7 years agoRevert "cfq: Remove special treatment for metadata rqs."
Jens Axboe [Fri, 19 Aug 2011 06:34:48 +0000]
Revert "cfq: Remove special treatment for metadata rqs."

We have a kernel build regression since 3.1-rc1, which is about 10%
regression. The kernel source is in an ext3 filesystem.
Alex Shi bisect it to commit:
commit a07405b7802691d29ab3b23bdc76ee6d006aad0b
Author: Justin TerAvest <teravest@google.com>
Date:   Sun Jul 10 22:09:19 2011 +0200

    cfq: Remove special treatment for metadata rqs.

Apparently this is caused by lack metadata preemption, where ext3/ext4
do use READ_META. I didn't see a way to fix the issue, so suggest
reverting the patch.

This reverts commit a07405b7802691d29ab3b23bdc76ee6d006aad0b.

Reported-by: Alex Shi<alex.shi@intel.com>
Reported-by: Shaohua Li<shaohua.li@intel.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblock: fix flush machinery for stacking drivers with differring flush flags
Jeff Moyer [Mon, 15 Aug 2011 19:37:25 +0000]
block: fix flush machinery for stacking drivers with differring flush flags

Commit ae1b1539622fb46e51b4d13b3f9e5f4c713f86ae, block: reimplement
FLUSH/FUA to support merge, introduced a performance regression when
running any sort of fsyncing workload using dm-multipath and certain
storage (in our case, an HP EVA).  The test I ran was fs_mark, and it
dropped from ~800 files/sec on ext4 to ~100 files/sec.  It turns out
that dm-multipath always advertised flush+fua support, and passed
commands on down the stack, where those flags used to get stripped off.
The above commit changed that behavior:

static inline struct request *__elv_next_request(struct request_queue *q)
{
        struct request *rq;

        while (1) {
-               while (!list_empty(&q->queue_head)) {
+               if (!list_empty(&q->queue_head)) {
                        rq = list_entry_rq(q->queue_head.next);
-                       if (!(rq->cmd_flags & (REQ_FLUSH | REQ_FUA)) ||
-                           (rq->cmd_flags & REQ_FLUSH_SEQ))
-                               return rq;
-                       rq = blk_do_flush(q, rq);
-                       if (rq)
-                               return rq;
+                       return rq;
                }

Note that previously, a command would come in here, have
REQ_FLUSH|REQ_FUA set, and then get handed off to blk_do_flush:

struct request *blk_do_flush(struct request_queue *q, struct request *rq)
{
        unsigned int fflags = q->flush_flags; /* may change, cache it */
        bool has_flush = fflags & REQ_FLUSH, has_fua = fflags & REQ_FUA;
        bool do_preflush = has_flush && (rq->cmd_flags & REQ_FLUSH);
        bool do_postflush = has_flush && !has_fua && (rq->cmd_flags &
        REQ_FUA);
        unsigned skip = 0;
...
        if (blk_rq_sectors(rq) && !do_preflush && !do_postflush) {
                rq->cmd_flags &= ~REQ_FLUSH;
if (!has_fua)
rq->cmd_flags &= ~REQ_FUA;
        return rq;
}

So, the flush machinery was bypassed in such cases (q->flush_flags == 0
&& rq->cmd_flags & (REQ_FLUSH|REQ_FUA)).

Now, however, we don't get into the flush machinery at all.  Instead,
__elv_next_request just hands a request with flush and fua bits set to
the scsi_request_fn, even if the underlying request_queue does not
support flush or fua.

The agreed upon approach is to fix the flush machinery to allow
stacking.  While this isn't used in practice (since there is only one
request-based dm target, and that target will now reflect the flush
flags of the underlying device), it does future-proof the solution, and
make it function as designed.

In order to make this work, I had to add a field to the struct request,
inside the flush structure (to store the original req->end_io).  Shaohua
had suggested overloading the union with rb_node and completion_data,
but the completion data is used by device mapper and can also be used by
other drivers.  So, I didn't see a way around the additional field.

I tested this patch on an HP EVA with both ext4 and xfs, and it recovers
the lost performance.  Comments and other testers, as always, are
appreciated.

Cheers,
Jeff

Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblock: improve rq_affinity placement
Shaohua Li [Thu, 11 Aug 2011 08:39:04 +0000]
block: improve rq_affinity placement

This patch reverts commit 35ae66e0a09ab70ed(block: Make rq_affinity = 1
work as expected). The purpose is to avoid an unnecessary IPI.
Let's take an example. My test box has cpu 0-7, one socket. Say request is
added from CPU 1, blk_complete_request() occurs at CPU 7. Without the reverted
patch, softirq will be done at CPU 7. With it, an IPI will be directed to CPU
0, and softirq will be done at CPU 0. In this case, doing softirq at CPU 0 and
CPU 7 have no difference from cache sharing point view and we can avoid an
ipi if doing it in CPU 7.
An immediate concern is this is just like QUEUE_FLAG_SAME_FORCE, but actually
not. blk_complete_request() is running in interrupt handler, and currently
I/O controller doesn't support multiple interrupts (I checked several LSI
cards and AHCI), so only one CPU can run blk_complete_request(). This is
still quite different as QUEUE_FLAG_SAME_FORCE.
Since only one CPU runs softirq, the only difference with below patch is
softirq not always runs at the first CPU of a group.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblktrace: add FLUSH/FUA support
Namhyung Kim [Thu, 11 Aug 2011 08:36:05 +0000]
blktrace: add FLUSH/FUA support

Add FLUSH/FUA support to blktrace. As FLUSH precedes WRITE and/or
FUA follows WRITE, use the same 'F' flag for both cases and
distinguish them by their (relative) position. The end results
look like (other flags might be shown also):

 - WRITE:            W
 - WRITE_FLUSH:      FW
 - WRITE_FUA:        WF
 - WRITE_FLUSH_FUA:  FWF

Note that we reuse TC_BARRIER due to lack of bit space of act_mask
so that the older versions of blktrace tools will report flush
requests as barriers from now on.

Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoMove some REQ flags to the common bio/request area
Matthew Wilcox [Thu, 11 Aug 2011 08:36:03 +0000]
Move some REQ flags to the common bio/request area

REQ_SECURE, REQ_FLUSH and REQ_FUA may all be set on a bio as well as
on a request, so relocate them to the shared part of the enum.

Signed-off-by: Matthew Wilcox <willy@linux.intel.com>
Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoMerge branch 'stable/for-jens' of git://git.kernel.org/pub/scm/linux/kernel/git/konra...
Jens Axboe [Tue, 9 Aug 2011 18:43:26 +0000]
Merge branch 'stable/for-jens' of git://git./linux/kernel/git/konrad/xen into for-linus

7 years agoallow blk_flush_policy to return REQ_FSEQ_DATA independent of *FLUSH
Jeff Moyer [Tue, 9 Aug 2011 18:32:09 +0000]
allow blk_flush_policy to return REQ_FSEQ_DATA independent of *FLUSH

blk_insert_flush has the following check:

/*
 * If there's data but flush is not necessary, the request can be
 * processed directly without going through flush machinery.  Queue
 * for normal execution.
 */
if ((policy & REQ_FSEQ_DATA) &&
    !(policy & (REQ_FSEQ_PREFLUSH | REQ_FSEQ_POSTFLUSH))) {
list_add_tail(&rq->queuelist, &q->queue_head);
return;
}

However, blk_flush_policy will not return with policy set to only
REQ_FSEQ_DATA:

static unsigned int blk_flush_policy(unsigned int fflags, struct request *rq)
{
unsigned int policy = 0;

if (fflags & REQ_FLUSH) {
if (rq->cmd_flags & REQ_FLUSH)
policy |= REQ_FSEQ_PREFLUSH;
if (blk_rq_sectors(rq))
policy |= REQ_FSEQ_DATA;
if (!(fflags & REQ_FUA) && (rq->cmd_flags & REQ_FUA))
policy |= REQ_FSEQ_POSTFLUSH;
}
return policy;
}

Notice that REQ_FSEQ_DATA is only set if REQ_FLUSH is set.  Fix this
mismatch by moving the setting of REQ_FSEQ_DATA outside of the REQ_FLUSH
check.

Tejun notes:

  Hmmm... yes, this can become a correctness issue if (and only if)
  blk_queue_flush() is called to change q->flush_flags while requests
  are in-flight; otherwise, requests wouldn't reach the function at all.
  Also, I think it would be a generally good idea to always set
  FSEQ_DATA if the request has data.

