deliverable/linux.git
10 years agocciss: Fallback to MSI rather than to INTx if MSI-X failed
Alexander Gordeev [Wed, 26 Feb 2014 09:02:41 +0000 (10:02 +0100)] 
cciss: Fallback to MSI rather than to INTx if MSI-X failed

Currently the driver falls back to INTx mode when MSI-X
initialization failed. This is a suboptimal behaviour
for chips that also support MSI. This update changes that
behaviour and falls back to MSI mode in case MSI-X mode
initialization failed.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Mike Miller <mike.miller@hp.com>
Cc: iss_storagedev@hp.com
Cc: Jens Axboe <axboe@kernel.dk>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoswim3: fix interruptible_sleep_on race
Arnd Bergmann [Wed, 26 Feb 2014 11:01:44 +0000 (12:01 +0100)] 
swim3: fix interruptible_sleep_on race

interruptible_sleep_on is racy and going away. This replaces the one
caller in the swim3 driver with the equivalent race-free
wait_event_interruptible call. Since we're here already, this
also fixes the case where we get interrupted from atomic context,
which used to just spin in the loop.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoataflop: fix sleep_on races
Arnd Bergmann [Wed, 26 Feb 2014 11:01:41 +0000 (12:01 +0100)] 
ataflop: fix sleep_on races

sleep_on() is inherently racy, and has been deprecated for a long time.
This fixes two instances in the atari floppy driver:

* fdc_wait/fdc_busy becomes an open-coded mutex. We cannot use the
  regular mutex since it gets released in interrupt context. The
  open-coded version using wait_event() and cmpxchg() is equivalent
  to the existing code but does the checks atomically, and we can
  now safely check the condition with irqs enabled.

* format_wait becomes a completion, which is the natural structure
  here. The format ioctl waits for the background task to either
  complete or abort.

This does not attempt to fix the preexisting bug of calling schedule
with local interrupts disabled.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Michael Schmitz <schmitz@biophys.uni-duesseldorf.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoDAC960: remove sleep_on usage
Arnd Bergmann [Wed, 26 Feb 2014 11:01:43 +0000 (12:01 +0100)] 
DAC960: remove sleep_on usage

sleep_on and its variants are going away. The use of sleep_on() in
DAC960_V2_ExecuteUserCommand seems to be bogus because the command
by the time we get there, the command has completed already and
we just enter the timeout. Based on this interpretation, I concluded
that we can replace it with a simple msleep(1000) and rearrange the
code around it slightly.

The interruptible_sleep_on_timeout in DAC960_gam_ioctl seems equivalent
to the race-free version using wait_event_interruptible_timeout.
I left the driver to return -EINTR rather than -ERESTARTSYS to preserve
the timeout behavior.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agomtip32xx: Use pci_enable_msi() instead of pci_enable_msi_range()
Alexander Gordeev [Tue, 25 Feb 2014 21:33:18 +0000 (22:33 +0100)] 
mtip32xx: Use pci_enable_msi() instead of pci_enable_msi_range()

Commit "mtip32xx: Use pci_enable_msix_range() instead of
pci_enable_msix()" was unnecessary, since pci_enable_msi()
function is not deprecated and is still preferable for
enabling the single MSI mode. This update reverts usage of
pci_enable_msi() function.

Besides, the changelog for that commit was bogus, since
mtip32xx driver uses MSI interrupt, not MSI-X.

Cc: Jens Axboe <axboe@kernel.dk>
Cc: Asai Thambi S P <asamymuthupa@micron.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoskd: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Wed, 19 Feb 2014 08:58:21 +0000 (09:58 +0100)] 
skd: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoskd: Use unified access to skdev->msix_entries throughout the code
Alexander Gordeev [Wed, 19 Feb 2014 08:58:20 +0000 (09:58 +0100)] 
skd: Use unified access to skdev->msix_entries throughout the code

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoskd: Fix incomplete cleanup of MSI-X interrupt
Alexander Gordeev [Wed, 19 Feb 2014 08:58:19 +0000 (09:58 +0100)] 
skd: Fix incomplete cleanup of MSI-X interrupt

When enabling MSI-X interrupts fails due to lack of memory
the call to pci_disable_msix() is missed and the device is
left with MSI-X interrupts enabled while the driver assumes
otherwise. This update fixes the described misbehaviour and
cleans up the code of skd_release_msix() function.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agoskd: Fix out of array boundary access
Alexander Gordeev [Wed, 19 Feb 2014 08:58:18 +0000 (09:58 +0100)] 
skd: Fix out of array boundary access

When enabling MSI-X, interrupts are requested for SKD_MAX_MSIX_COUNT
entries in skdev->msix_entries array, while the number of actually
allocated entries is skdev->msix_count. This might lead to an out of
boundary access in case number of allocated entries is less than
SKD_MAX_MSIX_COUNT. This update fixes the described misbehaviour.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agomtip32xx: Use pci_enable_msix_range() instead of pci_enable_msix()
Alexander Gordeev [Wed, 19 Feb 2014 08:58:16 +0000 (09:58 +0100)] 
mtip32xx: Use pci_enable_msix_range() instead of pci_enable_msix()

As result of deprecation of MSI-X/MSI enablement functions
pci_enable_msix() and pci_enable_msi_block() all drivers
using these two interfaces need to be updated to use the
new pci_enable_msi_range() and pci_enable_msix_range()
interfaces.

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Asai Thambi S P <asamymuthupa@micron.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agomtip32xx: Remove superfluous call to pci_disable_msi()
Alexander Gordeev [Wed, 19 Feb 2014 08:58:15 +0000 (09:58 +0100)] 
mtip32xx: Remove superfluous call to pci_disable_msi()

There is no need to call pci_disable_msi() in case
the previous call to pci_enable_msi() failed

Signed-off-by: Alexander Gordeev <agordeev@redhat.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Asai Thambi S P <asamymuthupa@micron.com>
Cc: linux-pci@vger.kernel.org
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agodrbd: Fix future possible NULL pointer dereference
Andreas Gruenbacher [Wed, 19 Feb 2014 09:49:07 +0000 (10:49 +0100)] 
drbd: Fix future possible NULL pointer dereference

Right now every resource has exactly one connection. But we are preparing
for dynamic connections. I.e. in the future thre can be resources without
connections.

However smatch points this out as 'variable dereferenced before check',
which is correct.

