Pull block driver updates from Jens Axboe:
"Here are the block driver changes for 4.4. This pull request
contains:
- NVMe:
- Refactor and moving of code to prepare for proper target
support. From Christoph and Jay.
- 32-bit nvme warning fix from Arnd.
- Error initialization fix from me.
- Proper namespace removal and reference counting support from
Keith.
- Device resume fix on IO failure, also from Keith.
- Dependency fix from Keith, now that nvme isn't under the
umbrella of the block anymore.
- Target location and maintainers update from Jay.
- From Ming Lei, the long awaited DIO/AIO support for loop.
- Enable BD-RE writeable opens, from Georgios"
* 'for-4.4/drivers' of git://git.kernel.dk/linux-block: (24 commits)
Update target repo for nvme patch contributions
NVMe: initialize error to '0'
nvme: use an integer value to Linux errno values
nvme: fix 32-bit build warning
NVMe: Add explicit block config dependency
nvme: include <linux/types.ĥ> in <linux/nvme.h>
nvme: move to a new drivers/nvme/host directory
nvme.h: add missing nvme_id_ctrl endianess annotations
nvme: move hardware structures out of the uapi version of nvme.h
nvme: add a local nvme.h header
nvme: properly handle partially initialized queues in nvme_create_io_queues
nvme: merge nvme_dev_start, nvme_dev_resume and nvme_async_probe
nvme: factor reset code into a common helper
nvme: merge nvme_dev_reset into nvme_reset_failed_dev
nvme: delete dev from dev_list in nvme_reset
NVMe: Simplify device resume on io queue failure
NVMe: Namespace removal simplifications
NVMe: Reference count open namespaces
cdrom: Random writing support for BD-RE media
block: loop: support DIO & AIO
...
- Improve balloon driver memory hotplug placement.
- Use unpopulated hotplugged memory for foreign pages (if
supported/enabled).
- Support 64 KiB guest pages on arm64.
- CPU hotplug support on arm/arm64.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJWOeSkAAoJEFxbo/MsZsTRph0H/0nE8Tx0GyGtOyCYfBdInTvI
WgjvL8VR1XrweZMVis3668MzhLSYg6b5lvJsoi+L3jlzYRyze43iHXsKfvp+8p0o
TVUhFnlHEHF8ASEtPydAi6HgS7Dn9OQ9LaZ45R1Gk0rHnwJjIQonhTn2jB0yS9Am
Hf4aZXP2NVZphjYcloqNsLH0G6mGLtgq8cS0uKcVO2YIrR4Dr3sfj9qfq9mflf8n
sA/5ifoHRfOUD1vJzYs4YmIBUv270jSsprWK/Mi2oXIxUTBpKRAV1RVCAPW6GFci
HIZjIJkjEPWLsvxWEs0dUFJQGp3jel5h8vFPkDWBYs3+9rILU2DnLWpKGNDHx3k=
=vUfa
-----END PGP SIGNATURE-----
Merge tag 'for-linus-4.4-rc0-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen updates from David Vrabel:
- Improve balloon driver memory hotplug placement.
- Use unpopulated hotplugged memory for foreign pages (if
supported/enabled).
- Support 64 KiB guest pages on arm64.
- CPU hotplug support on arm/arm64.
* tag 'for-linus-4.4-rc0-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip: (44 commits)
xen: fix the check of e_pfn in xen_find_pfn_range
x86/xen: add reschedule point when mapping foreign GFNs
xen/arm: don't try to re-register vcpu_info on cpu_hotplug.