Cheers,
Jeff

Signed-off-by: Jeff Moyer <jmoyer@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoxen/blkback: Make description more obvious.
Konrad Rzeszutek Wilk [Wed, 3 Aug 2011 15:12:17 +0000]
xen/blkback: Make description more obvious.

With the frontend having Xen but the backend not, it just looks odd:

  <*>   Xen virtual block device support
  <*>   Block-device backend driver

Fix it to have the 'Xen' in front of it.

Reported-by: Sander Eikelenboom <linux@eikelenboom.it>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

7 years agocfq-iosched: Add documentation about idling
Vivek Goyal [Fri, 5 Aug 2011 07:42:20 +0000]
cfq-iosched: Add documentation about idling

There are always questions about why CFQ is idling on various conditions.
Recent ones is Christoph asking again why to idle on REQ_NOIDLE. His
assertion is that XFS is relying more and more on workqueues and is
concerned that CFQ idling on IO from every workqueue will impact
XFS badly.

So he suggested that I add some more documentation about CFQ idling
and that can provide more clarity on the topic and also gives an
opprotunity to poke a hole in theory and lead to improvements.

So here is my attempt at that. Any comments are welcome.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblock: Make rq_affinity = 1 work as expected
Tao Ma [Fri, 5 Aug 2011 07:37:10 +0000]
block: Make rq_affinity = 1 work as expected

Commit 5757a6d76c introduced a new rq_affinity = 2 so as to make
the request completed in the __make_request cpu. But it makes the
old rq_affinity = 1 not work any more. The root cause is that
if the 'cpu' and 'req->cpu' is in the same group and cpu != req->cpu,
ccpu will be the same as group_cpu, so the completion will be
excuted in the 'cpu' not 'group_cpu'.

This patch fix problem by simpling removing group_cpu and the codes
are more explicit now. If ccpu == cpu, we complete in cpu, otherwise
we raise_blk_irq to ccpu.

Cc: Christoph Hellwig <hch@infradead.org>
Cc: Roland Dreier <roland@purestorage.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Jens Axboe <jaxboe@fusionio.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Reviewed-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblock: swim3: fix unterminated of_device_id table
Axel Lin [Wed, 3 Aug 2011 13:02:55 +0000]
block: swim3: fix unterminated of_device_id table

of_device_id structures need a NULL terminating entry, add it.

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblock/genhd.c: remove useless cast in diskstats_show()
Herbert Poetzl [Tue, 2 Aug 2011 10:43:50 +0000]
block/genhd.c: remove useless cast in diskstats_show()

Remove the (unsigned long long) cast in diskstats_show() and adjusts the
seq_printf() format string to 'unsigned long'

diskstats_show() uses part_stat_read() to get the stats, which either
accesses the specified field in the struct disk_stats directly (non SMP)
or sums up the per CPU values in a variable of the same type as the field,
so in any case the result will have the same type and range as the
specified field which for all disk_stats entries is unsigned long

Also, for unsigned long ranges the output of %lu should be identical to
the one of %llu, so no change in the actual proc entry contents.

Signed-off-by: Herbert Poetzl <herbert@13thfloor.at>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agodrivers/cdrom/cdrom.c: relax check on dvd manufacturer value
Andrew Morton [Tue, 2 Aug 2011 10:43:50 +0000]
drivers/cdrom/cdrom.c: relax check on dvd manufacturer value

The report has an ISO which has a very long manufacturer ID.  It seems
that Linux is wrong, not the ISO maker.

Relax the check for the length of this field: emit a warning and truncate
the incoming data to 2048 bytes rather than rejecting the entire thing.

dvd_manufact.value isn't null-terminated.  I'm not even sure if it's a
string.  The kernel doesn't apepar to use it anyway.

Addresses https://bugzilla.kernel.org/show_bug.cgi?id=39062

Reported-by: <ale.goujon@gmail.com>
Tested-by: <ale.goujon@gmail.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agodrivers/block/drbd/drbd_nl.c: use bitmap_parse instead of __bitmap_parse
H Hartley Sweeten [Tue, 2 Aug 2011 10:43:49 +0000]
drivers/block/drbd/drbd_nl.c: use bitmap_parse instead of __bitmap_parse

The buffer 'sc.cpu_mask' is a kernel buffer.  If bitmap_parse is used
instead of __bitmap_parse the extra parameter that indicates a kernel
buffer is not needed.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Cc: Lars Ellenberg <drbd-dev@lists.linbit.com>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agobsg-lib: add module.h include
Jens Axboe [Tue, 2 Aug 2011 08:43:35 +0000]
bsg-lib: add module.h include

Due to conflicts with the moduleh tree in linux-next, we
run into an include file mess. We really need export.h
in that tree, but if we add module.h locally then the
issue is easier to resolve.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agocfq-iosched: Reduce linked group count upon group destruction
Vivek Goyal [Tue, 2 Aug 2011 07:24:09 +0000]
cfq-iosched: Reduce linked group count upon group destruction

FQ keeps track of number of groups which are linked on blkcg->blkg_list.
This is useful to avoid races between queue exit and cgroup exit code
paths. So if at the request queue exit time linked group count is not
zero, that means there are some group out there which is yet to be
deleted under rcu read period and queue exit code should wait for
on rcu period.

In my previous patch I forgot to decrease the number of group count.
So in current form, we nr_blkcg_linked_grps is always non-zero and
we will always wait one rcu period (if BLK_CGROUP=y). The side effect
of this is that it can increase boot time. I am surprised, nobody
complained so far.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblk-throttle: correctly determine sync bio
Shaohua Li [Mon, 1 Aug 2011 08:31:06 +0000]
blk-throttle: correctly determine sync bio

read request is always sync. Using rw_is_sync() to determine
if a bio is sync.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoloop: fix deadlock when sysfs and LOOP_CLR_FD race against each other
Kay Sievers [Sun, 31 Jul 2011 20:21:35 +0000]
loop: fix deadlock when sysfs and LOOP_CLR_FD race against each other

LOOP_CLR_FD takes lo->lo_ctl_mutex and tries to remove the loop sysfs
files. Sysfs calls show() and waits for lo->lo_ctl_mutex. LOOP_CLR_FD
waits for show() to finish to remove the sysfs file.

  cat /sys/class/block/loop0/loop/backing_file
    mutex_lock_nested+0x176/0x350
    ? loop_attr_do_show_backing_file+0x2f/0xd0 [loop]
    ? loop_attr_do_show_backing_file+0x2f/0xd0 [loop]
    loop_attr_do_show_backing_file+0x2f/0xd0 [loop]
    dev_attr_show+0x1b/0x60
    ? sysfs_read_file+0x86/0x1a0
    ? __get_free_pages+0x12/0x50
    sysfs_read_file+0xaf/0x1a0

  ioctl(LOOP_CLR_FD):
    wait_for_common+0x12c/0x180
    ? try_to_wake_up+0x2a0/0x2a0
    wait_for_completion+0x18/0x20
    sysfs_deactivate+0x178/0x180
    ? sysfs_addrm_finish+0x43/0x70
    ? sysfs_addrm_start+0x1d/0x20
    sysfs_addrm_finish+0x43/0x70
    sysfs_hash_and_remove+0x85/0xa0
    sysfs_remove_group+0x59/0x100
    loop_clr_fd+0x1dc/0x3f0 [loop]
    lo_ioctl+0x223/0x7a0 [loop]

Instead of taking the lo_ctl_mutex from sysfs code, take the inner
lo->lo_lock, to protect the access to the backing_file data.

Thanks to Tejun for help debugging and finding a solution.

Cc: Milan Broz <mbroz@redhat.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Cc: stable@kernel.org
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoloop: add BLK_DEV_LOOP_MIN_COUNT=%i to allow distros 0 pre-allocated loop devices
Kay Sievers [Sun, 31 Jul 2011 20:08:04 +0000]
loop: add BLK_DEV_LOOP_MIN_COUNT=%i to allow distros 0 pre-allocated loop devices

Instead of unconditionally creating a fixed number of dead loop
devices which need to be investigated by storage handling services,
even when they are never used, we allow distros start with 0
loop devices and have losetup(8) and similar switch to the dynamic
/dev/loop-control interface instead of searching /dev/loop%i for free
devices.