This issue was introduced in
drbd: get_one_status(): Iterate over resource->devices instead of connection->peer_devices

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
10 years agodrbd: Add drbd_thread->resource and make drbd_thread->connection optional
Andreas Gruenbacher [Thu, 21 Jul 2011 11:45:21 +0000 (13:45 +0200)] 
drbd: Add drbd_thread->resource and make drbd_thread->connection optional

In the drbd_thread "infrastructure" functions, only use the resource instead of
the connection.  Make the connection field of drbd_thread optional.  This will
allow to introduce threads which are not associated with a connection.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Use the right peer device
Andreas Gruenbacher [Tue, 13 Sep 2011 08:39:41 +0000 (10:39 +0200)] 
drbd: Use the right peer device

in w_e_ (peer request) callbacks and in peer request I/O completion handlers

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Remove unused parameter of wire_flags_to_bio()
Andreas Gruenbacher [Tue, 30 Aug 2011 14:22:33 +0000 (16:22 +0200)] 
drbd: Remove unused parameter of wire_flags_to_bio()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Get rid of first_peer_device() in handle_write_conflicts()
Andreas Gruenbacher [Tue, 30 Aug 2011 13:38:04 +0000 (15:38 +0200)] 
drbd: Get rid of first_peer_device() in handle_write_conflicts()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: In the worker thread, process drbd_work instead of drbd_device_work items
Andreas Gruenbacher [Fri, 26 Aug 2011 21:50:08 +0000 (23:50 +0200)] 
drbd: In the worker thread, process drbd_work instead of drbd_device_work items

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Turn w_make_ov_request and make_resync_request into "normal" functions
Andreas Gruenbacher [Thu, 25 Aug 2011 14:59:58 +0000 (16:59 +0200)] 
drbd: Turn w_make_ov_request and make_resync_request into "normal" functions

These functions are not used as drbd_work callbacks.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Make w_make_resync_request() static
Andreas Gruenbacher [Thu, 25 Aug 2011 14:56:34 +0000 (16:56 +0200)] 
drbd: Make w_make_resync_request() static

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: struct drbd_peer_request: Use drbd_work instead of drbd_device_work
Andreas Gruenbacher [Thu, 25 Aug 2011 13:49:40 +0000 (15:49 +0200)] 
drbd: struct drbd_peer_request: Use drbd_work instead of drbd_device_work

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: struct after_conn_state_chg_work: Use drbd_work instead of drbd_device_work
Andreas Gruenbacher [Thu, 25 Aug 2011 13:14:48 +0000 (15:14 +0200)] 
drbd: struct after_conn_state_chg_work: Use drbd_work instead of drbd_device_work

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Turn conn_flush_workqueue() into drbd_flush_workqueue()
Andreas Gruenbacher [Thu, 28 Jul 2011 13:56:02 +0000 (15:56 +0200)] 
drbd: Turn conn_flush_workqueue() into drbd_flush_workqueue()

The new function can flush any work queue, not just the work queue of the data
socket of a connection.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Create a dedicated struct drbd_device_work
Andreas Gruenbacher [Thu, 28 Jul 2011 13:27:51 +0000 (15:27 +0200)] 
drbd: Create a dedicated struct drbd_device_work

drbd_device_work is a work item that has a reference to a device,
while drbd_work is a more generic work item that does not carry
a reference to a device.

All callbacks get a pointer to a drbd_work instance, those callbacks
that expect a drbd_device_work use the container_of macro to get it.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename w_prev_work_done -> w_complete
Andreas Gruenbacher [Mon, 25 Jul 2011 11:51:18 +0000 (13:51 +0200)] 
drbd: Rename w_prev_work_done -> w_complete

Also move it to drbd_receiver.c and make it static.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Move string function prototypes from linux/drbd.h to drbd_string.h
Andreas Gruenbacher [Thu, 1 Sep 2011 11:18:31 +0000 (13:18 +0200)] 
drbd: Move string function prototypes from linux/drbd.h to drbd_string.h

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Remove useless assertion
Andreas Gruenbacher [Wed, 24 Aug 2011 06:19:41 +0000 (08:19 +0200)] 
drbd: Remove useless assertion

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Kill drbd_task_to_thread_name()
Andreas Gruenbacher [Wed, 10 Aug 2011 13:05:02 +0000 (15:05 +0200)] 
drbd: Kill drbd_task_to_thread_name()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Pass a peer device to a number of fuctions
Andreas Gruenbacher [Mon, 8 Aug 2011 22:47:13 +0000 (00:47 +0200)] 
drbd: Pass a peer device to a number of fuctions

These functions actually operate on a peer device, or
need a peer device.

drbd_prepare_command(), drbd_send_command(), drbd_send_sync_param()
drbd_send_uuids(), drbd_gen_and_send_sync_uuid(), drbd_send_sizes()
drbd_send_state(), drbd_send_current_state(), and drbd_send_state_req()
drbd_send_sr_reply(), drbd_send_ack(), drbd_send_drequest(),
drbd_send_drequest_csum(), drbd_send_ov_request(), drbd_send_dblock()
drbd_send_block(), drbd_send_out_of_sync(), recv_dless_read()
drbd_drain_block(), receive_bitmap_plain(), recv_resync_read()
read_in_block(), read_for_csum(), drbd_alloc_pages(), drbd_alloc_peer_req()
need_peer_seq(), update_peer_seq(), wait_for_and_update_peer_seq()
drbd_sync_handshake(), drbd_asb_recover_{0,1,2}p(), drbd_connected()
drbd_disconnected(), decode_bitmap_c() and recv_bm_rle_bits()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Replace vnr_to_mdev() with conn_peer_device()
Andreas Gruenbacher [Tue, 9 Aug 2011 01:54:55 +0000 (03:54 +0200)] 
drbd: Replace vnr_to_mdev() with conn_peer_device()

The new function returns a peer device, which allows us to eliminate a few
instances of first_peer_device().

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: drbd_csum_bio(), drbd_csum_ee(): Remove unused device argument
Andreas Gruenbacher [Tue, 9 Aug 2011 00:49:01 +0000 (02:49 +0200)] 
drbd: drbd_csum_bio(), drbd_csum_ee(): Remove unused device argument

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Function prototype cleanups
Andreas Gruenbacher [Fri, 22 Jul 2011 09:14:41 +0000 (11:14 +0200)] 
drbd: Function prototype cleanups

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename drbdd_init() -> drbd_receiver()
Andreas Gruenbacher [Fri, 22 Jul 2011 09:04:36 +0000 (11:04 +0200)] 
drbd: Rename drbdd_init() -> drbd_receiver()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Move cpu_mask from connection to resource
Andreas Gruenbacher [Fri, 22 Jul 2011 12:29:02 +0000 (14:29 +0200)] 
drbd: Move cpu_mask from connection to resource

Also fix drbd_calc_cpu_mask() to spread resources equally over all online cpus
independent of device minor numbers.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Define the size of res_opts->cpu_mask in a single place
Andreas Gruenbacher [Fri, 22 Jul 2011 11:53:19 +0000 (13:53 +0200)] 
drbd: Define the size of res_opts->cpu_mask in a single place

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Move susp, susp_nod, susp_fen from connection to resource
Andreas Gruenbacher [Thu, 7 Jul 2011 23:19:44 +0000 (01:19 +0200)] 
drbd: Move susp, susp_nod, susp_fen from connection to resource

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Move conf_mutex from connection to resource
Andreas Gruenbacher [Thu, 7 Jul 2011 12:19:42 +0000 (14:19 +0200)] 
drbd: Move conf_mutex from connection to resource

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: drbd_adm_prepare(): Only set adm_ctx.connection when a connection is requested
Andreas Gruenbacher [Wed, 6 Jul 2011 13:05:58 +0000 (15:05 +0200)] 
drbd: drbd_adm_prepare(): Only set adm_ctx.connection when a connection is requested

Also change drbd_adm_connect() to expect a resource after it requested one.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Iterate over all connections
Andreas Gruenbacher [Wed, 6 Jul 2011 13:03:31 +0000 (15:03 +0200)] 
drbd: Iterate over all connections

in drbd_adm_down(), drbd_create_device() and drbd_set_role()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename net_conf variables old_conf -> old_net_conf and new_conf -> new_net_conf
Andreas Gruenbacher [Wed, 6 Jul 2011 12:42:39 +0000 (14:42 +0200)] 
drbd: Rename net_conf variables old_conf -> old_net_conf and new_conf -> new_net_conf

