openwrt/staging/blogic.git
15 years agoide-tape: unify r/w init paths
Tejun Heo [Sat, 18 Apr 2009 23:46:02 +0000 (08:46 +0900)]
ide-tape: unify r/w init paths

Impact: cleanup

Read and write init paths are almost identical.  Unify them into
idetape_init_rw().

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-tape: kill idetape_bh
Tejun Heo [Sat, 18 Apr 2009 23:46:02 +0000 (08:46 +0900)]
ide-tape: kill idetape_bh

Impact: kill now unnecessary idetape_bh

With everything using standard mechanisms, there is no need for
idetape_bh anymore.  Kill it and use tape->buf, cur and valid to
describe data buffer instead.

Changes worth mentioning are...

* idetape_queue_rq_tail() now always queue tape->buf and and adjusts
  buffer state properly before completion.

* idetape_pad_zeros() clears the buffer only once.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-tape: use standard data transfer mechanism
Tejun Heo [Sat, 18 Apr 2009 23:46:02 +0000 (08:46 +0900)]
ide-tape: use standard data transfer mechanism

Impact: use standard way to transfer data

ide-tape uses rq in an interesting way.  For r/w requests, rq->special
is used to carry a private buffer management structure idetape_bh and
rq->nr_sectors and current_nr_sectors are initialized to the number of
idetape blocks which isn't necessary 512 bytes.  Also,
rq->current_nr_sectors is used to report back the residual count in
units of idetape blocks.

This peculiarity taxes both block layer and ide.  ide-atapi has
different paths and hooks to accomodate it and what a rq means becomes
quite confusing and making changes at the block layer becomes quite
difficult and error-prone.

This patch makes ide-tape use bio instead.  With the previous patch,
ide-tape currently is using single contiguos buffer so replacing it
isn't difficult.  Data buffer is mapped into bio using
blk_rq_map_kern() in idetape_queue_rw_tail().  idetape_io_buffers()
and idetape_update_buffers() are dropped and pc->bh is set to null to
tell ide-atapi to use standard data transfer mechanism and idetape_bh
byte counts are updated by the issuer on completion using the residual
count.

This change also nicely removes the FIXME in ide_pc_intr() where
ide-tape rqs need to be completed using ide_rq_bytes() instead of
blk_rq_bytes() (although this didn't really matter as the request
didn't have bio).

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <jens.axboe@oracle.com>
15 years agoide-tape: use single continuous buffer
Tejun Heo [Sat, 18 Apr 2009 23:46:02 +0000 (08:46 +0900)]
ide-tape: use single continuous buffer

Impact: simpler buffer allocation and handling, kills OOM, fix DMA transfers

ide-tape has its own multiple buffer mechanism using struct
idetape_bh.  It allocates buffer with decreasing order-of-two
allocations so that it results in minimum number of segments.
However, the implementation is quite complex and works in a way that
no other block or ide driver works necessitating a lot of special case
handling.

The benefit this complex allocation scheme brings is questionable as
PIO or DMA the number of segments (16 maximum) doesn't make any
noticeable difference and it also doesn't negate the need for multiple
order allocation which can fail under memory pressure or high
fragmentation although it does lower the highest order necessary by
one when the buffer size isn't power of two.

As the first step to remove the custom buffer management, this patch
makes ide-tape allocate single continous buffer.  The maximum order is
four.  I doubt the change would cause any trouble but if it ever
matters, it should be converted to regular sg mechanism like everyone
else and even in that case dropping custom buffer handling and moving
to standard mechanism first make sense as an intermediate step.

This patch makes the first bh to contain the whole buffer and drops
multi bh handling code.  Following patches will make further changes.

This patch has the side effect of killing OOM triggered by allocation
path and fixing DMA transfers.  Previously, bug in alloc path
triggered OOM on command issue and commands were passed to DMA engine
without DMA-mapping all the segments.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-atapi,tape,floppy: allow ->pc_callback() to change rq->data_len
Tejun Heo [Sat, 18 Apr 2009 23:46:02 +0000 (08:46 +0900)]
ide-atapi,tape,floppy: allow ->pc_callback() to change rq->data_len

Impact: allow residual count implementation in ->pc_callback()

rq->data_len has two duties - carrying the number of input bytes on
issue and carrying residual count back to the issuer on completion.
ide-atapi completion callback ->pc_callback() is the right place to do
this but currently ide-atapi depends on rq->data_len carrying the
original request size after calling ->pc_callback() to complete the pc
request.

This patch makes ide_pc_intr(), ide_tape_issue_pc() and
ide_floppy_issue_pc() cache length to complete before calling
->pc_callback() so that it can modify rq->data_len as necessary.

Note: As using rq->data_len for two purposes can make cases like this
      incorrect in subtle ways, future changes will introduce separate
      field for residual count.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <jens.axboe@oracle.com>
15 years agoide-tape,floppy: fix failed command completion after request sense
Tejun Heo [Sat, 18 Apr 2009 23:46:02 +0000 (08:46 +0900)]
ide-tape,floppy: fix failed command completion after request sense

Impact: fix infinite retry loop

After a command failed, ide-tape and floppy inserts REQUEST_SENSE in
front of the failed command and according to the result, sets
pc->retries, flags and errors.  After REQUEST_SENSE is complete, the
failed command is again at the front of the queue and if the verdict
was to terminate the request, the issue functions tries to complete it
directly by calling drive->pc_callback() and returning ide_stopped.

However, drive->pc_callback() doesn't complete a request.  It only
prepares for completion of the request.  As a result, this creates an
infinite loop where the failed request is retried perpetually.

Fix it by actually ending the request by calling ide_complete_rq().

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-pm: don't abuse rq->data
Tejun Heo [Sat, 18 Apr 2009 22:00:43 +0000 (07:00 +0900)]
ide-pm: don't abuse rq->data

Impact: cleanup rq->data usage

ide-pm uses rq->data to carry pointer to struct request_pm_state
through request queue and rq->special is used to carray pointer to
local struct ide_cmd, which isn't necessary.  Use rq->special for
request_pm_state instead and use local ide_cmd in
ide_start_power_step().

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
15 years agoide-cd,atapi: use bio for internal commands
Tejun Heo [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-cd,atapi: use bio for internal commands

Impact: unify request data buffer handling

rq->data is used mostly to pass kernel buffer through request queue
without using bio.  There are only a couple of places which still do
this in kernel and converting to bio isn't difficult.