xen, cpu_hotplug: call device_offline instead of cpu_down
xen/arm: Enable cpu_hotplug.c
xenbus: Support multiple grants ring with 64KB
xen/grant-table: Add an helper to iterate over a specific number of grants
xen/xenbus: Rename *RING_PAGE* to *RING_GRANT*
xen/arm: correct comment in enlighten.c
xen/gntdev: use types from linux/types.h in userspace headers
xen/gntalloc: use types from linux/types.h in userspace headers
xen/balloon: Use the correct sizeof when declaring frame_list
xen/swiotlb: Add support for 64KB page granularity
xen/swiotlb: Pass addresses rather than frame numbers to xen_arch_need_swiotlb
arm/xen: Add support for 64KB page granularity
xen/privcmd: Add support for Linux 64KB page granularity
net/xen-netback: Make it running on 64KB page granularity
net/xen-netfront: Make it running on 64KB page granularity
block/xen-blkback: Make it running on 64KB page granularity
block/xen-blkfront: Make it running on 64KB page granularity
...
rbd requires stable pages, as it performs a crc of the page data before
they are send to the OSDs.
But since kernel 3.9 (patch 1d1d1a7672
"mm: only enforce stable page writes if the backing device requires
it") it is not assumed anymore that block devices require stable pages.
This patch sets the necessary flag to get stable pages back for rbd.
In a ceph installation that provides multiple ext4 formatted rbd
devices "bad crc" messages appeared regularly (ca 1 message every 1-2
minutes on every OSD that provided the data for the rbd) in the
OSD-logs before this patch. After this patch this messages are pretty
much gone (only ca 1-2 / month / OSD).
Cc: stable@vger.kernel.org # 3.9+, needs backporting
Signed-off-by: Ronny Hegewald <Ronny.Hegewald@online.de>
[idryomov@gmail.com: require stable pages only in crc case, changelog]
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Pull block layer fixes from Jens Axboe:
"A final set of fixes for 4.3.
It is (again) bigger than I would have liked, but it's all been
through the testing mill and has been carefully reviewed by multiple
parties. Each fix is either a regression fix for this cycle, or is
marked stable. You can scold me at KS. The pull request contains:
- Three simple fixes for NVMe, fixing regressions since 4.3. From
Arnd, Christoph, and Keith.
- A single xen-blkfront fix from Cathy, fixing a NULL dereference if
an error is returned through the staste change callback.
- Fixup for some bad/sloppy code in nbd that got introduced earlier
in this cycle. From Markus Pargmann.
- A blk-mq tagset use-after-free fix from Junichi.
- A backing device lifetime fix from Tejun, fixing a crash.
- And finally, a set of regression/stable fixes for cgroup writeback
from Tejun"
* 'for-linus' of git://git.kernel.dk/linux-block:
writeback: remove broken rbtree_postorder_for_each_entry_safe() usage in cgwb_bdi_destroy()
NVMe: Fix memory leak on retried commands
block: don't release bdi while request_queue has live references
nvme: use an integer value to Linux errno values
blk-mq: fix use-after-free in blk_mq_free_tag_set()
nvme: fix 32-bit build warning
writeback: fix incorrect calculation of available memory for memcg domains
writeback: memcg dirty_throttle_control should be initialized with wb->memcg_completions
writeback: bdi_writeback iteration must not skip dying ones
writeback: fix bdi_writeback iteration in wakeup_dirtytime_writeback()
writeback: laptop_mode_timer_fn() needs rcu_read_lock() around bdi_writeback iteration
nbd: Add locking for tasks
xen-blkfront: check for null drvdata in blkback_changed (XenbusStateClosing)
Mapping an image with a long parent chain (e.g. image foo, whose parent
is bar, whose parent is baz, etc) currently leads to a kernel stack
overflow, due to the following recursion in the reply path:
rbd_osd_req_callback()
rbd_obj_request_complete()
rbd_img_obj_callback()
rbd_img_parent_read_callback()
rbd_obj_request_complete()
...
Limit the parent chain to 16 images, which is ~5K worth of stack. When
the above recursion is eliminated, this limit can be lifted.
Fixes: http://tracker.ceph.com/issues/12538
Cc: stable@vger.kernel.org # 3.10+, needs backporting for < 4.2
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Josh Durgin <jdurgin@redhat.com>
Currently we leak parent_spec and trigger a "parent reference
underflow" warning if rbd_dev_create() in rbd_dev_probe_parent() fails.