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: drbd_create_device(): Take a resource instead of a connection argument
Andreas Gruenbacher [Wed, 6 Jul 2011 12:20:49 +0000 (14:20 +0200)] 
drbd: drbd_create_device(): Take a resource instead of a connection argument

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: drbd_adm_new_resource(): Check if resource exists, not if it has any connections
Andreas Gruenbacher [Tue, 5 Jul 2011 16:26:00 +0000 (18:26 +0200)] 
drbd: drbd_adm_new_resource(): Check if resource exists, not if it has any connections

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: get_one_status(): Iterate over resource->devices instead of connection->peer_de...
Andreas Gruenbacher [Tue, 5 Jul 2011 16:23:07 +0000 (18:23 +0200)] 
drbd: get_one_status(): Iterate over resource->devices instead of connection->peer_devices

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename drbd_{create,delete}_minor -> drbd_{create,delete}_device
Andreas Gruenbacher [Sun, 3 Jul 2011 21:32:26 +0000 (23:32 +0200)] 
drbd: Rename drbd_{create,delete}_minor -> drbd_{create,delete}_device

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Add explicit device parameter to D_ASSERT
Andreas Gruenbacher [Mon, 27 Jun 2011 14:23:33 +0000 (16:23 +0200)] 
drbd: Add explicit device parameter to D_ASSERT

The implicit dependency on a variable inside the macro is problematic.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Replace and remove the obsolete conn_() macros
Andreas Gruenbacher [Wed, 6 Jul 2011 09:01:44 +0000 (11:01 +0200)] 
drbd: Replace and remove the obsolete conn_() macros

With the polymorphic drbd_() macros, we no longer need the connection
specific variants.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Turn drbd_printk() into a polymorphic macro
Andreas Gruenbacher [Wed, 6 Jul 2011 08:57:39 +0000 (10:57 +0200)] 
drbd: Turn drbd_printk() into a polymorphic macro

This allows drbd_alert(), drbd_err(), drbd_warn(), and drbd_info() to work for
a resource, device, or connection so that we don't have to introduce three
separate sets of macros for that.

The drbd_printk() macro itself is pretty ugly, but that problem is limited to
one place in the code.  Using drbd_printk() on an object type which it doesn't
understand results in an undefined drbd_printk_with_wrong_object_type symbol.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Remove the terrible DEV hack
Andreas Gruenbacher [Sun, 3 Jul 2011 15:53:52 +0000 (17:53 +0200)] 
drbd: Remove the terrible DEV hack

DRBD was using dev_err() and similar all over the code; instead of having to
write dev_err(disk_to_dev(device->vdisk), ...) to convert a drbd_device into a
kernel device, a DEV macro was used which implicitly references the device
variable.  This is terrible; introduce separate drbd_err() and similar macros
with an explicit device parameter instead.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Turn connection->volumes into connection->peer_devices
Andreas Gruenbacher [Tue, 21 Jun 2011 15:23:59 +0000 (17:23 +0200)] 
drbd: Turn connection->volumes into connection->peer_devices

Let connection->peer_devices point to peer devices; connection->volumes was
pointing to devices.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Move resource options from connection to resource
Andreas Gruenbacher [Tue, 21 Jun 2011 14:11:28 +0000 (16:11 +0200)] 
drbd: Move resource options from connection to resource

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: conn_try_disconnect(): Use parameter instead of the global variable
Andreas Gruenbacher [Tue, 21 Jun 2011 11:58:18 +0000 (13:58 +0200)] 
drbd: conn_try_disconnect(): Use parameter instead of the global variable

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Replace conn_get_by_name() with drbd_find_resource()
Andreas Gruenbacher [Mon, 13 Jun 2011 12:27:45 +0000 (14:27 +0200)] 
drbd: Replace conn_get_by_name() with drbd_find_resource()

So far, connections and resources always come in pairs, but in the future with
multiple connections per resource, the names will stick with the resources.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Add struct drbd_resource->devices
Andreas Gruenbacher [Wed, 8 Jun 2011 23:40:48 +0000 (01:40 +0200)] 
drbd: Add struct drbd_resource->devices

This allows to access the volumes of a resource by number.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Minor cleanup in conn_new_minor()
Andreas Gruenbacher [Thu, 23 May 2013 12:57:17 +0000 (14:57 +0200)] 
drbd: Minor cleanup in conn_new_minor()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Add struct drbd_device->resource
Andreas Gruenbacher [Wed, 8 Jun 2011 23:38:00 +0000 (01:38 +0200)] 
drbd: Add struct drbd_device->resource

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: drbd_adm_down(): Move valid resource name check to drbd_adm_prepare()
Andreas Gruenbacher [Wed, 8 Jun 2011 23:26:16 +0000 (01:26 +0200)] 
drbd: drbd_adm_down(): Move valid resource name check to drbd_adm_prepare()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Add struct drbd_resource
Andreas Gruenbacher [Wed, 8 Jun 2011 20:17:38 +0000 (22:17 +0200)] 
drbd: Add struct drbd_resource

In a first step, each resource has exactly one connection, and both objects are
allocated at the same time.  The final result will be one resource and zero or
more connections.

Only allow to delete a resource if all its connections are C_STANDALONE.
Stop the worker threads of all connections early enough.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Improve some function and variable naming
Andreas Gruenbacher [Tue, 7 Jun 2011 20:54:17 +0000 (22:54 +0200)] 
drbd: Improve some function and variable naming

Rename functions
conn_destroy() -> drbd_destroy_connection(),
drbd_minor_destroy() -> drbd_destroy_device()
drbd_adm_add_minor() -> drbd_adm_add_minor()
drbd_adm_delete_minor() -> drbd_adm_del_minor()

Rename global variable minors to drbd_devices

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Introduce "peer_device" object between "device" and "connection"
Andreas Gruenbacher [Tue, 31 May 2011 12:33:49 +0000 (14:33 +0200)] 
drbd: Introduce "peer_device" object between "device" and "connection"

In a setup where a device (aka volume) can replicate to multiple peers and one
connection can be shared between multiple devices, we need separate objects to
represent devices on peer nodes and network connections.

As a first step to introduce multiple connections per device, give each
drbd_device object a single drbd_peer_device object which connects it to a
drbd_connection object.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename drbd_tconn -> drbd_connection
Andreas Gruenbacher [Mon, 30 May 2011 14:32:41 +0000 (16:32 +0200)] 
drbd: Rename drbd_tconn -> drbd_connection

sed -i -e 's:all_tconn:connections:g' -e 's:tconn:connection:g'

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename "mdev" to "device"
Andreas Gruenbacher [Sun, 3 Jul 2011 11:26:43 +0000 (13:26 +0200)] 
drbd: Rename "mdev" to "device"

sed -i -e 's:mdev:device:g'

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Rename struct drbd_conf -> struct drbd_device
Andreas Gruenbacher [Mon, 30 May 2011 14:15:21 +0000 (16:15 +0200)] 
drbd: Rename struct drbd_conf -> struct drbd_device

sed -i -e 's:\<drbd_conf\>:drbd_device:g'

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Split off on-the-wire protocol definitions
Andreas Gruenbacher [Mon, 30 May 2011 09:47:37 +0000 (11:47 +0200)] 
drbd: Split off on-the-wire protocol definitions

Keep the protocol definitions separate from the kernel code; they are useful in
their own right.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Describe the future high-level structure of DRBD
Andreas Gruenbacher [Fri, 10 Jun 2011 15:02:53 +0000 (17:02 +0200)] 
drbd: Describe the future high-level structure of DRBD

The following commits will successively transform DRBD into the structure
described here.