This patch converts ide-cd and atapi to use bio instead of rq->data
for request sense and internal pc commands.  With previous change to
unify sense request handling, this is relatively easily achieved by
adding blk_rq_map_kern() during sense_rq prep and PC issue.

If blk_rq_map_kern() fails for sense, the error is deferred till sense
issue and aborts the failed command which triggered the sense.  Note
that this is a slim possibility as sense prep is done on each command
issue, so for the above condition to actually trigger, all preps since
the last sense issue till the issue of the request which would require
a sense should fail.

* do_request functions might sleep now.  This should be okay as ide
  request_fn - do_ide_request() - is invoked only from make_request
  and plug work.  Make sure this is the case by adding might_sleep()
  to do_ide_request().

* Functions which access the read sense data before the sense request
  is complete now should access bio_data(sense_rq->bio) as the sense
  buffer might have been copied during blk_rq_map_kern().

* ide-tape updated to map sg.

* cdrom_do_block_pc() now doesn't have to deal with REQ_TYPE_ATA_PC
  special case.  Simplified.

* tp_ops->output/input_data path dropped from ide_pc_intr().

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-atapi: convert ide-{floppy,tape} to using preallocated sense buffer
Borislav Petkov [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-atapi: convert ide-{floppy,tape} to using preallocated sense buffer

Since we're issuing REQ_TYPE_SENSE now we need to allow those types of
rqs in the ->do_request callbacks. As a future improvement, sense_len
assignment might be unified across all ATAPI devices. Borislav to
check with specs and test.

As a result, get rid of ide_queue_pc_head() and
drive->request_sense_rq.

tj: * Init request sense ide_atapi_pc from sense request.  In the
      longer timer, it would probably better to fold
      ide_create_request_sense_cmd() into its only current user -
      ide_floppy_get_format_progress().

    * ide_retry_pc() no longer takes @disk.

CC: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
CC: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-cd: convert to using generic sense request
Borislav Petkov [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-cd: convert to using generic sense request

Preallocate a sense request in the ->do_request method and reinitialize
it only on demand, in case it's been consumed in the IRQ handler path.
The reason for this is that we don't want to be mapping rq to bio in
the IRQ path and introduce all kinds of unnecessary hacks to the block
layer.

tj: * Both user and kernel PC requests expect sense data to be stored
      in separate storage other than drive->sense_data.  Copy sense
      data to rq->sense on completion if rq->sense is not NULL.  This
      fixes bogus sense data on PC requests.

As a result, remove cdrom_queue_request_sense.

CC: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
CC: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide: add helpers for preparing sense requests
Borislav Petkov [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide: add helpers for preparing sense requests

This is in preparation of removing the queueing of a sense request out
of the IRQ handler path.

Use struct request_sense as a general sense buffer for all ATAPI
devices ide-{floppy,tape,cd}.

tj: * blk_get_request(__GFP_WAIT) can't be called from do_request() as
      it can cause deadlock.  Converted to use inline struct request
      and blk_rq_init().

    * Added xfer / cdb len selection depending on device type.

    * All sense prep logics folded into ide_prep_sense() which never
      fails.

    * hwif->rq clearing and sense_rq used handling moved into
      ide_queue_sense_rq().

    * blk_rq_map_kern() conversion is moved to later patch.

CC: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
CC: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Borislav Petkov <petkovbb@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-cd: don't abuse rq->buffer
Tejun Heo [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-cd: don't abuse rq->buffer

Impact: rq->buffer usage cleanup

ide-cd uses rq->buffer to carry pointer to the original request when
issuing REQUEST_SENSE.  Use rq->special instead.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
15 years agoide-atapi: don't abuse rq->buffer
Tejun Heo [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-atapi: don't abuse rq->buffer

Impact: rq->buffer usage cleanup

ide-atapi uses rq->buffer as private opaque value for internal special
requests.  rq->special isn't used for these cases (the only case where
rq->special is used is for ide-tape rw requests).  Use rq->special
instead.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
15 years agoide-taskfile: don't abuse rq->buffer
Tejun Heo [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-taskfile: don't abuse rq->buffer

Impact: rq->buffer usage cleanup

ide_raw_taskfile() directly uses rq->buffer to carry pointer to the
data buffer.  This complicates both block interface and ide backend
request handling.  Use blk_rq_map_kern() instead and drop special
handling for REQ_TYPE_ATA_TASKFILE from ide_map_sg().

Note that REQ_RW setting is moved upwards as blk_rq_map_kern() uses it
to initialize bio rw flag.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
15 years agoide-floppy: block pc always uses bio
Tejun Heo [Sat, 18 Apr 2009 22:00:42 +0000 (07:00 +0900)]
ide-floppy: block pc always uses bio

Impact: remove unnecessary code path

Block pc requests always use bio and rq->data is always NULL.  No need
to worry about !rq->bio cases in idefloppy_block_pc_cmd().  Note that
ide-atapi uses ide_pio_bytes() for bio PIO transfer which handle sg
fine.

Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Jens Axboe <axboe@kernel.dk>
15 years agoide-cd: clear sense buffer before issuing request sense
Tejun Heo [Sat, 18 Apr 2009 22:00:41 +0000 (07:00 +0900)]
ide-cd: clear sense buffer before issuing request sense

Impact: code simplification

ide_cd_request_sense_fixup() clears the tail of the sense buffer if
the device didn't completely fill it.  This patch makes
cdrom_queue_request_sense() clear the sense buffer before issuing the
command instead of clearing it afterwards.  This simplifies code and
eases future changes.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide kill unused ide_cmd->special
Tejun Heo [Sat, 18 Apr 2009 22:00:41 +0000 (07:00 +0900)]
ide kill unused ide_cmd->special

Impact: removal of unused field

No one uses ide_cmd->special anymore.  Kill it.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide: don't set REQ_SOFTBARRIER
Tejun Heo [Sat, 18 Apr 2009 22:00:41 +0000 (07:00 +0900)]
ide: don't set REQ_SOFTBARRIER

ide doesn't have to worry about REQ_SOFTBARRIER.  Don't set it.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide: use blk_run_queue() instead of blk_start_queueing()
Tejun Heo [Sat, 18 Apr 2009 22:00:41 +0000 (07:00 +0900)]
ide: use blk_run_queue() instead of blk_start_queueing()

blk_start_queueing() is being phased out in favor of
[__]blk_run_queue().  Switch.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoide-tape: remove back-to-back REQUEST_SENSE detection
Tejun Heo [Sat, 18 Apr 2009 22:00:41 +0000 (07:00 +0900)]
ide-tape: remove back-to-back REQUEST_SENSE detection

Impact: fix an oops which always triggers

ide_tape_issue_pc() assumed drive->pc isn't NULL on invocation when
checking for back-to-back request sense issues but drive->pc can be
NULL and even when it's not NULL, it's not safe to dereference it once
the previous command is complete because pc could have been freed or
was on stack.  Kill back-to-back REQUEST_SENSE detection.