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoloop: add management interface for on-demand device allocation
Kay Sievers [Sun, 31 Jul 2011 20:08:04 +0000]
loop: add management interface for on-demand device allocation

Loop devices today have a fixed pre-allocated number of usually 8.
The number can only be changed at module init time. To find a free
device to use, /dev/loop%i needs to be scanned, and all devices need
to be opened until a free one is possibly found.

This adds a new /dev/loop-control device node, that allows to
dynamically find or allocate a free device, and to add and remove loop
devices from the running system:
 LOOP_CTL_ADD adds a specific device. Arg is the number
 of the device. It returns the device i or a negative
 error code.

 LOOP_CTL_REMOVE removes a specific device, Arg is the
 number the device. It returns the device i or a negative
 error code.

 LOOP_CTL_GET_FREE finds the next unbound device or allocates
 a new one. No arg is given. It returns the device i or a
 negative error code.

The loop kernel module gets automatically loaded when
/dev/loop-control is accessed the first time. The alias
specified in the module, instructs udev to create this
'dead' device node, even when the module is not loaded.

Example:
 cfd = open("/dev/loop-control", O_RDWR);

 # add a new specific loop device
 err = ioctl(cfd, LOOP_CTL_ADD, devnr);

 # remove a specific loop device
 err = ioctl(cfd, LOOP_CTL_REMOVE, devnr);

 # find or allocate a free loop device to use
 devnr = ioctl(cfd, LOOP_CTL_GET_FREE);

 sprintf(loopname, "/dev/loop%i", devnr);
 ffd = open("backing-file", O_RDWR);
 lfd = open(loopname, O_RDWR);
 err = ioctl(lfd, LOOP_SET_FD, ffd);

Cc: Tejun Heo <tj@kernel.org>
Cc: Karel Zak <kzak@redhat.com>
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoloop: replace linked list of allocated devices with an idr index
Kay Sievers [Sun, 31 Jul 2011 20:08:04 +0000]
loop: replace linked list of allocated devices with an idr index

Replace the linked list, that keeps track of allocated devices, with an
idr index to allow a more efficient lookup of devices.

Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoblock: add bsg helper library
Mike Christie [Sun, 31 Jul 2011 20:05:09 +0000]
block: add bsg helper library

This moves the FC classes bsg code to the block layer and
makes it a lib so that other classes like iscsi and SAS can use it.

It is helpful because working with the request queue, bios,
creating scatterlists, etc are a pain that the LLD does not
have to worry about with normal IOs and should not have to
worry about for bsg requests.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: Jens Axboe <jaxboe@fusionio.com>

7 years agoMerge branch 'nfs-for-3.1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Sun, 31 Jul 2011 16:26:50 +0000]
Merge branch 'nfs-for-3.1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

* 'nfs-for-3.1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (28 commits)
  pnfsblock: write_pagelist handle zero invalid extents
  pnfsblock: note written INVAL areas for layoutcommit
  pnfsblock: bl_write_pagelist
  pnfsblock: bl_read_pagelist
  pnfsblock: cleanup_layoutcommit
  pnfsblock: encode_layoutcommit
  pnfsblock: merge rw extents
  pnfsblock: add extent manipulation functions
  pnfsblock: bl_find_get_extent
  pnfsblock: xdr decode pnfs_block_layout4
  pnfsblock: call and parse getdevicelist
  pnfsblock: merge extents
  pnfsblock: lseg alloc and free
  pnfsblock: remove device operations
  pnfsblock: add device operations
  pnfsblock: basic extent code
  pnfsblock: use pageio_ops api
  pnfsblock: add blocklayout Kconfig option, Makefile, and stubs
  pnfs: cleanup_layoutcommit
  pnfs: ask for layout_blksize and save it in nfs_server
  ...

7 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg...
Linus Torvalds [Sun, 31 Jul 2011 16:25:37 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/penberg/slab-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6:
  slab: use NUMA_NO_NODE
  slab: remove one NR_CPUS dependency

7 years agoMerge branch 'urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 31 Jul 2011 16:25:09 +0000]
Merge branch 'urgent-for-linus' of git://git./linux/kernel/git/wfg/writeback

* 'urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/wfg/writeback:
  don't busy retry the inode on failed grab_super_passive()

7 years agoMerge git://git.infradead.org/battery-2.6
Linus Torvalds [Sun, 31 Jul 2011 16:24:50 +0000]
Merge git://git.infradead.org/battery-2.6

* git://git.infradead.org/battery-2.6:
  gpio-charger: Fix checking return value of request_any_context_irq
  power_supply: MAX17042: Support additional properties
  max8903_charger: Allow platform data to be __initdata
  power_supply: Add charger driver for MAX8998/LP3974
  power_supply: Add charger driver for MAX8997/8966
  max17042_battery: Remove obsolete cleanup for clientdata
  twl4030_charger: Fix warnings
  wm831x_power: Support multiple instances
  wm831x_backup: Support multiple instances
  apm_power: Fix style error in macros
  s3c_adc_battery: Fix annotation for s3c_adc_battery_probe()
  bq20z75: Enable detection after registering
  bq20z75: Add support for external notification

7 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/brodo/cpupowerutils
Linus Torvalds [Sun, 31 Jul 2011 16:23:40 +0000]
Merge git://git./linux/kernel/git/brodo/cpupowerutils

* git://git.kernel.org/pub/scm/linux/kernel/git/brodo/cpupowerutils:
  cpupower: Do detect IDA (opportunistic processor performance) via cpuid
  cpupower: Show Intel turbo ratio support via ./cpupower frequency-info
  cpupowerutils: increase MAX_LINE_LEN
  cpupower: Rename package from cpupowerutils to cpupower
  cpupowerutils: Rename: libcpufreq->libcpupower
  cpupowerutils: use kernel version-derived version string
  cpupowerutils: utils - ConfigStyle bugfixes
  cpupowerutils: helpers - ConfigStyle bugfixes
  cpupowerutils: idle_monitor - ConfigStyle bugfixes
  cpupowerutils: lib - ConfigStyle bugfixes
  cpupowerutils: bench - ConfigStyle bugfixes
  cpupowerutils: do not update po files on each and every compile
  cpupowerutils: remove ccdv, use kernel quiet/verbose mechanism
  cpupowerutils: use COPYING, CREDITS from top-level directory
  cpupowerutils - cpufrequtils extended with quite some features

7 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6
Linus Torvalds [Sun, 31 Jul 2011 16:23:08 +0000]
Merge git://git./linux/kernel/git/brodo/pcmcia-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/brodo/pcmcia-2.6:
  smc91c92_cs.c: fix bogus compiler warning
  orinoco_cs: be more careful when matching cards with ID 0x0156:0x0002
  hostap_cs: support cards with "Version 01.02" as third product ID
  pcmcia: add PCMCIA_DEVICE_MANF_CARD_PROD_ID3
  pxa2xx pcmcia - stargate 2 use gpio array.
  pcmcia: pxa2xx: remove empty socket_init / socket_resume functions.
  drivers:pcmcia:soc_common: make socket_init and socket_suspend optional

7 years agopnfsblock: write_pagelist handle zero invalid extents
Peng Tao [Sun, 31 Jul 2011 00:52:56 +0000]
pnfsblock: write_pagelist handle zero invalid extents

For invalid extents, find other pages in the same fsblock and write them out.

[pnfsblock: write_begin]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: note written INVAL areas for layoutcommit
Fred Isaman [Sun, 31 Jul 2011 00:52:55 +0000]
pnfsblock: note written INVAL areas for layoutcommit

Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: bl_write_pagelist
Fred Isaman [Sun, 31 Jul 2011 00:52:54 +0000]
pnfsblock: bl_write_pagelist

Note: When upper layer's read/write request cannot be fulfilled, the block
layout driver shouldn't silently mark the page as error. It should do
what can be done and  leave the rest to the upper layer. To do so, we
should set rdata/wdata->res.count properly.

When upper layer re-send the read/write request to finish the rest
part of the request, pgbase is the position where we should start at.