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agoidr: Add new function idr_is_empty()
Andreas Gruenbacher [Mon, 8 Aug 2011 21:36:56 +0000 (23:36 +0200)] 
idr: Add new function idr_is_empty()

Signed-off-by: Andreas Gruenbacher <agruen@linbit.com>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrbd: Add missing error goto
Philipp Reisner [Thu, 1 Aug 2013 08:21:47 +0000 (10:21 +0200)] 
drbd: Add missing error goto

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>
10 years agodrivers: block: Remove unused function drbd_bm_write_lazy() in drbd_bitmap.c
Rashika Kheria [Thu, 19 Dec 2013 09:33:36 +0000 (15:03 +0530)] 
drivers: block: Remove unused function drbd_bm_write_lazy() in drbd_bitmap.c

Remove unused function drbd_bm_write_lazy() in drbd/drbd_bitmap.c.

This eliminates the following warning in drbd/drbd_bitmap.c:
drivers/block/drbd/drbd_bitmap.c:1208:5: warning: no previous prototype for â€˜drbd_bm_write_lazy’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark function seq_printf_with_thousands_grouping() as static in drbd_...
Rashika Kheria [Thu, 19 Dec 2013 09:34:47 +0000 (15:04 +0530)] 
drivers: block: Mark function seq_printf_with_thousands_grouping() as static in drbd_proc.c

Mark function seq_printf_with_thousands_grouping() as static in
drbd/drbd_proc.c because it is not used outside this file.

This eliminates the following warning in drbd/drbd_proc.c:
drivers/block/drbd/drbd_proc.c:49:6: warning: no previous prototype for â€˜seq_printf_with_thousands_grouping’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark the function as static in drbd_worker.c
Rashika Kheria [Thu, 19 Dec 2013 09:36:10 +0000 (15:06 +0530)] 
drivers: block: Mark the function as static in drbd_worker.c

Mark functions drbd_endio_read_sec_final(), drbd_send_barrier(),
need_to_send_barrier(), dequeue_work_batch(), dequeue_work_item() and
wait_for_work() as static in drbd/drbd_worker.c because they are not
used outside this file.

This eliminates the following warnings in drbd/drbd_worker.c:
drivers/block/drbd/drbd_worker.c:99:6: warning: no previous prototype for â€˜drbd_endio_read_sec_final’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_worker.c:1276:5: warning: no previous prototype for â€˜drbd_send_barrier’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_worker.c:1774:6: warning: no previous prototype for â€˜need_to_send_barrier’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_worker.c:1798:6: warning: no previous prototype for â€˜dequeue_work_batch’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_worker.c:1806:6: warning: no previous prototype for â€˜dequeue_work_item’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_worker.c:1815:6: warning: no previous prototype for â€˜wait_for_work’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Move prototype declaration to appropriate header file from drbd_main.c
Rashika Kheria [Thu, 19 Dec 2013 09:37:47 +0000 (15:07 +0530)] 
drivers: block: Move prototype declaration to appropriate header file from drbd_main.c

Move prototype declaration of functions drbdd_init() and drbd_asender()
from drbd/drbd_main.c to header file drbd/drbd_int.h because these
functions are used by more than one file.

This eliminates the following warning in drbd/drbd_receiver.c:
drivers/block/drbd/drbd_receiver.c:4836:5: warning: no previous prototype for â€˜drbdd_init’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_receiver.c:5245:5: warning: no previous prototype for â€˜drbd_asender’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark functions as static in drbd_receiver.c
Rashika Kheria [Thu, 19 Dec 2013 09:41:09 +0000 (15:11 +0530)] 
drivers: block: Mark functions as static in drbd_receiver.c

Mark functions conn_wait_active_ee_empty() and
drbd_crypto_alloc_digest_safe() as static in drbd/drbd_receiver.c
because they are not used outside this file.

This eliminates the following warning in drbd/drbd_receiver.c:
drivers/block/drbd/drbd_receiver.c:1401:6: warning: no previous prototype for â€˜conn_wait_active_ee_empty’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_receiver.c:3259:21: warning: no previous prototype for â€˜drbd_crypto_alloc_digest_safe’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark functions as static in drbd_req.c
Rashika Kheria [Thu, 19 Dec 2013 09:42:27 +0000 (15:12 +0530)] 
drivers: block: Mark functions as static in drbd_req.c

Mark functions drbd_request_prepare() and find_oldest_request() as
static in drbd/drbd_req.c because they are not used outside this file.

This eliminates the following warnings in drbd/drbd_req.c:
drivers/block/drbd/drbd_req.c:1037:1: warning: no previous prototype for â€˜drbd_request_prepare’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_req.c:1323:22: warning: no previous prototype for â€˜find_oldest_request’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Move prototype declaration of function tl_abort_disk_io() to appropri...
Rashika Kheria [Thu, 19 Dec 2013 09:45:02 +0000 (15:15 +0530)] 
drivers: block: Move prototype declaration of function tl_abort_disk_io() to appropriate header file from drbd_state.c

Move the prototype declaration of function tl_abort_disk_io() from
drbd/drbd_state.c to appropriate header file drbd/drbd_int.h because it
is used by more than 2 files.

This eliminates the following warnings in drbd/drbd_main.c:
drivers/block/drbd/drbd_main.c:310:6: warning: no previous prototype for â€˜tl_abort_disk_io’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark function as static in drbd_actlog.c
Rashika Kheria [Thu, 19 Dec 2013 09:43:37 +0000 (15:13 +0530)] 
drivers: block: Mark function as static in drbd_actlog.c

Mark the function drbd_al_begin_io_prepare() as static in
drbd/drbd_actlog.c because it is not used outside this file.

This eliminates the following warnings in drbd/drbd_actlog.c:
drivers/block/drbd/drbd_actlog.c:277:6: warning: no previous prototype for â€˜drbd_al_begin_io_prepare’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark functions as static in drbd_nl.c
Rashika Kheria [Thu, 19 Dec 2013 09:47:33 +0000 (15:17 +0530)] 
drivers: block: Mark functions as static in drbd_nl.c

Mark functions conn_khelper(), nla_put_drbd_cfg_context(),
nla_put_status_info() and get_one_status() as static in drbd/drbd_nl.c
because they are not used outside this file.