Signed-off-by: Tejun Heo <tj@kernel.org>
15 years agoblock: clear req->errors on bio completion only for fs requests
Tejun Heo [Sat, 18 Apr 2009 22:00:41 +0000 (07:00 +0900)]
block: clear req->errors on bio completion only for fs requests

Impact: subtle behavior change

For fs requests, rq is only carrier of bios and rq error status as a
whole doesn't mean much.  This is the reason why rq->errors is being
cleared on each partial completion of a request as on each partial
completion the error status is transferred to the respective bios.

For pc requests, rq->errors is used to carry error status to the
issuer and thus __end_that_request_first() doesn't clear it on such
cases.

The condition was fine till now as only fs and pc requests have used
bio and thus the bio completion path.  However, future changes will
unify data accesses to bio and all non fs users care about rq error
status.  Clear rq->errors on bio completion only for fs requests.

In general, the implicit clearing is a bit too subtle especially as
the meaning of rq->errors is completely dependent on low level
drivers.  Unifying / cleaning up rq->errors usage and letting llds
manage it would be better.  TODO comment added.

Signed-off-by: Tejun Heo <tj@kernel.org>
Acked-by: Jens Axboe <axboe@kernel.dk>
15 years agocs5536: define dma_sff_read_status() method
Sergei Shtylyov [Sat, 18 Apr 2009 15:42:20 +0000 (17:42 +0200)]
cs5536: define dma_sff_read_status() method

The driver somehow got merged with the initializer for the dma_sff_read_status()
method missing which caused kernel panic on bootup.

This should fix the kernel.org bug #13026...

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Reported-by: Arnd Hannemann <hannemann@nets.rwth-aachen.de>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
15 years agoide: fix barriers support
Bartlomiej Zolnierkiewicz [Sat, 18 Apr 2009 15:42:20 +0000 (17:42 +0200)]
ide: fix barriers support

Freeing non-slab objects is bad and results in an oops.  Fix it.

Reported-and-tested-by: Andrew Price <andy@andrewprice.me.uk>
Cc: Theodore Tso <tytso@mit.edu>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
15 years agoide: Remove void casts
Jack Stone [Sat, 18 Apr 2009 15:42:19 +0000 (17:42 +0200)]
ide: Remove void casts

Remove uneeded void casts

Signed-off-by: Jack Stone <jwjstone@fastmail.fm>
Cc: jeff@garzik.org
Cc: kernel-janitors@vger.kernel.org
Cc: Jack Stone <jwjstone@fastmail.fm>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
15 years agohpt366: use ATA_DMA_* constants
Sergei Shtylyov [Sat, 18 Apr 2009 15:42:19 +0000 (17:42 +0200)]
hpt366: use ATA_DMA_* constants

Use ATA_DMA_* constants instead of the bare numbers for the BMIDE register bits.

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
15 years agohpt366: fix HPT370 DMA timeouts
Sergei Shtylyov [Sat, 18 Apr 2009 15:42:19 +0000 (17:42 +0200)]
hpt366: fix HPT370 DMA timeouts

The big driver change in 2.4.19-rc1 introduced a regression for many HPT370[A]
chips -- DMA stopped to work completely, only causing endless timeouts...

The culprit has been identified (at last!): it turned to be the code resetting
the DMA state machine before each transfer. Stop doing it now as this counter-
measure has clearly caused more harm than good.

This should fix the kernel.org bug #7703.

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>
15 years agoMerge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzi...
Linus Torvalds [Fri, 17 Apr 2009 23:20:29 +0000 (16:20 -0700)]
Merge branch 'upstream-linus' of git://git./linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  [libata] fix build error on drivers/ata/pata_legacy.c
  pata_via: Cache and rewrite the device bit
  sata_mv: workaround for multi_count errata sata24
  sata_mv: tidy up qc->tf usage in qc_prep() functions

15 years ago[libata] fix build error on drivers/ata/pata_legacy.c
Zhenwen Xu [Fri, 17 Apr 2009 07:32:59 +0000 (15:32 +0800)]
[libata] fix build error on drivers/ata/pata_legacy.c

fix those errors:

drivers/ata/pata_legacy.c: In function ‘pdc_data_xfer_vlb’:
drivers/ata/pata_legacy.c:289: error: ‘ap’ undeclared (first use in this function)
drivers/ata/pata_legacy.c:289: error: (Each undeclared identifier is reported only once
drivers/ata/pata_legacy.c:289: error: for each function it appears in.)
drivers/ata/pata_legacy.c: At top level:
drivers/ata/pata_legacy.c:869: error: ‘ATA_PFLAG_PIO32_CHANGE’ undeclared here (not in a
+function)
make[2]: *** [drivers/ata/pata_legacy.o] Error 1
make[1]: *** [drivers/ata] Error 2

Signed-off-by: Zhenwen Xu <helight.xu@gmail.com>
Acked-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
15 years agopata_via: Cache and rewrite the device bit
Alan Cox [Fri, 17 Apr 2009 11:21:21 +0000 (12:21 +0100)]
pata_via: Cache and rewrite the device bit

Some VIA chipsets will reset the DEV bit after IEN changes on ctl. Our
optimised write path avoids doing this but we need to remove the
optimisation on these devices.

[Identified and some original patches proposed by Josehn Chan @ VIA but
discussion then all ground to a halt so given a test case I dug it back out]

Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk
Tested-by: Christoph Bisping (bug #13086)
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
15 years agosata_mv: workaround for multi_count errata sata24
Mark Lord [Mon, 13 Apr 2009 15:29:34 +0000 (11:29 -0400)]
sata_mv: workaround for multi_count errata sata24

Workaround for errata SATA#24 in sata_mv.
This errata affects WRITE_MULTI* commands when
the device multi_count produces a DRQ block size >= 4Kbytes.