The problem is we take the !parent out_err branch and that only drops
refcounts; parent_spec that would've been freed had we called
rbd_dev_unparent() remains and triggers rbd_warn() in
rbd_dev_parent_put() - at that point we have parent_spec != NULL and
parent_ref == 0, so counter ends up being -1 after the decrement.
Redo rbd_dev_probe_parent() to fix this.
Cc: stable@vger.kernel.org # 3.10+, needs backporting for < 4.2
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Linux may use a different page size than the size of grant. So make
clear that the order is actually in number of grant.
Signed-off-by: Julien Grall <julien.grall@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
The PV block protocol is using 4KB page granularity. The goal of this
patch is to allow a Linux using 64KB page granularity behaving as a
block backend on a non-modified Xen.
It's only necessary to adapt the ring size and the number of request per
indirect frames. The rest of the code is relying on the grant table
code.
Note that the grant table code is allocating a Linux page per grant
which will result to waste 6OKB for every grant when Linux is using 64KB
page granularity. This could be improved by sharing the page between
multiple grants.
Signed-off-by: Julien Grall <julien.grall@citrix.com>
Acked-by: "Roger Pau Monné" <roger.pau@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
The PV block protocol is using 4KB page granularity. The goal of this
patch is to allow a Linux using 64KB page granularity using block
device on a non-modified Xen.
The block API is using segment which should at least be the size of a
Linux page. Therefore, the driver will have to break the page in chunk
of 4K before giving the page to the backend.
When breaking a 64KB segment in 4KB chunks, it is possible that some
chunks are empty. As the PV protocol always require to have data in the
chunk, we have to count the number of Xen page which will be in use and
avoid sending empty chunks.
Note that, a pre-defined number of grants are reserved before preparing
the request. This pre-defined number is based on the number and the
maximum size of the segments. If each segment contains a very small
amount of data, the driver may reserve too many grants (16 grants is
reserved per segment with 64KB page granularity).
Furthermore, in the case of persistent grants we allocate one Linux page
per grant although only the first 4KB of the page will be effectively
in use. This could be improved by sharing the page with multiple grants.
Signed-off-by: Julien Grall <julien.grall@citrix.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Prepare the code to support 64KB page granularity. The first
implementation will use a full Linux page per indirect and persistent
grant. When non-persistent grant is used, each page of a bio request
may be split in multiple grant.
Furthermore, the field page of the grant structure is only used to copy
data from persistent grant or indirect grant. Avoid to set it for other
use case as it will have no meaning given the page will be split in
multiple grant.
Provide 2 functions, to setup indirect grant, the other for bio page.
Signed-off-by: Julien Grall <julien.grall@citrix.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
All the usage of the field pfn are done using the same idiom:
pfn_to_page(grant->pfn)
This will return always the same page. Store directly the page in the
grant to clean up the code.
Signed-off-by: Julien Grall <julien.grall@citrix.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
Currently, blkif_queue_request has 2 distinct execution path:
- Send a discard request
- Send a read/write request
The function is also allocating grants to use for generating the
request. Although, this is only used for read/write request.
Rather than having a function with 2 distinct execution path, separate
the function in 2. This will also remove one level of tabulation.
Signed-off-by: Julien Grall <julien.grall@citrix.com>
Reviewed-by: Roger Pau Monné <roger.pau@citrix.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
This covers only the simplest case - an object size sized write, but
it's still useful in tiering setups when EC is used for the base tier
as writefull op can be proxied, saving an object promotion.
Even though updating ceph_osdc_new_request() to allow writefull should
just be a matter of fixing an assert, I didn't do it because its only
user is cephfs. All other sites were updated.
Reflects ceph.git commit 7bfb7f9025a8ee0d2305f49bf0336d2424da5b5b.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Commit 30e2bc08b2 ("Revert "block: remove artifical max_hw_sectors
cap"") restored a clamp on max_sectors. It's now 2560 sectors instead
of 1024, but it's not good enough: we set max_hw_sectors to rbd object
size because we don't want object sized I/Os to be split, and the
default object size is 4M.