This eliminates the following warnings in drbd/drbd_nl.c:
drivers/block/drbd/drbd_nl.c:365:5: warning: no previous prototype for â€˜conn_khelper’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_nl.c:2727:5: warning: no previous prototype for â€˜nla_put_drbd_cfg_context’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_nl.c:2753:5: warning: no previous prototype for â€˜nla_put_status_info’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_nl.c:2895:5: warning: no previous prototype for â€˜get_one_status’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agodrivers: block: Mark functions as static in drbd_main.c
Rashika Kheria [Thu, 19 Dec 2013 09:46:22 +0000 (15:16 +0530)] 
drivers: block: Mark functions as static in drbd_main.c

Mark functions _drbd_send_uuids(), fill_bitmap_rle_bits() and
init_submitter() as static in drbd/drbd_main.c because they are
not used outside this file.

This eliminates the following warnings in drbd/drbd_main.c:
drivers/block/drbd/drbd_main.c:826:5: warning: no previous prototype for â€˜_drbd_send_uuids’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_main.c:1070:5: warning: no previous prototype for â€˜fill_bitmap_rle_bits’ [-Wmissing-prototypes]
drivers/block/drbd/drbd_main.c:2592:5: warning: no previous prototype for â€˜init_submitter’ [-Wmissing-prototypes]

Signed-off-by: Rashika Kheria <rashika.kheria@gmail.com>
Reviewed-by: Josh Triplett <josh@joshtriplett.org>
Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
10 years agoLinux 3.14-rc3
Linus Torvalds [Sun, 16 Feb 2014 21:30:25 +0000 (13:30 -0800)] 
Linux 3.14-rc3

10 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux...
Linus Torvalds [Sun, 16 Feb 2014 19:05:27 +0000 (11:05 -0800)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/linux-btrfs

Pull btrfs fixes from Chris Mason:
 "We have a small collection of fixes in my for-linus branch.

  The big thing that stands out is a revert of a new ioctl.  Users
  haven't shipped yet in btrfs-progs, and Dave Sterba found a better way
  to export the information"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: use right clone root offset for compressed extents
  btrfs: fix null pointer deference at btrfs_sysfs_add_one+0x105
  Btrfs: unset DCACHE_DISCONNECTED when mounting default subvol
  Btrfs: fix max_inline mount option
  Btrfs: fix a lockdep warning when cleaning up aborted transaction
  Revert "btrfs: add ioctl to export size of global metadata reservation"

10 years agoMerge tag 'dt-fixes-for-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/robh...
Linus Torvalds [Sun, 16 Feb 2014 19:03:58 +0000 (11:03 -0800)] 
Merge tag 'dt-fixes-for-3.14' of git://git./linux/kernel/git/robh/linux

Pull devicetree fixes from Rob Herring:
 "Fix booting on PPC boards.  Changes to of_match_node matching caused
  the serial port on some PPC boards to stop working.  Reverted the
  change and reimplement to split matching between new style compatible
  only matching and fallback to old matching algorithm"

* tag 'dt-fixes-for-3.14' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  of: search the best compatible match first in __of_match_node()
  Revert "OF: base: match each node compatible against all given matches first"

10 years agoof: search the best compatible match first in __of_match_node()
Kevin Hao [Fri, 14 Feb 2014 05:22:46 +0000 (13:22 +0800)] 
of: search the best compatible match first in __of_match_node()

Currently, of_match_node compares each given match against all node's
compatible strings with of_device_is_compatible.

To achieve multiple compatible strings per node with ordering from
specific to generic, this requires given matches to be ordered from
specific to generic. For most of the drivers this is not true and also
an alphabetical ordering is more sane there.

Therefore, this patch introduces a function to match each of the node's
compatible strings against all given compatible matches without type and
name first, before checking the next compatible string. This implies
that node's compatibles are ordered from specific to generic while
given matches can be in any order. If we fail to find such a match
entry, then fall-back to the old method in order to keep compatibility.

Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Kevin Hao <haokexin@gmail.com>
Tested-by: Stephen Chivers <schivers@csc.com>
Signed-off-by: Rob Herring <robh@kernel.org>
10 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending
Linus Torvalds [Sun, 16 Feb 2014 00:18:47 +0000 (16:18 -0800)] 
Merge git://git./linux/kernel/git/nab/target-pending

Pull SCSI target fixes from Nicholas Bellinger:
 "Mostly minor fixes this time to v3.14-rc1 related changes.  Also
  included is one fix for a free after use regression in persistent
  reservations UNREGISTER logic that is CC'ed to >= v3.11.y stable"

* git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  Target/sbc: Fix protection copy routine
  IB/srpt: replace strict_strtoul() with kstrtoul()
  target: Simplify command completion by removing CMD_T_FAILED flag
  iser-target: Fix leak on failure in isert_conn_create_fastreg_pool
  iscsi-target: Fix SNACK Type 1 + BegRun=0 handling
  target: Fix missing length check in spc_emulate_evpd_83()
  qla2xxx: Remove last vestiges of qla_tgt_cmd.cmd_list
  target: Fix 32-bit + CONFIG_LBDAF=n link error w/ sector_div
  target: Fix free-after-use regression in PR unregister

10 years agoMerge branch 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Sun, 16 Feb 2014 00:17:51 +0000 (16:17 -0800)] 
Merge branch 'i2c/for-current' of git://git./linux/kernel/git/wsa/linux

Pull i2c fixes from Wolfram Sang:
 "i2c has a bugfix and documentation improvements for you"

* 'i2c/for-current' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  Documentation: i2c: mention ACPI method for instantiating devices
  Documentation: i2c: describe devicetree method for instantiating devices
  i2c: mv64xxx: refactor message start to ensure proper initialization

10 years agoMerge branches 'irq-urgent-for-linus' and 'irq-core-for-linus' of git://git.kernel...
Linus Torvalds [Sun, 16 Feb 2014 00:06:12 +0000 (16:06 -0800)] 
Merge branches 'irq-urgent-for-linus' and 'irq-core-for-linus' of git://git./linux/kernel/git/tip/tip

Pull irq update from Thomas Gleixner:
 "Fix from the urgent branch: a trivial oneliner adding the missing
  Kconfig dependency curing build failures which have been discovered by
  several build robots.

  The update in the irq-core branch provides a new function in the
  irq/devres code, which is a prerequisite for driver developers to get
  rid of boilerplate code all over the place.

  Not a bugfix, but it has zero impact on the current kernel due to the
  lack of users.  It's simpler to provide the infrastructure to
  interested parties via your tree than fulfilling the wishlist of
  driver maintainers on which particular commit or tag this should be
  based on"

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq: Add missing irq_to_desc export for CONFIG_SPARSE_IRQ=n

* 'irq-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq: Add devm_request_any_context_irq()

10 years agoMerge branch 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 16 Feb 2014 00:04:42 +0000 (16:04 -0800)] 
Merge branch 'timers-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull timer fixes from Thomas Gleixner:
 "The following trilogy of patches brings you:

   - fix for a long standing math overflow issue with HZ < 60

   - an onliner fix for a corner case in the dreaded tick broadcast
     mechanism affecting a certain range of AMD machines which are
     infested with the infamous automagic C1E power control misfeature

   - a fix for one of the ARM platforms which allows the kernel to
     proceed and boot instead of stupidly panicing for no good reason.
     The patch is slightly larger than necessary, but it's less ugly
     than the alternative 5 liner"

* 'timers-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  tick: Clear broadcast pending bit when switching to oneshot
  clocksource: Kona: Print warning rather than panic
  time: Fix overflow when HZ is smaller than 60

10 years agoMerge tag 'trace-fixes-v3.14-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 15 Feb 2014 23:03:34 +0000 (15:03 -0800)] 
Merge tag 'trace-fixes-v3.14-rc2' of git://git./linux/kernel/git/rostedt/linux-trace

Pull twi tracing fixes from Steven Rostedt:
 "Two urgent fixes in the tracing utility.