We work around it here by converting such operations
into ordinary PIO_WRITEs instead.

Note that this might result in a PIO FUA write unavoidably being converted
into a non-FUA write.  In practice, any system using FUA is also going to be
using DMA rather than PIO, so this shouldn't affect anyone in the real world.

Signed-off-by: Mark Lord <mlord@pobox.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
15 years agosata_mv: tidy up qc->tf usage in qc_prep() functions
Mark Lord [Mon, 13 Apr 2009 15:27:18 +0000 (11:27 -0400)]
sata_mv: tidy up qc->tf usage in qc_prep() functions

Tidy up qc->tf accesses in the mv_qc_prep() functions.

Signed-off-by: Mark Lord <mlord@pobox.com>
Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6
Linus Torvalds [Fri, 17 Apr 2009 20:53:34 +0000 (13:53 -0700)]
Merge git://git./linux/kernel/git/gregkh/staging-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging-2.6: (43 commits)
  staging: slicoss: update README
  otus/zdusb.c: additional USB idnetifier
  Staging: go7007: fix build issues
  Staging: sxg: Fix leaks and checksum errors in transmit code path
  Staging: sxg: Fix sleep in atomic context warning while loading driver
  Staging: sxg: Use correct queue_id for transmitting non-TCP packets
  Staging: sxg: Fire watchdog timer at end of open routine to change the link
  Staging: Pohmelfs: Add load balancing between network states with the same priority.
  Staging: Pohmelfs: Added IO permissions and priorities.
  Staging: Pohmelfs: Added ->show_stats() callback.
  Staging: Pohmelfs: Drop ftrans debugging code.
  Staging: Pohmelfs: Use wait_on_page_timeout when waiting for remote directory sync instead of hardcoded 25 seconds.
  Staging: Pohmelfs: Reduce debugging noise about non-existing objects.
  Staging: Pohmelfs: Sync fs before killing it, since dentry cache is shrunk before writeback is invoked via generic_shutdown_super()
  Staging: Pohmelfs: Extend remount option.
  Staging: Pohmelfs: Set NETFS_INODE_REMOTE_SYNCED and clear NETFS_INODE_OWNED bits in the root inode.
  Staging: Pohmelfs: Added 'need_lock' variable into debug print.
  Staging: Pohmelfs: Disable read lock in pohmelfs_getattr().
  Staging: Pohmelfs: Move parent lock to the place where we really have to send a lookup request to the server.
  Staging: pohmelfs: Populate dentry cache when receiving the new readdir entry.
  ...

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6
Linus Torvalds [Fri, 17 Apr 2009 20:53:16 +0000 (13:53 -0700)]
Merge git://git./linux/kernel/git/gregkh/driver-core-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core-2.6:
  UIO: fix specific device driver missing statement for depmod
  Driver core: remove pr_fmt() from dynamic_dev_dbg() printk
  driver core: prevent device_for_each_child from oopsing
  dynamic debug: resurrect old pr_debug() semantics as pr_devel()
  Driver Core: early platform driver
  proc: mounts_poll() make consistent to mdstat_poll
  sysfs: sysfs poll keep the poll rule of regular file.
  driver core: allow non-root users to listen to uevents
  driver core: fix driver_match_device
  sysfs: don't use global workqueue in sysfs_schedule_callback()

15 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Fri, 17 Apr 2009 20:53:00 +0000 (13:53 -0700)]
Merge git://git./linux/kernel/git/gregkh/usb-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb-2.6: (22 commits)
  WUSB: correct format of wusb_chid sysfs file
  WUSB: fix oops when completing URBs for disconnected devices
  WUSB: disconnect all devices when stopping a WUSB HCD
  USB: whci-hcd: check return value of usb_hcd_link_urb_to_ep()
  USB: whci-hcd: provide a endpoint_reset method
  USB: add reset endpoint operations
  USB device codes for Motorola phone.
  usb-storage: fix mistake in Makefile
  USB: usb-serial ch341: support for DTR/RTS/CTS
  Revert USB: usb-serial ch341: support for DTR/RTS/CTS
  USB: musb: fix possible panic while resuming
  USB: musb: fix isochronous TXDMA (take 2)
  USB: musb: sanitize clearing TXCSR DMA bits (take 2)
  USB: musb: bugfixes for multi-packet TXDMA support
  USB: musb_host, fix ep0 fifo flushing
  USB: usb-storage: augment unusual_devs entry for Simple Tech/Datafab
  USB: musb_host, minor enqueue locking fix (v2)
  USB: fix oops in cdc-wdm in case of malformed descriptors
  USB: qcserial: Add extra device IDs
  USB: option: Add ids for D-Link DWM-652 3.5G modem
  ...

15 years agoMerge master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Fri, 17 Apr 2009 20:51:14 +0000 (13:51 -0700)]
Merge master.kernel.org:/home/rmk/linux-2.6-arm

* master.kernel.org:/home/rmk/linux-2.6-arm: (48 commits)
  [ARM] S3C24XX: ADC: Check pending queue before freeing adc client
  [ARM] S3C: Fix ADC driver sparse warning
  [ARM] Osiris: Fix double initialisation in machine block
  [ARM] Anubis: Fix sparse warnings for items that should be static
  [ARM] JIVE: Fix sparse warnings about items which should be static
  [ARM] S3C: Fix sparse warning from missing 's3c_device_hwmon'
  [ARM] S3C24XX: Fix sparse error in gpiolib.c
  [ARM] 5455/1: Fix IRQ noise from VIC code
  [ARM] 5454/1: ep93xx_eth: fix sparse warnings
  [ARM] remove .gitignore from include/asm-arm
  Update MAINTAINERS
  mxc defconfig updates
  mx31ads: Mark as having full regulatoion constraints with 1133-EV1 board
  mx31ads: Depend on all the WM8350 core dependencies for WM1133-EV1 board
  Fix ifdef in plat-mxc/irc.c
  MX1ADS: remove I2C ifdefs
  qong: remove AIPS[12] mappings from machine-specific iotable
  mx31ads: imoux pins should be passed in as unsigned int
  MXC: remove orphan imx_init_uart() definition
  mx31: pin definition for csi
  ...