[pnfsblock: bl_write_pagelist support functions]
[pnfsblock: bl_write_pagelist adjust for missing PG_USE_PNFS]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: handle errors when read or write pagelist.]
Signed-off-by: Zhang Jingwang <yyalone@gmail.com>
[pnfs-block: use new write_pagelist api]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>

[SQUASHME: pnfsblock: mds_offset is set in the generic layer]
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>

[pnfsblock: mark IO error with NFS_LAYOUT_{RW|RO}_FAILED]
Signed-off-by: Peng Tao <peng_tao@emc.com>
[pnfsblock: SQUASHME: adjust to API change]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: fixup blksize alignment in bl_setup_layoutcommit]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[pnfsblock: bl_write_pagelist adjust for missing PG_USE_PNFS]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: handle errors when read or write pagelist.]
Signed-off-by: Zhang Jingwang <yyalone@gmail.com>
[pnfs-block: use new write_pagelist api]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: bl_read_pagelist
Fred Isaman [Sun, 31 Jul 2011 00:52:53 +0000]
pnfsblock: bl_read_pagelist

Note: When upper layer's read/write request cannot be fulfilled, the block
layout driver shouldn't silently mark the page as error. It should do
what can be done and  leave the rest to the upper layer. To do so, we
should set rdata/wdata->res.count properly.

When upper layer re-send the read/write request to finish the rest
part of the request, pgbase is the position where we should start at.

[pnfsblock: mark IO error with NFS_LAYOUT_{RW|RO}_FAILED]
Signed-off-by: Peng Tao <peng_tao@emc.com>
[pnfsblock: read path error handling]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: handle errors when read or write pagelist.]
Signed-off-by: Zhang Jingwang <yyalone@gmail.com>
[pnfs-block: use new read_pagelist api]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: cleanup_layoutcommit
Fred Isaman [Sun, 31 Jul 2011 00:52:52 +0000]
pnfsblock: cleanup_layoutcommit

In blocklayout driver. There are two things happening
while layoutcommit/cleanup.
1. the modified extents are encoded.
2. On cleanup the extents are put back on the layout rw
   extents list, for reads.

In the new system where actual xdr encoding is done in
encode_layoutcommit() directly into xdr buffer, these are
the new commit stages:

1. On setup_layoutcommit, the range is adjusted as before
   and a structure is allocated for communication with
   bl_encode_layoutcommit && bl_cleanup_layoutcommit
   (Generic layer provides a void-star to hang it on)

2. bl_encode_layoutcommit is called to do the actual
   encoding directly into xdr. The commit-extent-list is not
   freed and is stored on above structure.
   FIXME: The code is not yet converted to the new XDR cleanup

3. On cleanup the commit-extent-list is put back by a call
   to set_to_rw() as before, but with no need for XDR decoding
   of the list as before. And the commit-extent-list is freed.
   Finally allocated structure is freed.

[rm inode and pnfs_layout_hdr args from cleanup_layoutcommit()]
Signed-off-by: Jim Rees <rees@umich.edu>
[pnfsblock: introduce bl_committing list]
Signed-off-by: Peng Tao <peng_tao@emc.com>
[pnfsblock: SQUASHME: adjust to API change]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[blocklayout: encode_layoutcommit implementation]
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
[pnfsblock: fix bug setting up layoutcommit.]
Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
[pnfsblock: cleanup_layoutcommit wants a status parameter]
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: encode_layoutcommit
Fred Isaman [Sun, 31 Jul 2011 00:52:51 +0000]
pnfsblock: encode_layoutcommit

In blocklayout driver. There are two things happening
while layoutcommit/cleanup.
1. the modified extents are encoded.
2. On cleanup the extents are put back on the layout rw
   extents list, for reads.

In the new system where actual xdr encoding is done in
encode_layoutcommit() directly into xdr buffer, these are
the new commit stages:

1. On setup_layoutcommit, the range is adjusted as before
   and a structure is allocated for communication with
   bl_encode_layoutcommit && bl_cleanup_layoutcommit
   (Generic layer provides a void-star to hang it on)

2. bl_encode_layoutcommit is called to do the actual
   encoding directly into xdr. The commit-extent-list is not
   freed and is stored on above structure.
   FIXME: The code is not yet converted to the new XDR cleanup

3. On cleanup the commit-extent-list is put back by a call
   to set_to_rw() as before, but with no need for XDR decoding
   of the list as before. And the commit-extent-list is freed.
   Finally allocated structure is freed.

[rm inode and pnfs_layout_hdr args from cleanup_layoutcommit()]
[pnfsblock: get rid of deprecated xdr macros]
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[blocklayout: encode_layoutcommit implementation]
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
[pnfsblock: fix bug setting up layoutcommit.]
Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
[pnfsblock: prevent commit list corruption]
[pnfsblock: fix layoutcommit with an empty opaque]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: merge rw extents
Fred Isaman [Sun, 31 Jul 2011 00:52:50 +0000]
pnfsblock: merge rw extents

Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: add extent manipulation functions
Fred Isaman [Sun, 31 Jul 2011 00:52:49 +0000]
pnfsblock: add extent manipulation functions

Adds working implementations of various support functions
to handle INVAL extents, needed by writes, such as
bl_mark_sectors_init and bl_is_sector_init.

[pnfsblock: fix 64-bit compiler warnings for extent manipulation]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[Implement release_inval_marks]
Signed-off-by: Zhang Jingwang <zhangjingwang@nrchpc.ac.cn>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: bl_find_get_extent
Fred Isaman [Sun, 31 Jul 2011 00:52:48 +0000]
pnfsblock: bl_find_get_extent

Implement bl_find_get_extent(), one of the core extent manipulation
routines.

[pnfsblock: Lookup list entry of layouts and tags in reverse order]
Signed-off-by: Zhang Jingwang <zhangjingwang@nrchpc.ac.cn>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Jim Rees <rees@umich.edu>

pnfsblock: fix print format warnings for sector_t and size_t

gcc spews warnings about these on x86_64, e.g.:
fs/nfs/blocklayout/blocklayout.c:74: warning: format ‘%Lu’ expects type ‘long long unsigned int’, but argument 2 has type ‘sector_t’
fs/nfs/blocklayout/blocklayout.c:388: warning: format ‘%d’ expects type ‘int’, but argument 5 has type ‘size_t’

Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: xdr decode pnfs_block_layout4
Fred Isaman [Sun, 31 Jul 2011 00:52:47 +0000]
pnfsblock: xdr decode pnfs_block_layout4

XDR decodes the block layout payload sent in LAYOUTGET result, storing
the result in an extent list.

[pnfsblock: get rid of deprecated xdr macros]
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: fix bug getting pnfs_layout_type in translate_devid().]
Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: call and parse getdevicelist
Fred Isaman [Sun, 31 Jul 2011 00:52:46 +0000]
pnfsblock: call and parse getdevicelist

Call GETDEVICELIST during mount, then call and parse GETDEVICEINFO
for each device returned.

[pnfsblock: get rid of deprecated xdr macros]
Signed-off-by: Jim Rees <rees@umich.edu>
[pnfsblock: fix pnfs_deviceid references]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: fix print format warnings for sector_t and size_t]
[pnfs-block: #include <linux/vmalloc.h>]
[pnfsblock: no PNFS_NFS_SERVER]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
[pnfsblock: fix bug determining size of striped volume]
[pnfsblock: fix oops when using multiple devices]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[pnfsblock: get rid of vmap and deviceid->area structure]
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: merge extents
Fred Isaman [Sun, 31 Jul 2011 00:52:45 +0000]
pnfsblock: merge extents

Replace a stub, so that extents underlying the layouts are properly
added, merged, or ignored as necessary.

Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: delete the new node before put it]
Signed-off-by: Mingyang Guo <guomingyang@nrchpc.ac.cn>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: lseg alloc and free
Fred Isaman [Sun, 31 Jul 2011 00:52:44 +0000]
pnfsblock: lseg alloc and free

Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfsblock: fix bug getting pnfs_layout_type in translate_devid().]
Signed-off-by: Tao Guo <guotao@nrchpc.ac.cn>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Zhang Jingwang <Jingwang.Zhang@emc.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: remove device operations
Jim Rees [Sun, 31 Jul 2011 00:52:43 +0000]
pnfsblock: remove device operations

Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[upcall bugfixes]
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: add device operations
Jim Rees [Sun, 31 Jul 2011 00:52:42 +0000]
pnfsblock: add device operations

Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[upcall bugfixes]
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: basic extent code
Fred Isaman [Sun, 31 Jul 2011 00:52:41 +0000]
pnfsblock: basic extent code

Adds structures and basic create/delete code for extents.

Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Zhang Jingwang <Jingwang.Zhang@emc.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: use pageio_ops api
Benny Halevy [Sun, 31 Jul 2011 00:52:40 +0000]
pnfsblock: use pageio_ops api

[pnfsblock: use pnfs_generic_pg_init_read/write]
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfsblock: add blocklayout Kconfig option, Makefile, and stubs
Fred Isaman [Sun, 31 Jul 2011 00:52:39 +0000]
pnfsblock: add blocklayout Kconfig option, Makefile, and stubs

Define a configuration variable to enable/disable compilation of the
block driver code.