  The first is a fix for the way the ring buffer stores timestamps.
  After a restructure of the code was done, the ring buffer timestamp
  logic missed the fact that the first event on a sub buffer is to have
  a zero delta, as the full timestamp is stored on the sub buffer
  itself.  But because the delta was not cleared to zero, the timestamp
  for that event will be calculated as the real timestamp + the delta
  from the last timestamp.  This can skew the timestamps of the events
  and have them say they happened when they didn't really happen.
  That's bad.

  The second fix is for modifying the function graph caller site.  When
  the stop machine was removed from updating the function tracing code,
  it missed updating the function graph call site location.  It is still
  modified as if it is being done via stop machine.  But it's not.  This
  can lead to a GPF and kernel crash if the function graph call site
  happens to lie between cache lines and one CPU is executing it while
  another CPU is doing the update.  It would be a very hard condition to
  hit, but the result is severe enough to have it fixed ASAP"

* tag 'trace-fixes-v3.14-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  ftrace/x86: Use breakpoints for converting function graph caller
  ring-buffer: Fix first commit on sub-buffer having non-zero delta

10 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 15 Feb 2014 23:02:28 +0000 (15:02 -0800)] 
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 EFI fixes from Peter Anvin:
 "A few more EFI-related fixes"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/efi: Check status field to validate BGRT header
  x86/efi: Fix 32-bit fallout

10 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Sat, 15 Feb 2014 23:01:33 +0000 (15:01 -0800)] 
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM SoC fixes from Kevin Hilman:
 "A collection of ARM SoC fixes for v3.14-rc1.

  Mostly a collection of Kconfig, device tree data and compilation fixes
  along with fix to drivers/phy that fixes a boot regression on some
  Marvell mvebu platforms"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  dma: mv_xor: Silence a bunch of LPAE-related warnings
  ARM: ux500: disable msp2 device tree node
  ARM: zynq: Reserve not DMAable space in front of the kernel
  ARM: multi_v7_defconfig: Select CONFIG_SOC_DRA7XX
  ARM: imx6: Initialize low-power mode early again
  ARM: pxa: fix various compilation problems
  ARM: pxa: fix compilation problem on AM300EPD board
  ARM: at91: add Atmel's SAMA5D3 Xplained board
  spi/atmel: document clock properties
  mmc: atmel-mci: document clock properties
  ARM: at91: enable USB host on at91sam9n12ek board
  ARM: at91/dt: fix sama5d3 ohci hclk clock reference
  ARM: at91/dt: sam9263: fix compatibility string for the I2C
  ata: sata_mv: Fix probe failures with optional phys
  drivers: phy: Add support for optional phys
  drivers: phy: Make NULL a valid phy reference
  ARM: fix HAVE_ARM_TWD selection for OMAP and shmobile
  ARM: moxart: move DMA_OF selection to driver
  ARM: hisi: fix kconfig warning on HAVE_ARM_TWD

10 years agoDocumentation: i2c: mention ACPI method for instantiating devices
Wolfram Sang [Sat, 15 Feb 2014 14:58:35 +0000 (15:58 +0100)] 
Documentation: i2c: mention ACPI method for instantiating devices

Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Acked-by: Guenter Roeck <linux@roeck-us.net>
10 years agoDocumentation: i2c: describe devicetree method for instantiating devices
Wolfram Sang [Mon, 10 Feb 2014 10:03:55 +0000 (11:03 +0100)] 
Documentation: i2c: describe devicetree method for instantiating devices

Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
10 years agoBtrfs: use right clone root offset for compressed extents
Filipe David Borba Manana [Sat, 15 Feb 2014 15:53:16 +0000 (15:53 +0000)] 
Btrfs: use right clone root offset for compressed extents

For non compressed extents, iterate_extent_inodes() gives us offsets
that take into account the data offset from the file extent items, while
for compressed extents it doesn't. Therefore we have to adjust them before
placing them in a send clone instruction. Not doing this adjustment leads to
the receiving end requesting for a wrong a file range to the clone ioctl,
which results in different file content from the one in the original send
root.

Issue reproducible with the following excerpt from the test I made for
xfstests:

  _scratch_mkfs
  _scratch_mount "-o compress-force=lzo"

  $XFS_IO_PROG -f -c "truncate 118811" $SCRATCH_MNT/foo
  $XFS_IO_PROG -c "pwrite -S 0x0d -b 39987 92267 39987" $SCRATCH_MNT/foo

  $BTRFS_UTIL_PROG subvolume snapshot -r $SCRATCH_MNT $SCRATCH_MNT/mysnap1

  $XFS_IO_PROG -c "pwrite -S 0x3e -b 80000 200000 80000" $SCRATCH_MNT/foo
  $BTRFS_UTIL_PROG filesystem sync $SCRATCH_MNT
  $XFS_IO_PROG -c "pwrite -S 0xdc -b 10000 250000 10000" $SCRATCH_MNT/foo
  $XFS_IO_PROG -c "pwrite -S 0xff -b 10000 300000 10000" $SCRATCH_MNT/foo

  # will be used for incremental send to be able to issue clone operations
  $BTRFS_UTIL_PROG subvolume snapshot -r $SCRATCH_MNT $SCRATCH_MNT/clones_snap

  $BTRFS_UTIL_PROG subvolume snapshot -r $SCRATCH_MNT $SCRATCH_MNT/mysnap2

  $FSSUM_PROG -A -f -w $tmp/1.fssum $SCRATCH_MNT/mysnap1
  $FSSUM_PROG -A -f -w $tmp/2.fssum -x $SCRATCH_MNT/mysnap2/mysnap1 \
      -x $SCRATCH_MNT/mysnap2/clones_snap $SCRATCH_MNT/mysnap2
  $FSSUM_PROG -A -f -w $tmp/clones.fssum $SCRATCH_MNT/clones_snap \
      -x $SCRATCH_MNT/clones_snap/mysnap1 -x $SCRATCH_MNT/clones_snap/mysnap2

  $BTRFS_UTIL_PROG send $SCRATCH_MNT/mysnap1 -f $tmp/1.snap
  $BTRFS_UTIL_PROG send $SCRATCH_MNT/clones_snap -f $tmp/clones.snap
  $BTRFS_UTIL_PROG send -p $SCRATCH_MNT/mysnap1 \
      -c $SCRATCH_MNT/clones_snap $SCRATCH_MNT/mysnap2 -f $tmp/2.snap

  _scratch_unmount
  _scratch_mkfs
  _scratch_mount

  $BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $tmp/1.snap
  $FSSUM_PROG -r $tmp/1.fssum $SCRATCH_MNT/mysnap1 2>> $seqres.full

  $BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $tmp/clones.snap
  $FSSUM_PROG -r $tmp/clones.fssum $SCRATCH_MNT/clones_snap 2>> $seqres.full

  $BTRFS_UTIL_PROG receive $SCRATCH_MNT -f $tmp/2.snap
  $FSSUM_PROG -r $tmp/2.fssum $SCRATCH_MNT/mysnap2 2>> $seqres.full

Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agobtrfs: fix null pointer deference at btrfs_sysfs_add_one+0x105
Anand Jain [Wed, 15 Jan 2014 09:22:28 +0000 (17:22 +0800)] 
btrfs: fix null pointer deference at btrfs_sysfs_add_one+0x105

bdev is null when disk has disappeared and mounted with
the degrade option

stack trace
---------
btrfs_sysfs_add_one+0x105/0x1c0 [btrfs]
open_ctree+0x15f3/0x1fe0 [btrfs]
btrfs_mount+0x5db/0x790 [btrfs]
? alloc_pages_current+0xa4/0x160
mount_fs+0x34/0x1b0
vfs_kern_mount+0x62/0xf0
do_mount+0x22e/0xa80
? __get_free_pages+0x9/0x40
? copy_mount_options+0x31/0x170
SyS_mount+0x7e/0xc0
system_call_fastpath+0x16/0x1b
---------

reproducer:
-------
mkfs.btrfs -draid1 -mraid1 /dev/sdc /dev/sdd
(detach a disk)
devmgt detach /dev/sdc [1]
mount -o degrade /dev/sdd /btrfs
-------

[1] github.com/anajain/devmgt.git

Signed-off-by: Anand Jain <Anand.Jain@oracle.com>
Tested-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agoi2c: mv64xxx: refactor message start to ensure proper initialization
Wolfram Sang [Thu, 13 Feb 2014 20:36:29 +0000 (21:36 +0100)] 
i2c: mv64xxx: refactor message start to ensure proper initialization

Because the offload mechanism can fall back to a standard transfer,
having two seperate initialization states is unfortunate. Let's just
have one state which does things consistently. This fixes a bug where
some preparation was missing when the fallback happened. And it makes
the code much easier to follow. To implement this, we put the check
if offload is possible at the top of the offload setup function.

Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Tested-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Cc: stable@vger.kernel.org # v3.12+
Fixes: 930ab3d403ae (i2c: mv64xxx: Add I2C Transaction Generator support)
10 years agoMerge tag 'usb-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sat, 15 Feb 2014 00:15:45 +0000 (16:15 -0800)] 
Merge tag 'usb-3.14-rc3' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here is a bunch of USB fixes for 3.14-rc3.  Most of these are xhci
  reverts, fixing a bunch of reported issues with USB 3 host controller
  issues that loads of people have been hitting (with the exception of
  kernel developers, all of our machines seem to be working fine, which
  is why these took so long to get resolved...)

  There are some other minor fixes and new device ids, as ususal.  All
  have been in linux-next successfully"

* tag 'usb-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (22 commits)
  usb: option: blacklist ZTE MF667 net interface
  Revert "usb: xhci: Link TRB must not occur within a USB payload burst"
  Revert "xhci: Avoid infinite loop when sg urb requires too many trbs"
  Revert "xhci: Set scatter-gather limit to avoid failed block writes."
  xhci 1.0: Limit arbitrarily-aligned scatter gather.
  Modpost: fixed USB alias generation for ranges including 0x9 and 0xA
  usb: core: Fix potential memory leak adding dyn USBdevice IDs
  USB: ftdi_sio: add Tagsys RFID Reader IDs
  usb: qcserial: add Netgear Aircard 340U
  usb-storage: enable multi-LUN scanning when needed
  USB: simple: add Dynastream ANT USB-m Stick device support
  usb-storage: add unusual-devs entry for BlackBerry 9000
  usb-storage: restrict bcdDevice range for Super Top in Cypress ATACB
  usb: phy: move some error messages to debug
  usb: ftdi_sio: add Mindstorms EV3 console adapter
  usb: dwc2: fix memory corruption in dwc2 driver
  usb: dwc2: fix role switch breakage
  usb: dwc2: bail out early when booting with "nousb"
  Revert "xhci: replace xhci_read_64() with readq()"
  Revert "xhci: replace xhci_write_64() with writeq()"
  ...

10 years agoMerge tag 'tty-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sat, 15 Feb 2014 00:15:03 +0000 (16:15 -0800)] 
Merge tag 'tty-3.14-rc3' of git://git./linux/kernel/git/gregkh/tty

Pull tty/serial driver fixes from Greg KH:
 "Here are a small number of tty/serial driver fixes to resolve reported
  issues with 3.14-rc and earlier (in the case of the vt bugfix).  Some
  of these have been tested and reported by a number of people as the
  tty bugfix was pretty commonly hit on some platforms.

  All have been in linux-next for a while"

* tag 'tty-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  vt: Fix secure clear screen
  serial: 8250: Support XR17V35x fraction divisor
  n_tty: Fix stale echo output
  serial: sirf: fix kernel panic caused by unpaired spinlock
  serial: 8250_pci: unbreak last serial ports on NetMos 9865 cards
  n_tty: Fix poll() when TIME_CHAR and MIN_CHAR == 0
  serial: omap: fix rs485 probe on defered pinctrl
  serial: 8250_dw: fix compilation warning when !CONFIG_PM_SLEEP
  serial: omap-serial: Move info message to probe function
  tty: Set correct tty name in 'active' sysfs attribute
  tty: n_gsm: Fix for modems with brk in modem status control
  drivers/tty/hvc: don't use module_init in non-modular hyp. console code

10 years agoMerge tag 'staging-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 15 Feb 2014 00:14:11 +0000 (16:14 -0800)] 
Merge tag 'staging-3.14-rc3' of git://git./linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are a number (lots, I know) of fixes for staging drivers to
  resolve a bunch of reported issues.

  The largest patches here is one revert of a patch that is in 3.14-rc1
  to fix reported problems, and a sync of a usb host driver that
  required some ARM patches to go in before it could be accepted (which
  is why it missed -rc1)

  All of these have been in linux-next for a while with no reported
  issues"

* tag 'staging-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging: (56 commits)
  staging/rtl8821ae: fix build, depends on MAC80211
  iio: max1363: Use devm_regulator_get_optional for optional regulator
  iio:accel:bma180: Use modifier instead of index in channel specification
  iio: adis16400: Set timestamp as the last element in chan_spec
  iio: ak8975: Fix calculation formula for convert micro tesla to gauss unit
  staging:iio:ad799x fix typo in ad799x_events[]
  iio: mxs-lradc: remove useless scale_available files
  iio: mxs-lradc: fix buffer overflow
  iio:magnetometer:mag3110: Fix output of decimal digits in show_int_plus_micros()
  iio:magnetometer:mag3110: Report busy in _read_raw() / write_raw() when buffer is enabled
  wlags49_h2: Fix overflow in wireless_set_essid()
  xlr_net: Fix missing trivial allocation check
  staging: r8188eu: overflow in rtw_p2p_get_go_device_address()
  staging: r8188eu: array overflow in rtw_mp_ioctl_hdl()
  staging: r8188eu: Fix typo in USB_DEVICE list
  usbip/userspace/libsrc/names.c: memory leak
  gpu: ion: dereferencing an ERR_PTR
  staging: comedi: usbduxsigma: fix unaligned dereferences
  staging: comedi: fix too early cleanup in comedi_auto_config()
  staging: android: ion: dummy: fix an error code
  ...

