 years ago[netdrvr acenic] fix RX descriptor memory ordering
Janice M. Girouard [Wed, 14 Jul 2004 11:13:01 +0000 (07:13 -0400)]
[netdrvr acenic] fix RX descriptor memory ordering

20 years ago[wireless airo] fix alignment problem (particularly on ARM)
Guennadi Liakhovetski [Wed, 14 Jul 2004 11:06:52 +0000 (07:06 -0400)]
[wireless airo] fix alignment problem (particularly on ARM)

Bugzilla #2905

20 years ago[netdrvr tg3] bump version and reldate
Jeff Garzik [Wed, 14 Jul 2004 11:01:55 +0000 (07:01 -0400)]
[netdrvr tg3] bump version and reldate

20 years ago[PATCH] tg3 bug
Anton Blanchard [Wed, 14 Jul 2004 11:00:50 +0000 (07:00 -0400)]
[PATCH] tg3 bug

During receive processing, the tg3 card updates rx_producer (the
hardware position in the receive ring) and the opaque cookie. Due to PCI
rules the stores happen in order. However the cpu may reorder the reads.

In these sort of cases there is usually a data dependency between
reading the index and looking up the data (since we use the index to
load the data). I think all cpus except alpha guarantee the reads
happen in order in this case.

However in this particular case we load hw_idx, compare it to sw_idx and
then use sw_idx to locate the opaque cookie. There is no data dependency
in this case. We need a read memory barrier between the read of
rx_producer and the opaque cookie to enforce ordering.

Thanks to Olof Johansson, Michael Chan and Broadcom for their assistance
in finding this bug.

Signed-off-by: Anton Blanchard <anton@samba.org>
20 years ago[PATCH] fix airo oops-on-removal
Andrew Morton [Wed, 14 Jul 2004 10:42:35 +0000 (06:42 -0400)]
[PATCH] fix airo oops-on-removal

From: Bill Nottingham <notting@redhat.com>

airo creates /proc/driver/aironet/<device name> on device activation.
However, the device can be renamed - then on teardown it tries to remove
the wrong directory.  The removal of /proc/driver/aironet then runs afoul
of the BUG_ON() in remove_proc_entry.

This fixes it by keeping a copy of the name of the directory it created.

(It doesn't actually solve the problem of the stats directory still being
/proc/driver/aironet/eth0 when you rename the device to, say, 'joe'.  But
that patch would be a little less trivial.)

Signed-off-by: Andrew Morton <akpm@osdl.org>
20 years ago[PATCH] prism54 Fix wrong type for BSSID
Margit Schubert-While [Wed, 14 Jul 2004 10:42:25 +0000 (06:42 -0400)]
[PATCH] prism54 Fix wrong type for BSSID

2004-07-12 Margit Schubert-While <margitsw@t-online.de>

* The OID type for BSSID was incorrectly set to type SSID.
  It should be type RAW. This lead to interesting reporting
  by "iwpriv ethX g_bssid".
  (Which caused garbage output and possibly an out of bound)

* Be ultra-cautious in reporting SSID by changing the "%s"
  to "%.*s" and passing the length. (Prompted by the false
  type above, whereby length = 0 and a %s on a garbage field)

20 years agoMerge pobox.com:/spare/repo/netdev-2.6/sis900
Jeff Garzik [Wed, 14 Jul 2004 10:33:02 +0000 (06:33 -0400)]
Merge pobox.com:/spare/repo/netdev-2.6/sis900
into pobox.com:/spare/repo/net-drivers-2.6