Add the minimal structure for a pnfs block layout driver, and empty
list-heads that will hold the extent data

[pnfsblock: make NFS_V4_1 select PNFS_BLOCK]
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
[pnfs-block: fix CONFIG_PNFS_BLOCK dependencies]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[pnfsblock: SQUASHME: adjust to API change]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfs: move pnfs_layout_type inline in nfs_inode]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
[blocklayout: encode_layoutcommit implementation]
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[pnfsblock: layout alloc and free]
Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
[pnfs: move pnfs_layout_type inline in nfs_inode]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[pnfsblock: define module alias]
Signed-off-by: Peng Tao <peng_tao@emc.com>
[rm inode and pnfs_layout_hdr args from cleanup_layoutcommit()]
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: cleanup_layoutcommit
Andy Adamson [Sun, 31 Jul 2011 00:52:38 +0000]
pnfs: cleanup_layoutcommit

This gives layout driver a chance to cleanup structures they put in at
encode_layoutcommit.

Signed-off-by: Andy Adamson <andros@netapp.com>
[fixup layout header pointer for layoutcommit]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
[rm inode and pnfs_layout_hdr args from cleanup_layoutcommit()]
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: ask for layout_blksize and save it in nfs_server
Fred Isaman [Sun, 31 Jul 2011 00:52:37 +0000]
pnfs: ask for layout_blksize and save it in nfs_server

Block layout needs it to determine IO size.

Signed-off-by: Fred Isaman <iisaman@citi.umich.edu>
Signed-off-by: Tao Guo <glorioustao@gmail.com>
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: add set-clear layoutdriver interface
Benny Halevy [Sun, 31 Jul 2011 00:52:36 +0000]
pnfs: add set-clear layoutdriver interface

To allow layout driver to issue getdevicelist at mount time, and clean up
at umount time.

[fixup non NFS_V4_1 set_pnfs_layoutdriver definition]
[pnfs: pass mntfh down the init_pnfs path]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: GETDEVICELIST
Andy Adamson [Sun, 31 Jul 2011 00:52:35 +0000]
pnfs: GETDEVICELIST

The block driver uses GETDEVICELIST

Signed-off-by: Andy Adamson <andros@netapp.com>
[pass struct nfs_server * to getdevicelist]
[get machince creds for getdevicelist]
[fix getdevicelist decode sizing]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>
Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: use lwb as layoutcommit length
Peng Tao [Sun, 31 Jul 2011 00:52:34 +0000]
pnfs: use lwb as layoutcommit length

Using NFS4_MAX_UINT64 will break current protocol.

[Needed in v3.0]
CC: Stable Tree <stable@kernel.org>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: let layoutcommit handle a list of lseg
Peng Tao [Sun, 31 Jul 2011 00:52:33 +0000]
pnfs: let layoutcommit handle a list of lseg

There can be multiple lseg per file, so layoutcommit should be
able to handle it.

[Needed in v3.0]
CC: Stable Tree <stable@kernel.org>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: save layoutcommit cred at layout header init
Peng Tao [Sun, 31 Jul 2011 00:52:32 +0000]
pnfs: save layoutcommit cred at layout header init

No need to save it for every lseg.
No need to save it at every pnfs_set_layoutcommit.

[Needed in v3.0]
CC: Stable Tree <stable@kernel.org>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agopnfs: save layoutcommit lwb at layout header
Peng Tao [Sun, 31 Jul 2011 00:52:31 +0000]
pnfs: save layoutcommit lwb at layout header

No need to save it for every lseg.

[Needed in v3.0]
CC: Stable Tree <stable@kernel.org>
Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Jim Rees <rees@umich.edu>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agoslab: use NUMA_NO_NODE
Andrew Morton [Thu, 28 Jul 2011 20:59:49 +0000]
slab: use NUMA_NO_NODE

Use the nice enumerated constant.

Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>

7 years agodon't busy retry the inode on failed grab_super_passive()
Wu Fengguang [Sat, 30 Jul 2011 04:14:35 +0000]
don't busy retry the inode on failed grab_super_passive()

This fixes a soft lockup on conditions

a) the flusher is working on a work by __bdi_start_writeback(), while

b) someone else calls writeback_inodes_sb*() or sync_inodes_sb(), which
   grab sb->s_umount and enqueue a new work for the flusher to execute

The s_umount grabbed by (b) will fail the grab_super_passive() in (a).
Then if the inode is requeued, wb_writeback() will busy retry on it.
As a result, wb_writeback() loops for ever without releasing
wb->list_lock, which further blocks other tasks.

Fix the busy loop by redirtying the inode. This may undesirably delay
the writeback of the inode, however most likely it will be picked up
soon by the queued work by writeback_inodes_sb*(), sync_inodes_sb() or
even writeback_inodes_wb().

bug url: http://www.spinics.net/lists/linux-fsdevel/msg47292.html

Reported-by: Christoph Hellwig <hch@infradead.org>
Tested-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>

7 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial
Linus Torvalds [Sat, 30 Jul 2011 20:51:11 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/trivial

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial:
  sungem: fix compile failure caused by trivial #include consolidation

7 years agoMerge branch 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groec...
Linus Torvalds [Sat, 30 Jul 2011 18:57:57 +0000]
Merge branch 'hwmon-for-linus' of git://git./linux/kernel/git/groeck/staging

* 'hwmon-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/staging: (24 commits)
  hwmon: (lm90) Refactor reading of config2 register
  hwmon: (lm90) Make SA56004 detection more robust
  hwmon: (lm90) Simplify handling of extended local temp register
  hwmon: (pmbus) Add client driver for LM25066, LM5064, and LM5066
  hwmon: (max34440) Add support for peak attributes
  hwmon: (max8688) Add support for peak attributes
  hwmon: (max16064) Add support for peak attributes
  hwmon: (adm1275) Add support for peak attributes
  hwmon: (pmbus) Add support for peak attributes
  hwmon: Add new attributes to sysfs ABI
  hwmon: (pmbus) Strengthen check for status register existence
  hwmon: (pmbus) Add support for virtual pages
  hwmon: (pmbus) Support reading and writing of word registers in device specific code
  hwmon: (pmbus) Increase attribute name size
  hwmon: (pmbus) Add ADP4000, NCP4200 and NCP4208 to list of supported devices
  hwmon: (pmbus) Add support for VID output voltage mode
  hwmon: (pmbus) Move PMBus drivers to drivers/hwmon/pmbus
  hwmon: (coretemp) Add core/pkg threshold support to Coretemp
  hwmon: (lm95241) Add support for LM95231
  hwmon: LM95245 driver
  ...

7 years agoshm: optimize locking and ipc_namespace getting
Vasiliy Kulikov [Thu, 28 Jul 2011 23:56:40 +0000]
shm: optimize locking and ipc_namespace getting

shm_lock() does a lookup of shm segment in shm_ids(ns).ipcs_idr, which
is redundant as we already know shmid_kernel address.  An actual lock is
also not required for reads until we really want to destroy the segment.

exit_shm() and shm_destroy_orphaned() may avoid the loop by checking
whether there is at least one segment in current ipc_namespace.

The check of nsproxy and ipc_ns against NULL is redundant as exit_shm()
is called from do_exit() before the call to exit_notify(), so the
dereferencing current->nsproxy->ipc_ns is guaranteed to be safe.

Reported-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

7 years agoshm: handle separate PID namespaces case
Vasiliy Kulikov [Thu, 28 Jul 2011 23:55:31 +0000]
shm: handle separate PID namespaces case

shm_try_destroy_orphaned() and shm_try_destroy_current() didn't handle
the case of separate PID namespaces, but a single IPC namespace.  If
there are tasks with the same PID values using the same shmem object,
the wrong destroy decision could be reached.

On shm segment creation store the pointer to the creator task in
shmid_kernel->shm_creator field and zero it on task exit.  Then
use the ->shm_creator insread of shm_cprid in both functions.  As
shmid_kernel object is already locked at this stage, no additional
locking is needed.

Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Acked-by: Serge Hallyn <serge.hallyn@canonical.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

7 years agoAdditional readdir cookie loop information
Bryan Schumaker [Fri, 29 Jul 2011 15:49:06 +0000]
Additional readdir cookie loop information

Print out the name of the file that triggers the cookie loop  message to
make it slightly easier to track down the cause.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6
Linus Torvalds [Sat, 30 Jul 2011 18:36:02 +0000]
Merge git://git./linux/kernel/git/jejb/scsi-misc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi-misc-2.6: (71 commits)
  [SCSI] fcoe: cleanup cpu selection for incoming requests
  [SCSI] fcoe: add fip retry to avoid missing critical keep alive
  [SCSI] libfc: fix warn on in lport retry
  [SCSI] libfc: Remove the reference to FCP packet from scsi_cmnd in case of error
  [SCSI] libfc: cleanup sending SRR request
  [SCSI] libfc: two minor changes in comments
  [SCSI] libfc, fcoe: ignore rx frame with wrong xid info
  [SCSI] libfc: release exchg cache
  [SCSI] libfc: use FC_MAX_ERROR_CNT
  [SCSI] fcoe: remove unused ptype field in fcoe_rcv_info
  [SCSI] bnx2fc: Update copyright and bump version to 1.0.4
  [SCSI] bnx2fc: Tx BDs cache in write tasks
  [SCSI] bnx2fc: Do not arm CQ when there are no CQEs
  [SCSI] bnx2fc: hold tgt lock when calling cmd_release
  [SCSI] bnx2fc: Enable support for sequence level error recovery
  [SCSI] bnx2fc: HSI changes for tape
  [SCSI] bnx2fc: Handle REC_TOV error code from firmware
  [SCSI] bnx2fc: REC/SRR link service request and response handling
  [SCSI] bnx2fc: Support 'sequence cleanup' task
  [SCSI] dh_rdac: Associate HBA and storage in rdac_controller to support partitions in storage
  ...

7 years agoNFS: Fix spurious readdir cookie loop messages
Trond Myklebust [Sat, 30 Jul 2011 16:45:35 +0000]
NFS: Fix spurious readdir cookie loop messages

If the directory contents change, then we have to accept that the
file->f_pos value may shrink if we do a 'search-by-cookie'. In that
case, we should turn off the loop detection and let the NFS client
try to recover.

The patch also fixes a second loop detection bug by ensuring
that after turning on the ctx->duped flag, we read at least one new
cookie into ctx->dir_cookie before attempting to match with
ctx->dup_cookie.

Reported-by: Petr Vandrovec <petr@vandrovec.name>
Cc: stable@kernel.org [2.6.39+]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

7 years agoMerge branch 'slub/lockless' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg...
Linus Torvalds [Sat, 30 Jul 2011 18:21:48 +0000]
Merge branch 'slub/lockless' of git://git./linux/kernel/git/penberg/slab-2.6

* 'slub/lockless' of git://git.kernel.org/pub/scm/linux/kernel/git/penberg/slab-2.6: (21 commits)
  slub: When allocating a new slab also prep the first object
  slub: disable interrupts in cmpxchg_double_slab when falling back to pagelock
  Avoid duplicate _count variables in page_struct
  Revert "SLUB: Fix build breakage in linux/mm_types.h"
  SLUB: Fix build breakage in linux/mm_types.h
  slub: slabinfo update for cmpxchg handling
  slub: Not necessary to check for empty slab on load_freelist
  slub: fast release on full slab
  slub: Add statistics for the case that the current slab does not match the node
  slub: Get rid of the another_slab label
  slub: Avoid disabling interrupts in free slowpath
  slub: Disable interrupts in free_debug processing
  slub: Invert locking and avoid slab lock
  slub: Rework allocator fastpaths
  slub: Pass kmem_cache struct to lock and freeze slab
  slub: explicit list_lock taking
  slub: Add cmpxchg_double_slab()
  mm: Rearrange struct page
  slub: Move page->frozen handling near where the page->freelist handling occurs
  slub: Do not use frozen page flag but a bit in the page counters
  ...

7 years agoMerge branch 'kconfig' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild-2.6
Linus Torvalds [Sat, 30 Jul 2011 07:17:06 +0000]
Merge branch 'kconfig' of git://git./linux/kernel/git/mmarek/kbuild-2.6

* 'kconfig' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild-2.6: (25 commits)
  kconfig: Introduce IS_ENABLED(), IS_BUILTIN() and IS_MODULE()
  xconfig: Abort close if configuration cannot be saved
  kconfig: fix missing "0x" prefix from S_HEX symbol in autoconf.h
  kconfig/nconf: remove useless conditionnal
  kconfig/nconf: prevent segfault on empty menu
  kconfig/nconf: use the generic menu_get_ext_help()
  nconfig: Avoid Wunused-but-set warning
  kconfig/conf: mark xfgets() private
  kconfig: remove pending prototypes for kconfig_load()
  kconfig/conf: add command line options' description
  kconfig/conf: reduce the scope of `defconfig_file'
  kconfig: use calloc() for expr allocation
  kconfig: introduce specialized printer
  kconfig: do not overwrite symbol direct dependency in assignment
  kconfig/gconf: silent missing prototype warnings
  kconfig/gconf: kill deadcode
  kconfig: nuke LKC_DIRECT_LINK cruft
  kconfig: nuke reference to SWIG
  kconfig: add missing <stdlib.h> inclusion
  kconfig: add missing <ctype.h> inclusion
  ...

Fix up conflicts in scripts/kconfig/Makefile

7 years agoMerge branch 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Sat, 30 Jul 2011 07:08:53 +0000]
Merge branch 'v4l_for_linus' of git://git./linux/kernel/git/mchehab/linux-2.6

* 'v4l_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-2.6: (430 commits)
  [media] ir-mce_kbd-decoder: include module.h for its facilities
  [media] ov5642: include module.h for its facilities
  [media] em28xx: Fix DVB-C maxsize for em2884
  [media] tda18271c2dd: Fix saw filter configuration for DVB-C @6MHz
  [media] v4l: mt9v032: Fix Bayer pattern
  [media] V4L: mt9m111: rewrite set_pixfmt
  [media] V4L: mt9m111: fix missing return value check mt9m111_reg_clear
  [media] V4L: initial driver for ov5642 CMOS sensor
  [media] V4L: sh_mobile_ceu_camera: fix Oops when USERPTR mapping fails
  [media] V4L: soc-camera: remove soc-camera bus and devices on it
  [media] V4L: soc-camera: un-export the soc-camera bus
  [media] V4L: sh_mobile_csi2: switch away from using the soc-camera bus notifier
  [media] V4L: add media bus configuration subdev operations
  [media] V4L: soc-camera: group struct field initialisations together
  [media] V4L: soc-camera: remove now unused soc-camera specific PM hooks
  [media] V4L: pxa-camera: switch to using standard PM hooks
  [media] NetUP Dual DVB-T/C CI RF: force card hardware revision by module param
  [media] Don't OOPS if videobuf_dvb_get_frontend return NULL
  [media] NetUP Dual DVB-T/C CI RF: load firmware according card revision
  [media] omap3isp: Support configurable HS/VS polarities
  ...

Fix up conflicts:
 - arch/arm/mach-omap2/board-rx51-peripherals.c:
     cleanup regulator supply definitions in mach-omap2
   vs
     OMAP3: RX-51: define vdds_csib regulator supply
 - drivers/staging/tm6000/tm6000-alsa.c (trivial)

7 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Sat, 30 Jul 2011 06:45:06 +0000]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: Fix build breakage in xfs_iops.c when CONFIG_FS_POSIX_ACL is not set

7 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/ecryptfs/ecryptf...
Linus Torvalds [Sat, 30 Jul 2011 06:43:50 +0000]
Merge branch 'next' of git://git./linux/kernel/git/ecryptfs/ecryptfs-2.6

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/ecryptfs/ecryptfs-2.6:
  ecryptfs: Make inode bdi consistent with superblock bdi
  eCryptfs: Unlock keys needed by ecryptfsd

7 years agoMerge branch 'stable/bug.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 30 Jul 2011 06:43:32 +0000]
Merge branch 'stable/bug.fixes' of git://git./linux/kernel/git/konrad/xen

* 'stable/bug.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen:
  xen/balloon: Fix compile errors - missing header files.
  xen/grant: Fix compile warning.
  xen/pciback: remove duplicated #include

7 years agofrv: remove unnecessary code
Greg Dietsche [Fri, 29 Jul 2011 15:46:16 +0000]
frv: remove unnecessary code

remove unnecessary code that matches this coccinelle pattern

if (...)
return ret;
return ret;

Signed-off-by: Greg Dietsche <Gregory.Dietsche@cuw.edu>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

7 years agoramoops: fix compile failure on parisc
James Bottomley [Fri, 29 Jul 2011 13:11:32 +0000]
ramoops: fix compile failure on parisc

Fixes this:

  drivers/char/ramoops.c: In function 'ramoops_init':
  drivers/char/ramoops.c:221: error: implicit declaration of function 'IS_ERR'
  drivers/char/ramoops.c:222: error: implicit declaration of function 'PTR_ERR'

If it actually builds on other platforms, it's probably getting
linux/err.h via some other #include.

Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

7 years agoMerge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes...
Linus Torvalds [Sat, 30 Jul 2011 06:35:05 +0000]
Merge branch 'linux-next' of git://git./linux/kernel/git/jbarnes/pci-2.6

* 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
  PCI: remove printks about disabled bridge windows
  PCI: fold pci_calc_resource_flags() into decode_bar()
  PCI: treat mem BAR type "11" (reserved) as 32-bit, not 64-bit, BAR
  PCI: correct pcie_set_readrq write size
  PCI: pciehp: change wait time for valid configuration access
  x86/PCI: Preserve existing pci=bfsort whitelist for Dell systems
  PCI: ARI is a PCIe v2 feature
  x86/PCI: quirks: Use pci_dev->revision
  PCI: Make the struct pci_dev * argument of pci_fixup_irqs const.
  PCI hotplug: cpqphp: use pci_dev->vendor
  PCI hotplug: cpqphp: use pci_dev->subsystem_{vendor|device}
  x86/PCI: config space accessor functions should not ignore the segment argument
  PCI: Assign values to 'pci_obff_signal_type' enumeration constants
  x86/PCI: reduce severity of host bridge window conflict warnings
  PCI: enumerate the PCI device only removed out PCI hieratchy of OS when re-scanning PCI
  PCI: PCIe AER: add aer_recover_queue
  x86/PCI: select direct access mode for mmconfig option
  PCI hotplug: Rename is_ejectable which also exists in dock.c

7 years agoMerge branch 'upstream/xen-tracing2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 30 Jul 2011 06:33:40 +0000]
Merge branch 'upstream/xen-tracing2' of git://git./linux/kernel/git/jeremy/xen

* 'upstream/xen-tracing2' of git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen:
  xen/tracing: fix compile errors when tracing is disabled.

7 years agoMerge branch 'at91/cleanup' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Sat, 30 Jul 2011 06:32:53 +0000]
Merge branch 'at91/cleanup' of git://git./linux/kernel/git/arm/linux-arm-soc

* 'at91/cleanup' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux-arm-soc:
  at91: add arch specific ioremap support
  at91: factorize sram init
  at91: move register clocks to soc generic init
  at91: move clock subsystem init to soc generic init
  at91: use structure to store the current soc
  at91: remove AT91_DBGU offset from dbgu register macro
  at91: factorize at91 interrupts init to soc
  at91: introduce commom AT91_BASE_SYS

7 years agoMerge branch 'next/dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux...
Linus Torvalds [Sat, 30 Jul 2011 06:32:02 +0000]
Merge branch 'next/dt' of git://git./linux/kernel/git/arm/linux-arm-soc

* 'next/dt' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux-arm-soc: (21 commits)
  arm/dt: tegra devicetree support
  arm/versatile: Add device tree support
  dt/irq: add irq_domain_generate_simple() helper
  irq: add irq_domain translation infrastructure
  dmaengine: imx-sdma: add device tree probe support
  dmaengine: imx-sdma: sdma_get_firmware does not need to copy fw_name
  dmaengine: imx-sdma: use platform_device_id to identify sdma version
  mmc: sdhci-esdhc-imx: add device tree probe support
  mmc: sdhci-pltfm: dt device does not pass parent to sdhci_alloc_host
  mmc: sdhci-esdhc-imx: get rid of the uses of cpu_is_mx()
  mmc: sdhci-esdhc-imx: do not reference platform data after probe
  mmc: sdhci-esdhc-imx: extend card_detect and write_protect support for mx5
  net/fec: add device tree probe support
  net: ibm_newemac: convert it to use of_get_phy_mode
  dt/net: add helper function of_get_phy_mode
  net/fec: gasket needs to be enabled for some i.mx
  serial/imx: add device tree probe support
  serial/imx: get rid of the uses of cpu_is_mx1()
  arm/dt: Add dtb make rule
  arm/dt: Add skeleton dtsi file
  ...

7 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux-arm-soc
Linus Torvalds [Sat, 30 Jul 2011 06:31:22 +0000]
Merge branch 'fixes' of git://git./linux/kernel/git/arm/linux-arm-soc

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/linux-arm-soc:
  ARM: zynq: remove incorrectly deleted file
  ARM: tegra: only select MACH_HAS_SND_SOC_TEGRA_WM8903 if SND_SOC

7 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-watchdog
Linus Torvalds [Sat, 30 Jul 2011 06:28:47 +0000]
Merge git://git./linux/kernel/git/wim/linux-watchdog

* git://git.kernel.org/pub/scm/linux/kernel/git/wim/linux-watchdog: (25 commits)
  watchdog: WatchDog Timer Driver Core - Add minimum and max timeout
  watchdog: WatchDog Timer Driver Core - Add ioctl call
  watchdog: WatchDog Timer Driver Core - Add nowayout feature
  watchdog: WatchDog Timer Driver Core - Add Magic Close feature
  watchdog: WatchDog Timer Driver Core - Add WDIOC_SETTIMEOUT and WDIOC_GETTIMEOUT ioctl
  watchdog: WatchDog Timer Driver Core - Add WDIOC_SETOPTIONS ioctl
  watchdog: WatchDog Timer Driver Core - Add WDIOC_KEEPALIVE ioctl
  watchdog: WatchDog Timer Driver Core - Add basic ioctl functionality
  watchdog: WatchDog Timer Driver Core - Add basic framework
  watchdog: hpwdt: add next gen HP servers
  watchdog: it8712f_wdt.c: improve includes
  watchdog: at91sam9/wdt: move register header to drivers
  watchdog: Add Xilinx watchdog timer driver
  watchdog: remove empty pm-functions
  watchdog: sp805: Flush posted writes in enable/disable.
  watchdog: sp805: Don't write 0 to the load value register.
  watchdog: imx2_wdt: add device tree probe support
  watchdog: s3c2410: Add support for device tree based probe
  watchdog: mpcore_wdt: Add suspend/resume support.
  watchdog: mtx1-wdt: use dev_{err,info} instead of printk()
  ...

7 years agoext2: remove duplicate 'ext2_get_acl()' define
Linus Torvalds [Sat, 30 Jul 2011 06:21:50 +0000]
ext2: remove duplicate 'ext2_get_acl()' define

When commit 4e34e719e457 ("fs: take the ACL checks to common code")
changed the xyz_check_acl() functions into the more natural
xyz_get_acl() interface, we grew two copies of the

  #define ext2_get_acl NULL

define for the non-acl case.

Remove the extra one.

Reported-by: Marco Stornelli <marco.stornelli@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

7 years agokconfig: Introduce IS_ENABLED(), IS_BUILTIN() and IS_MODULE()
Michal Marek [Wed, 20 Jul 2011 15:38:57 +0000]
kconfig: Introduce IS_ENABLED(), IS_BUILTIN() and IS_MODULE()

Replace the config_is_*() macros with a variant that allows for grepping
for usage of CONFIG_* options in the code. Usage:

  if (IS_ENABLED(CONFIG_NUMA))

or

  #if IS_ENABLED(CONFIG_NUMA)

The IS_ENABLED() macro evaluates to 1 if the argument is set (to either 'y'
or 'm'), IS_BUILTIN() tests if the option is 'y' and IS_MODULE() test if
the option is 'm'. Only boolean and tristate options are supported.

Reviewed-by: Arnaud Lacombe <lacombar@gmail.com>
Acked-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Michal Marek <mmarek@suse.cz>

7 years agocpupower: Do detect IDA (opportunistic processor performance) via cpuid
Thomas Renninger [Thu, 21 Jul 2011 09:54:54 +0000]
cpupower: Do detect IDA (opportunistic processor performance) via cpuid

IA32-Intel Devel guide Volume 3A - 14.3.2.1
-------------------------------------------
...
Opportunistic processor performance operation can be disabled by setting bit 38 of
IA32_MISC_ENABLES. This mechanism is intended for BIOS only. If
IA32_MISC_ENABLES[38] is set, CPUID.06H:EAX[1] will return 0.