15 years agoMerge branch 'defconfig-s3c2410' of git://aeryn.fluff.org.uk/bjdooks/linux
Russell King [Fri, 17 Apr 2009 20:35:56 +0000 (21:35 +0100)]
Merge branch 'defconfig-s3c2410' of git://aeryn.fluff.org.uk/bjdooks/linux

15 years agoMerge branch 'v2630-rc2-fixes' of git://aeryn.fluff.org.uk/bjdooks/linux
Russell King [Fri, 17 Apr 2009 20:35:44 +0000 (21:35 +0100)]
Merge branch 'v2630-rc2-fixes' of git://aeryn.fluff.org.uk/bjdooks/linux

15 years agostaging: slicoss: update README
Stephen Hemminger [Wed, 15 Apr 2009 23:52:16 +0000 (16:52 -0700)]
staging: slicoss: update README

I looked, I gagged, I left

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agootus/zdusb.c: additional USB idnetifier
Daniele Napolitano [Tue, 14 Apr 2009 23:58:33 +0000 (16:58 -0700)]
otus/zdusb.c: additional USB idnetifier

Provide support for WN111v2 USB 802.11n adapter.

Signed-off-by: Daniele Napolitano <dnax88@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: go7007: fix build issues
Greg Kroah-Hartman [Mon, 13 Apr 2009 20:16:54 +0000 (13:16 -0700)]
Staging: go7007: fix build issues

Now that TUNER_SET_TYPE_ADDR is gone from the tree, the older code kicks
in and tries to use TUNER_SET_TYPE, which went away a long time ago.

This patch removes all of this logic, as it should not be needed anymore
now, and by doing so, fixes the build.

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: sxg: Fix leaks and checksum errors in transmit code path
Mithlesh Thukral [Wed, 25 Mar 2009 10:21:49 +0000 (15:51 +0530)]
Staging: sxg: Fix leaks and checksum errors in transmit code path

Fix the transmit function for the following:
* Free XmtCmd in the error code path. This use to leak memory in
  error conditions.
* Do pci mapping after the checksum operations are over. They can
  reallocate the skb at a different location.
* Fix UDP checksum errors which were seen in wireshark

Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: sxg: Fix sleep in atomic context warning while loading driver
Mithlesh Thukral [Wed, 25 Mar 2009 10:21:14 +0000 (15:51 +0530)]
Staging: sxg: Fix sleep in atomic context warning while loading driver

Leave a spinlock before calling request_irq(). request_irq() calls kmalloc
which can sleep. This was generating a warning dump while driver is loaded.

Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: sxg: Use correct queue_id for transmitting non-TCP packets
Mithlesh Thukral [Wed, 25 Mar 2009 10:20:34 +0000 (15:50 +0530)]
Staging: sxg: Use correct queue_id for transmitting non-TCP packets

Use correct queue_id while transmitting non-TCP packets.
They should always use queue 0.

Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
15 years agoStaging: sxg: Fire watchdog timer at end of open routine to change the link
Mithlesh Thukral [Wed, 25 Mar 2009 10:19:58 +0000 (15:49 +0530)]
Staging: sxg: Fire watchdog timer at end of open routine to change the link

The watchdog timer which updates the link status was not fired at the
end of sxg_entry_open(). Add that.

Signed-off-by: LinSysSoft Sahara Team <saharaproj@linsyssoft.com>
Signed-off-by: Mithlesh Thukral <mithlesh@linsyssoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Add load balancing between network states with the same priority.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:30 +0000 (15:04 +0300)]
Staging: Pohmelfs: Add load balancing between network states with the same priority.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Added IO permissions and priorities.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:29 +0000 (15:04 +0300)]
Staging: Pohmelfs: Added IO permissions and priorities.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Added ->show_stats() callback.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:28 +0000 (15:04 +0300)]
Staging: Pohmelfs: Added ->show_stats() callback.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Drop ftrans debugging code.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:27 +0000 (15:04 +0300)]
Staging: Pohmelfs: Drop ftrans debugging code.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Use wait_on_page_timeout when waiting for remote directory sync...
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:26 +0000 (15:04 +0300)]
Staging: Pohmelfs: Use wait_on_page_timeout when waiting for remote directory sync instead of hardcoded 25 seconds.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Reduce debugging noise about non-existing objects.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:25 +0000 (15:04 +0300)]
Staging: Pohmelfs: Reduce debugging noise about non-existing objects.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Sync fs before killing it, since dentry cache is shrunk before...
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:24 +0000 (15:04 +0300)]
Staging: Pohmelfs: Sync fs before killing it, since dentry cache is shrunk before writeback is invoked via generic_shutdown_super()

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Extend remount option.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:23 +0000 (15:04 +0300)]
Staging: Pohmelfs: Extend remount option.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Set NETFS_INODE_REMOTE_SYNCED and clear NETFS_INODE_OWNED bits...
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:22 +0000 (15:04 +0300)]
Staging: Pohmelfs: Set NETFS_INODE_REMOTE_SYNCED and clear NETFS_INODE_OWNED bits in the root inode.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Added 'need_lock' variable into debug print.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:21 +0000 (15:04 +0300)]
Staging: Pohmelfs: Added 'need_lock' variable into debug print.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Disable read lock in pohmelfs_getattr().
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:20 +0000 (15:04 +0300)]
Staging: Pohmelfs: Disable read lock in pohmelfs_getattr().

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: Pohmelfs: Move parent lock to the place where we really have to send a looku...
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:19 +0000 (15:04 +0300)]
Staging: Pohmelfs: Move parent lock to the place where we really have to send a lookup request to the server.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: pohmelfs: Populate dentry cache when receiving the new readdir entry.
Evgeniy Polyakov [Fri, 27 Mar 2009 12:04:18 +0000 (15:04 +0300)]
Staging: pohmelfs: Populate dentry cache when receiving the new readdir entry.