10 years agoMerge tag 'driver-core-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 15 Feb 2014 00:13:40 +0000 (16:13 -0800)] 
Merge tag 'driver-core-3.14-rc3' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fix from Greg KH:
 "Here is a single driver core patch for 3.14-rc3 for the component code
  that Russell has found and fixed"

* tag 'driver-core-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  drivers/base: fix devres handling for master device

10 years agoMerge tag 'char-misc-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregk...
Linus Torvalds [Sat, 15 Feb 2014 00:13:00 +0000 (16:13 -0800)] 
Merge tag 'char-misc-3.14-rc3' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc fixes from Greg KH:
 "Here are some small char/misc driver fixes, along with some
  documentation updates, for 3.14-rc3.  Nothing major, just a number of
  fixes for reported issues"

* tag 'char-misc-3.14-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  Revert "misc: eeprom: sunxi: Add new compatibles"
  Revert "ARM: sunxi: dt: Convert to the new SID compatibles"
  misc: mic: fix possible signed underflow (undefined behavior) in userspace API
  ARM: sunxi: dt: Convert to the new SID compatibles
  misc: eeprom: sunxi: Add new compatibles
  misc: genwqe: Fix potential memory leak when pinning memory
  Documentation:Update Documentation/zh_CN/arm64/memory.txt
  Documentation:Update Documentation/zh_CN/arm64/booting.txt
  Documentation:Chinese translation of Documentation/arm64/tagged-pointers.txt
  raw: set range for MAX_RAW_DEVS
  raw: test against runtime value of max_raw_minors
  Drivers: hv: vmbus: Don't timeout during the initial connection with host
  Drivers: hv: vmbus: Specify the target CPU that should receive notification
  VME: Correct read/write alignment algorithm
  mei: don't unset read cb ptr on reset
  mei: clear write cb from waiting list on reset

10 years agoBtrfs: unset DCACHE_DISCONNECTED when mounting default subvol
Josef Bacik [Fri, 14 Feb 2014 18:43:48 +0000 (13:43 -0500)] 
Btrfs: unset DCACHE_DISCONNECTED when mounting default subvol

A user was running into errors from an NFS export of a subvolume that had a
default subvol set.  When we mount a default subvol we will use d_obtain_alias()
to find an existing dentry for the subvolume in the case that the root subvol
has already been mounted, or a dummy one is allocated in the case that the root
subvol has not already been mounted.  This allows us to connect the dentry later
on if we wander into the path.  However if we don't ever wander into the path we
will keep DCACHE_DISCONNECTED set for a long time, which angers NFS.  It doesn't
appear to cause any problems but it is annoying nonetheless, so simply unset
DCACHE_DISCONNECTED in the get_default_root case and switch btrfs_lookup() to
use d_materialise_unique() instead which will make everything play nicely
together and reconnect stuff if we wander into the defaul subvol path from a
different way.  With this patch I'm no longer getting the NFS errors when
exporting a volume that has been mounted with a default subvol set.  Thanks,

cc: bfields@fieldses.org
cc: ebiederm@xmission.com
Signed-off-by: Josef Bacik <jbacik@fb.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agoBtrfs: fix max_inline mount option
Mitch Harder [Thu, 13 Feb 2014 15:13:16 +0000 (09:13 -0600)] 
Btrfs: fix max_inline mount option

Currently, the only mount option for max_inline that has any effect is
max_inline=0.  Any other value that is supplied to max_inline will be
adjusted to a minimum of 4k.  Since max_inline has an effective maximum
of ~3900 bytes due to page size limitations, the current behaviour
only has meaning for max_inline=0.

This patch will allow the the max_inline mount option to accept non-zero
values as indicated in the documentation.

Signed-off-by: Mitch Harder <mitch.harder@sabayonlinux.org>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agoBtrfs: fix a lockdep warning when cleaning up aborted transaction
Liu Bo [Sat, 8 Feb 2014 07:33:08 +0000 (15:33 +0800)] 
Btrfs: fix a lockdep warning when cleaning up aborted transaction

Given now we have 2 spinlock for management of delayed refs,
CONFIG_DEBUG_SPINLOCK=y helped me find this,

[ 4723.413809] BUG: spinlock wrong CPU on CPU#1, btrfs-transacti/2258
[ 4723.414882]  lock: 0xffff880048377670, .magic: dead4ead, .owner: btrfs-transacti/2258, .owner_cpu: 2
[ 4723.417146] CPU: 1 PID: 2258 Comm: btrfs-transacti Tainted: G        W  O 3.12.0+ #4
[ 4723.421321] Call Trace:
[ 4723.421872]  [<ffffffff81680fe7>] dump_stack+0x54/0x74
[ 4723.422753]  [<ffffffff81681093>] spin_dump+0x8c/0x91
[ 4723.424979]  [<ffffffff816810b9>] spin_bug+0x21/0x26
[ 4723.425846]  [<ffffffff81323956>] do_raw_spin_unlock+0x66/0x90
[ 4723.434424]  [<ffffffff81689bf7>] _raw_spin_unlock+0x27/0x40
[ 4723.438747]  [<ffffffffa015da9e>] btrfs_cleanup_one_transaction+0x35e/0x710 [btrfs]
[ 4723.443321]  [<ffffffffa015df54>] btrfs_cleanup_transaction+0x104/0x570 [btrfs]
[ 4723.444692]  [<ffffffff810c1b5d>] ? trace_hardirqs_on_caller+0xfd/0x1c0
[ 4723.450336]  [<ffffffff810c1c2d>] ? trace_hardirqs_on+0xd/0x10
[ 4723.451332]  [<ffffffffa015e5ee>] transaction_kthread+0x22e/0x270 [btrfs]
[ 4723.452543]  [<ffffffffa015e3c0>] ? btrfs_cleanup_transaction+0x570/0x570 [btrfs]
[ 4723.457833]  [<ffffffff81079efa>] kthread+0xea/0xf0
[ 4723.458990]  [<ffffffff81079e10>] ? kthread_create_on_node+0x140/0x140
[ 4723.460133]  [<ffffffff81692aac>] ret_from_fork+0x7c/0xb0
[ 4723.460865]  [<ffffffff81079e10>] ? kthread_create_on_node+0x140/0x140
[ 4723.496521] ------------[ cut here ]------------

----------------------------------------------------------------------

The reason is that we get to call cond_resched_lock(&head_ref->lock) while
still holding @delayed_refs->lock.

So it's different with __btrfs_run_delayed_refs(), where we do drop-acquire
dance before and after actually processing delayed refs.

Here we don't drop the lock, others are not able to add new delayed refs to
head_ref, so cond_resched_lock(&head_ref->lock) is not necessary here.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Chris Mason <clm@fb.com>
10 years agoRevert "btrfs: add ioctl to export size of global metadata reservation"
Chris Mason [Fri, 14 Feb 2014 21:42:13 +0000 (13:42 -0800)] 
Revert "btrfs: add ioctl to export size of global metadata reservation"

This reverts commit 01e219e8069516cdb98594d417b8bb8d906ed30d.

David Sterba found a different way to provide these features without adding a new
ioctl.  We haven't released any progs with this ioctl yet, so I'm taking this out
for now until we finalize things.

Signed-off-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.cz>
CC: Jeff Mahoney <jeffm@suse.com>
This page took 0.080842 seconds and 5 git commands to generate.