Better detect things via cpuid, this cleans up the code a bit
and the MSR parts were not working correctly anyway.

Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: lenb@kernel.org
CC: linux@dominikbrodowski.net
CC: cpufreq@vger.kernel.org
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupower: Show Intel turbo ratio support via ./cpupower frequency-info
Thomas Renninger [Thu, 21 Jul 2011 09:54:53 +0000]
cpupower: Show Intel turbo ratio support via ./cpupower frequency-info

This adds the last piece missing from turbostat (if called with -v).
It shows on Intel machines supporting Turbo Boost how many cores
have to be active/idle to enter which boost mode (frequency).

Whether the HW really enters these boost modes can be verified via
./cpupower monitor.

Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: lenb@kernel.org
CC: linux@dominikbrodowski.net
CC: cpufreq@vger.kernel.org
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agoxfs: Fix build breakage in xfs_iops.c when CONFIG_FS_POSIX_ACL is not set
Markus Trippelsdorf [Tue, 26 Jul 2011 09:15:20 +0000]
xfs: Fix build breakage in xfs_iops.c when CONFIG_FS_POSIX_ACL is not set

commit 4e34e719e45, that takes the ACL checks to common code,
accidentely broke the build when CONFIG_FS_POSIX_ACL is not set:

CC fs/xfs/linux-2.6/xfs_iops.o
fs/xfs/linux-2.6/xfs_iops.c:1025:14: error: ‘xfs_get_acl’ undeclared here (not in a function)

Fix this by declaring xfs_get_acl a static inline function.

Signed-off-by: Markus Trippelsdorf <markus@trippelsdorf.de>
Signed-off-by: Alex Elder <aelder@sgi.com>

7 years agocpupowerutils: increase MAX_LINE_LEN
Roman Vasiyarov [Mon, 25 Apr 2011 17:34:23 +0000]
cpupowerutils: increase MAX_LINE_LEN

larger sysfs data (>255 bytes) was truncated and thus used improperly

[linux@dominikbrodowski.net: adapted to cpupowerutils]
Signed-off-by: Roman Vasiyarov <rvasiyarov@gmail.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupower: Rename package from cpupowerutils to cpupower
Thomas Renninger [Thu, 21 Apr 2011 15:50:26 +0000]
cpupower: Rename package from cpupowerutils to cpupower

Signed-off-by: Thomas Renninger <trenn@suse.de>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: Rename: libcpufreq->libcpupower
Thomas Renninger [Thu, 21 Apr 2011 15:50:25 +0000]
cpupowerutils: Rename: libcpufreq->libcpupower

[linux@dominikbrodowski.net: fix .gitignore]
Signed-off-by: Thomas Renninger <trenn@suse.de>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: use kernel version-derived version string
Dominik Brodowski [Wed, 20 Apr 2011 18:01:39 +0000]
cpupowerutils: use kernel version-derived version string

As cpupowerutils is intended to be included into the kernel sources,
use the kernel versioning instead of a custom version.

The script utils/version-gen.sh is largely based on the script already
found in tools/perf/util/PERF-VERSION-GEN .

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: utils - ConfigStyle bugfixes
Dominik Brodowski [Tue, 19 Apr 2011 18:33:50 +0000]
cpupowerutils: utils - ConfigStyle bugfixes

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: helpers - ConfigStyle bugfixes
Dominik Brodowski [Tue, 19 Apr 2011 18:16:05 +0000]
cpupowerutils: helpers - ConfigStyle bugfixes

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: idle_monitor - ConfigStyle bugfixes
Dominik Brodowski [Tue, 19 Apr 2011 17:58:59 +0000]
cpupowerutils: idle_monitor - ConfigStyle bugfixes

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: lib - ConfigStyle bugfixes
Dominik Brodowski [Tue, 19 Apr 2011 17:42:33 +0000]
cpupowerutils: lib - ConfigStyle bugfixes

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: bench - ConfigStyle bugfixes
Dominik Brodowski [Tue, 19 Apr 2011 17:20:12 +0000]
cpupowerutils: bench - ConfigStyle bugfixes

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: do not update po files on each and every compile
Dominik Brodowski [Tue, 19 Apr 2011 16:52:44 +0000]
cpupowerutils: do not update po files on each and every compile

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: remove ccdv, use kernel quiet/verbose mechanism
Dominik Brodowski [Tue, 19 Apr 2011 07:03:52 +0000]
cpupowerutils: remove ccdv, use kernel quiet/verbose mechanism

Use the quiet/verbose mechanism found in kernel tools, without
relying on the special tool "ccdv"

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils: use COPYING, CREDITS from top-level directory
Dominik Brodowski [Tue, 12 Apr 2011 20:50:19 +0000]
cpupowerutils: use COPYING, CREDITS from top-level directory

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agocpupowerutils - cpufrequtils extended with quite some features
Dominik Brodowski [Wed, 30 Mar 2011 14:30:11 +0000]
cpupowerutils - cpufrequtils extended with quite some features

CPU power consumption vs performance tuning is no longer
limited to CPU frequency switching anymore: deep sleep states,
traditional dynamic frequency scaling and hidden turbo/boost
frequencies are tied close together and depend on each other.
The first two exist on different architectures like PPC, Itanium and
ARM, the latter (so far) only on X86. On X86 the APU (CPU+GPU) will
only run most efficiently if CPU and GPU has proper power management
in place.

Users and Developers want to have *one* tool to get an overview what
their system supports and to monitor and debug CPU power management
in detail. The tool should compile and work on as many architectures
as possible.

Once this tool stabilizes a bit, it is intended to replace the
Intel-specific tools in tools/power/x86

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agosmc91c92_cs.c: fix bogus compiler warning
Dominik Brodowski [Fri, 29 Jul 2011 16:27:34 +0000]
smc91c92_cs.c: fix bogus compiler warning

On a recent compile, I got this warning:

CC [M] drivers/net/pcmcia/smc91c92_cs.o
drivers/net/pcmcia/smc91c92_cs.c: In function ‘smc91c92_probe’:
drivers/net/pcmcia/smc91c92_cs.c:812:12: warning: ‘j’ may be used uninitialized in this function

However, "j" is only used in a branch which has the same condition as
a previous branch, where j is set, e.g.

int j;

if (CONDITION)
    j = VALUE

...

if (CONDITION)
   printk(j)

Still, avoid this warning, as it is easy to circumvent.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agoorinoco_cs: be more careful when matching cards with ID 0x0156:0x0002
Pavel Roskin [Tue, 26 Jul 2011 22:52:48 +0000]
orinoco_cs: be more careful when matching cards with ID 0x0156:0x0002

Without CONFIG_HERMES_PRISM, only match cards that have "Version 01.01"
as the third product ID.  Those have Agere firmware.

With CONFIG_HERMES_PRISM, match all 0x0156:0x0002 cards.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agohostap_cs: support cards with "Version 01.02" as third product ID
Pavel Roskin [Tue, 26 Jul 2011 22:52:41 +0000]
hostap_cs: support cards with "Version 01.02" as third product ID

Cards with numeric ID 0x0156:0x0002 and third ID "Version 01.02" can be
assumed to have Intersil firmware.  Cards with Agere firmware use
"Version 01.01".

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agopcmcia: add PCMCIA_DEVICE_MANF_CARD_PROD_ID3
Pavel Roskin [Tue, 26 Jul 2011 22:52:35 +0000]
pcmcia: add PCMCIA_DEVICE_MANF_CARD_PROD_ID3

This is needed to match wireless cards with Intersil firmware that have
ID 0x0156:0x0002 and the third ID "Version 01.02".  Such cards are
currently matched by orinoco_cs, which doesn't support WPA.  They should
be matched by hostap_cs.

The first and the second product ID vary widely, so there are few users
with some particular IDs.  Of those, very few can submit a patch for
hostap_cs or write a useful bugreport.  It's still important to support
their hardware properly.

With PCMCIA_DEVICE_MANF_CARD_PROD_ID3, it should be possible to cover
the remaining Intersil based designs that kept the numeric ID and the
"version" of the reference design.

Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>

7 years agopxa2xx pcmcia - stargate 2 use gpio array.
Jonathan Cameron [Wed, 13 Jul 2011 14:54:58 +0000]
pxa2xx pcmcia - stargate 2 use gpio array.

Kill off the buff gpio as not used anywhere.

Signed-off-by: Jonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>