Signed-off-by: Evgeniy Polyakov <zbr@ioremap.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoRevert Staging: SLICOSS: use gfp_kernel where possible
Lior Dotan [Wed, 11 Feb 2009 11:35:10 +0000 (13:35 +0200)]
Revert Staging: SLICOSS: use gfp_kernel where possible

Revert commit 2bb347361e2c19799431f56488a3f64de40a3aa6

This commit has been reported to cause problems:
  Mar 24 11:50:31 linuxdev kernel: BUG: sleeping function called from invalid context at mm/slub.c:1599 Mar 24 11:50:31 linuxdev kernel: in_atomic(): 1, irqs_disabled(): 0, pid: 3251, name: avahi-daemon

Cc: Lior Dotan <liodot@gmail.com>
Cc: Christopher Harrer <charrer@alacritech.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: STLC45XX should depend on GENERIC_HARDIRQS
Geert Uytterhoeven [Mon, 6 Apr 2009 13:15:55 +0000 (15:15 +0200)]
Staging: STLC45XX should depend on GENERIC_HARDIRQS

m68k allmodconfig:
| drivers/staging/stlc45xx/stlc45xx.c: In function 'stlc45xx_probe':
| drivers/staging/stlc45xx/stlc45xx.c:2456: error: implicit declaration of function 'set_irq_type'
| make[6]: *** [drivers/staging/stlc45xx/stlc45xx.o] Error 1

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Defer flush and release operations to avoid deadlocks.
Arve Hjønnevåg [Mon, 6 Apr 2009 22:13:00 +0000 (15:13 -0700)]
Staging: binder: Defer flush and release operations to avoid deadlocks.

If a transaction that contains a file descriptor fails on a later object,
the new file descriptor needs to be closed. If this is a binder file
descriptor we would deadlock in flush. If there were no other references to
the file at this point release would also be called.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Prevent the wrong thread from adding a transaction to the stack.
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:59 +0000 (15:12 -0700)]
Staging: binder: Prevent the wrong thread from adding a transaction to the stack.

If a thread is part of a transaction stack, it is only allowed to make
another call if it was the target of the top transaction on the stack.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Cast to uintptr_t instead of size_t when aligning pointers
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:58 +0000 (15:12 -0700)]
Staging: binder: Cast to uintptr_t instead of size_t when aligning pointers

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Keep a reference to the files_struct while the driver is mmapped
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:57 +0000 (15:12 -0700)]
Staging: binder: Keep a reference to the files_struct while the driver is mmapped

This prevents breaking fget_light if a single threaded application
allows incoming file descriptors (in replies or on nodes).
Should also prevent inserting a file in the wrong files_struct if the
receving process execs in the middle of a transaction (between
task_get_unused_fd_flags and task_fd_install).

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Add more offset validation.
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:56 +0000 (15:12 -0700)]
Staging: binder: Add more offset validation.

Check that datasize is not smaller than one flat_binder_object.
Check that offsets are aligned.
Check that offsets_size is aligned.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: mmap fixes.
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:55 +0000 (15:12 -0700)]
Staging: binder: mmap fixes.

Only allow a binder file pointer to be mmapped once. The buffer management
code cannot deal with more then one area.
Also remove leftover mutex_unlock if mmap fails.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Don't create two proc entries with the same name if the driver is...
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:54 +0000 (15:12 -0700)]
Staging: binder: Don't create two proc entries with the same name if the driver is opened twice in one process.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: binder: Remove VM_EXEC check.
Arve Hjønnevåg [Mon, 6 Apr 2009 22:12:53 +0000 (15:12 -0700)]
Staging: binder: Remove VM_EXEC check.

Many platforms do not support mappings without VM_EXEC.

Signed-off-by: Arve Hjønnevåg <arve@android.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: aten2011: Clean up some tty bits
Alan Cox [Tue, 7 Apr 2009 17:43:43 +0000 (18:43 +0100)]
Staging: aten2011: Clean up some tty bits

Minor fixes for tty layer stuff in this driver

Signed-off-by: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: rt2870: add ID for Sitecom WL-315
Daniele Napolitano [Wed, 1 Apr 2009 21:53:39 +0000 (14:53 -0700)]
Staging: rt2870: add ID for Sitecom WL-315

Signed-off-by: Daniele Napolitano <dnax88@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: line6: convert to snd_card_create()
Alexander Beregalov [Sun, 29 Mar 2009 15:59:13 +0000 (19:59 +0400)]
Staging: line6: convert to snd_card_create()

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: wlan-ng: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:31:28 +0000 (19:31 +0400)]
Staging: wlan-ng: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: sxg: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:30:46 +0000 (19:30 +0400)]
Staging: sxg: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: slicoss: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:30:08 +0000 (19:30 +0400)]
Staging: slicoss: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: rt3070: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:29:18 +0000 (19:29 +0400)]
Staging: rt3070: convert to netdev_ops

Also remove unused variable.

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: rt2870: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:28:11 +0000 (19:28 +0400)]
Staging: rt2870: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: rt2860: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:27:21 +0000 (19:27 +0400)]
Staging: rt2860: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: otus: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:26:25 +0000 (19:26 +0400)]
Staging: otus: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: et131x: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:25:33 +0000 (19:25 +0400)]
Staging: et131x: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: epl: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:24:32 +0000 (19:24 +0400)]
Staging: epl: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoStaging: at76: convert to netdev_ops
Alexander Beregalov [Sun, 29 Mar 2009 15:23:34 +0000 (19:23 +0400)]
Staging: at76: convert to netdev_ops

Signed-off-by: Alexander Beregalov <a.beregalov@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoWUSB: correct format of wusb_chid sysfs file
David Vrabel [Wed, 8 Apr 2009 17:36:30 +0000 (17:36 +0000)]
WUSB: correct format of wusb_chid sysfs file

Make the wusb_chid sysfs file match the ABI documentation.

Print all zeros if the WUSB host is stopped (instead of an empty file)
and end the file with a newline.

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoWUSB: fix oops when completing URBs for disconnected devices
David Vrabel [Wed, 8 Apr 2009 17:36:32 +0000 (17:36 +0000)]
WUSB: fix oops when completing URBs for disconnected devices

Fix an oops in wusbhc_giveback_urb() if the wusb device had disconnected
while an urb was in progress. Also release the ref count obtained here.

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoWUSB: disconnect all devices when stopping a WUSB HCD
David Vrabel [Wed, 8 Apr 2009 17:36:33 +0000 (17:36 +0000)]
WUSB: disconnect all devices when stopping a WUSB HCD

Make sure all WUSB devices are disconnected when stopping a WUSB HCD so
that we don't leak the devices' wusb_dev structures.

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: whci-hcd: check return value of usb_hcd_link_urb_to_ep()
David Vrabel [Wed, 8 Apr 2009 17:36:31 +0000 (17:36 +0000)]
USB: whci-hcd: check return value of usb_hcd_link_urb_to_ep()

Check the return value of usb_hcd_link_urb_to_ep() and do not add the
urb to the ASL/PZL if it returns an error.

Omitting the check results in urbs that appear to be submitted
successfully but then cannot be unliked (because
usb_hcd_check_unlink_urb() returns an error).  This can cause khubd (for
example) to block forever in usb_kill_urb().

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: whci-hcd: provide a endpoint_reset method
David Vrabel [Wed, 8 Apr 2009 17:36:29 +0000 (17:36 +0000)]
USB: whci-hcd: provide a endpoint_reset method

Provide a endpoint_reset method to reset sequence number and current
window.  This QHead information can only be changed while the qset is
not in a schedule.

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: add reset endpoint operations
David Vrabel [Wed, 8 Apr 2009 17:36:28 +0000 (17:36 +0000)]
USB: add reset endpoint operations

Wireless USB endpoint state has a sequence number and a current
window and not just a single toggle bit.  So allow HCDs to provide a
endpoint_reset method and call this or clear the software toggles as
required (after a clear halt, set configuration etc.).

usb_settoggle() and friends are then HCD internal and are moved into
core/hcd.h and all device drivers call usb_reset_endpoint() instead.

If the device endpoint state has been reset (with a clear halt) but
the host endpoint state has not then subsequent data transfers will
not complete. The device will only work again after it is reset or
disconnected.

Signed-off-by: David Vrabel <david.vrabel@csr.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB device codes for Motorola phone.
Dr. Greg Wettstein [Sat, 11 Apr 2009 14:12:08 +0000 (09:12 -0500)]
USB device codes for Motorola phone.

The v950 appears to be a ruggedized version of the Motorola Razor
phone.  Tethering to the phone to use it in 'phone as modem' mode
requires the use of the specialized moto-modem driver which layers
over the usb-serial driver.  Support for the v950 was added simply
adding the device ID's for the phone.

Signed-off-by: Dr. Greg Wettstein <greg@enjellic.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agousb-storage: fix mistake in Makefile
Alan Stern [Thu, 9 Apr 2009 18:59:44 +0000 (14:59 -0400)]
usb-storage: fix mistake in Makefile

This patch (as1228) fixes a Makefile error introduced when the
subdrivers in usb-storage were split out into separate modules.  The
intention is that when CONFIG_USB_LIBUSUAL is set, libusual.o and
usual-tables.o should be combined into a single object file (called
usb-libusual).  The current Makefile will instead create two separate
objects, and the result won't load properly.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-and-tested-by: Alan Jenkins <sourcejedi.lkml@googlemail.com>
Reported-and-tested-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: usb-serial ch341: support for DTR/RTS/CTS
Werner Cornelius [Fri, 16 Jan 2009 20:02:41 +0000 (21:02 +0100)]
USB: usb-serial ch341: support for DTR/RTS/CTS

commit 664d5df92e88b6ef091048a802b3750f4e989180 upstream.

Fixup of Werner Cornelius patch to the ch341 USB-serial driver, which adds:
- support all baudrates, not just a hard-coded set
- support for controlling DTR, RTS and CTS

Features still missing:
- character length other than 8 bits
- parity settings
- break control

I adapted his patch for the new usb_serial API introduced in 2.6.25-git8 by
Alan Cox on 22 July 2008. Non-compliance to the new API was a reason for
refusing a similar patch from Tollef Fog Heen.

Usage example by Tollef Fog Heen :
        TEMPer USB thermometer <http://err.no/src/TEMPer.c>

based on a patch by:

From: Tollef Fog Heen <tfheen@err.no>

* Implement support for all baud rates rather than just a hard
  coded set.
* Make it possible to control status and control lines
* Grab a bunch of #defines from FreeBSD to reduce the number of
  magic numbers in the file

Signed-off-by: Werner Cornelius <Werner.Cornelius@cornelius-consult.de>
Signed-off-by: Boris Hajduk <boris@hajduk.org>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Tollef Fog Heen <tfheen@err.no>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoRevert USB: usb-serial ch341: support for DTR/RTS/CTS
Greg Kroah-Hartman [Tue, 7 Apr 2009 18:56:32 +0000 (11:56 -0700)]
Revert USB: usb-serial ch341: support for DTR/RTS/CTS

Reverts commit 664d5df92e88b6ef091048a802b3750f4e989180 as the commit
log information was not complete, and we didn't have a proper
signed-off-by by the author of the original BSD code.

Cc: Werner Cornelius <Werner.Cornelius@cornelius-consult.de>
Cc: Boris Hajduk <boris@hajduk.org>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: musb: fix possible panic while resuming
Kim Kyuwon [Fri, 27 Mar 2009 01:56:51 +0000 (18:56 -0700)]
USB: musb: fix possible panic while resuming

During driver resume processing, musb could cause a kernel panic.
Fix by enabling the clock earlier, with the resume_early method.

Signed-off-by: Kim Kyuwon <q1.kim@samsung.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: musb: fix isochronous TXDMA (take 2)
Sergei Shtylyov [Fri, 27 Mar 2009 01:29:19 +0000 (18:29 -0700)]
USB: musb: fix isochronous TXDMA (take 2)

Multi-frame isochronous TX URBs transfers in DMA mode never
complete with CPPI DMA because musb_host_tx() doesn't restart
DMA on the second frame, only emitting a debug message.
With Inventra DMA they complete, but in PIO mode.  To fix:

 - Factor out programming of the DMA transfer from
   musb_ep_program() into musb_tx_dma_program();

 - Reorder the code at the end of musb_host_tx() to
   facilitate the fallback to PIO iff DMA fails;

 - Handle the buffer offset consistently for both
   PIO and DMA modes;

 - Add an argument to musb_ep_program() for the same
   reason (it only worked correctly with non-zero
   offset of the first frame in PIO mode);

 - Set the completed isochronous frame descriptor's
   'actual_length' and 'status' fields correctly in
   DMA mode.

Also, since CPPI reportedly doesn't like sending isochronous
packets in the RNDIS mode, change the criterion for this
mode to be used only for multi-packet transfers.  (There's
no need for that mode in the single-packet case anyway.)

[ dbrownell@users.sourceforge.net: split comment paragraph
into bullet list, shrink patch delta, style tweaks ]

Signed-off-by: Pavel Kiryukhin <pkiryukhin@ru.mvista.com>
Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: musb: sanitize clearing TXCSR DMA bits (take 2)
Sergei Shtylyov [Fri, 27 Mar 2009 01:27:47 +0000 (18:27 -0700)]
USB: musb: sanitize clearing TXCSR DMA bits (take 2)

The MUSB code clears TXCSR_DMAMODE incorrectly in several
places, either asserting that TXCSR_DMAENAB is clear (when
sometimes it isn't) or clearing both bits together.  Recent
versions of the programmer's guide require DMAENAB to be
cleared first, although some older ones didn't.

Fix this and while at it:

 - In musb_gadget::txstate(), stop clearing the AUTOSET
   and DMAMODE bits for the CPPI case since they never
   get set anyway (the former bit is reserved on DaVinci);
   but do clear the DMAENAB bit on the DMA error path.

 - In musb_host::musb_ep_program(), remove the duplicate
   DMA controller specific code code clearing the TXCSR
   previous state, add the code to clear TXCSR DMA bits
   on the Inventra DMA error path, to replace such code
   (executed late) on the PIO path.

 - In musbhsdma::dma_channel_abort()/dma_controller_irq(),
   add/use the 'offset' variable to avoid MUSB_EP_OFFSET()
   invocations on every RXCSR/TXCSR access.

[dbrownell@users.sourceforge.net: don't introduce CamelCase,
shrink diff]

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: musb: bugfixes for multi-packet TXDMA support
Sergei Shtylyov [Fri, 27 Mar 2009 01:26:40 +0000 (18:26 -0700)]
USB: musb: bugfixes for multi-packet TXDMA support

We really want to use DMA mode 1 for all multi-packet transfers;
that's one IRQ on DMA completion, instead of one per packet.

There is an important issue with such transfers, especially on
the host side:  when such transfers end with a full-size packet,
we must defer musb_dma_completion() calls until the FIFO empties.
Else we report URB completions too soon, and may clobber data in
the FIFO fifo when writing the next packet (losing data).

The Inventra DMA support uses DMA mode 1, but it ignores that
issue.  The CPPI DMA support uses mode 0, but doesn't handle
its TXPKTRDY interrupts quite right either; it can get stale
"packet ready" interrupts, and report transfer completion too
early using slightly different code paths, also losing data.

So I'm solving it in a generic way -- by adding a sort of the
"interrupt filter" into musb_host_tx(), catching these cases
where a DMA completion IRQ doesn't suffice and removing some
needlessly controller-specific logic.  When a TXDMA interrupt
happens and DMA request mode 1 is active, that filter resets
to mode 0 and defers URB completion processing until TXPKTRDY,
unless the FIFO is already empty.  Related filtering logic in
Inventra and CPPI code gets removed.

Since it should be competely safe now to use the DMA request
mode 1 for host side transfers with the CPPI DMA controller,
set it in musb_h_tx_dma_start() ... now renamed (and shared).

[ dbrownell@users.sourceforge.net: don't introduce more
CamElCase; use more concise explanations ]

Signed-off-by: Sergei Shtylyov <sshtylyov@ru.mvista.com>
Cc: Felipe Balbi <felipe.balbi@nokia.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: musb_host, fix ep0 fifo flushing
David Brownell [Fri, 27 Mar 2009 00:38:30 +0000 (17:38 -0700)]
USB: musb_host, fix ep0 fifo flushing

The MUSB host side can't share generic TX FIFO flush logic
with EP0; the EP0 TX status register bits are different
from those for other entpoints.

Resolve this issue by providing a new EP0-specific routine
to flush and reset the FIFO, which pays careful attention to
restrictions listed in the latest programmer's guide.  This
gets rid of an open issue whereby the usbtest control write
test (#14) failed.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: usb-storage: augment unusual_devs entry for Simple Tech/Datafab
Alan Stern [Tue, 24 Mar 2009 14:39:13 +0000 (10:39 -0400)]
USB: usb-storage: augment unusual_devs entry for Simple Tech/Datafab

This patch (as1227) adds the MAX_SECTORS_64 flag to the unusual_devs
entry for the Simple Tech/Datafab controller.  This fixes Bugzilla
#12882.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-and-tested-by: binbin <binbinsh@gmail.com>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: musb_host, minor enqueue locking fix (v2)
David Brownell [Fri, 27 Mar 2009 00:36:57 +0000 (17:36 -0700)]
USB: musb_host, minor enqueue locking fix (v2)

Someone noted that the enqueue path used an unlocked access
for usb_host_endpoint->hcpriv ... fix that, by being safe
and always accessing it under spinlock protection.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: fix oops in cdc-wdm in case of malformed descriptors
Oliver Neukum [Sat, 4 Apr 2009 07:25:15 +0000 (09:25 +0200)]
USB: fix oops in cdc-wdm in case of malformed descriptors

cdc-wdm needs to ignore extremely malformed descriptors.

Signed-off-by: Oliver Neukum <oliver@neukum.org>
Cc: stable <stable@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: qcserial: Add extra device IDs
Matthew Garrett [Sat, 4 Apr 2009 16:24:24 +0000 (17:24 +0100)]
USB: qcserial: Add extra device IDs

Add a set of device IDs from the Windows drivers. These aren't complete
(there's a couple of cases where a QDL device is identified without the
associated modem being identified), but it's better than the current
situation.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
15 years agoUSB: option: Add ids for D-Link DWM-652 3.5G modem
Pascal Terjan [Tue, 7 Apr 2009 12:40:42 +0000 (14:40 +0200)]
USB: option: Add ids for D-Link DWM-652 3.5G modem

This patch allows D-Link DWM-652 3.5G modem to work.
It is an express card but was only tested with the provided usb adapter as I
don't have machines with express card connector.

/dev/ttyUSB{0,1,2} get created, and using comgt on ttyUSB1 works fine :

[root@plop tmp]# comgt -d /dev/ttyUSB1 -e

Enter PIN number: XXXX
Waiting for Registration..(120 sec max).
Registered on Home network: "Orange France",2
Signal Quality: 15,99

From: Pascal Terjan <pterjan@mandriva.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>