So, set max_sectors to max_hw_sectors in rbd at queue init time.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Reviewed-by: Alex Elder <elder@linaro.org>
Resources are reallocated for requeued commands, so unmap and release
the iod for the failed command.
It's a pretty bad memory leak and causes a kernel hang if you remove a
drive because of a busy dma pool. You'll get messages spewing like this:
nvme 0000:xx:xx.x: dma_pool_destroy prp list 256, ffff880420dec000 busy
and lock up pci and the driver since removal never completes while
holding a lock.
Cc: stable@vger.kernel.org
Cc: <stable@vger.kernel.org> # 4.0.x-
Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Use a separate integer variable to hold the signed Linux errno
values we pass back to the block layer. Note that for pass through
commands those might still be NVMe values, but those fit into the
int as well.
Fixes: f4829a9b7a: ("blk-mq: fix racy updates of rq->errors")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Compiling the nvme driver on 32-bit warns about a cast from a __u64
variable to a pointer:
drivers/block/nvme-core.c: In function 'nvme_submit_io':
drivers/block/nvme-core.c:1847:4: warning: cast to pointer from integer of different size [-Wint-to-pointer-cast]
(void __user *)io.addr, length, NULL, 0);
The cast here is intentional and safe, so we can shut up the
gcc warning by adding an intermediate cast to 'uintptr_t'.
I had previously submitted a patch to fix this problem in the
nvme driver, but it was accepted on the same day that two new
warnings got added.
For clarification, I also change the third instance of this cast
to use uintptr_t instead of unsigned long now.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: d29ec8241c ("nvme: submit internal commands through the block layer")
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patch moves the NVMe driver from drivers/block/ to its own new
drivers/nvme/host/ directory. This is in preparation of splitting the
current monolithic driver up and add support for the upcoming NVMe
over Fabrics standard. The drivers/nvme/host/ is chose to leave space
for a NVMe target implementation in addition to this host side driver.
Signed-off-by: Jay Sternberg <jay.e.sternberg@intel.com>
[hch: rebased, renamed core.c to pci.c, slight tweaks]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Currently all NVMe command and completion structures are exposed to userspace
through the uapi version of nvme.h. They are not an ABI between the kernel
and userspace, and will change in C-incompatible way for future versions of
the spec. Move them to the kernel version of the file and rename the uapi
header to nvme_ioctl.h so that userspace can easily detect the presence of
the new clean header. Nvme-cli already carries a local copy of the header,
so it won't be affected by this move.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Add a new drivers/block/nvme.h which contains all the driver internal
interface.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
This avoids having to clean up later in a seemingly unrelated place.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
And give the resulting function a sensible name. This keeps all the
error handling in a single place and will allow for further improvements
to it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
And give the resulting function a more descriptive name.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Device resets need to delete the device from the device list before
kicking of the reset an re-probe, otherwise we get the device added
to the list twice. nvme_reset is the only side missing this deletion
at the moment, and this patch adds it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Releasing IO queues and disks was done in a work queue outside the
controller resume context to delete namespaces if the controller failed
after a resume from suspend. This is unnecessary since we can resume
a device asynchronously.
This patch makes resume use probe_work so it can directly remove
namespaces if the device is manageable but not IO capable. Since the
deleting disks was the only reason we had the convoluted "reset_workfn",
this patch removes that unnecessary indirection.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This liberates namespace removal from the device, allowing gendisk
references to be closed independent of the nvme controller reference
count.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Dynamic namespace attachment means the namespace may be removed at any
time, so the namespace reference count can not be tied to the device
reference count. This fixes a NULL dereference if an opened namespace
is detached from a controller.
Signed-off-by: Keith Busch <keith.busch@intel.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
The timeout handling introduced in
7e2893a16d (nbd: Fix timeout detection)
introduces a race condition which may lead to killing of tasks that are
not in nbd context anymore. This was not observed or reproducable yet.
This patch adds locking to critical use of task_recv and task_send to
avoid killing tasks that already left the NBD thread functions. This
lock is only acquired if a timeout occures or the nbd device
starts/stops.
Reported-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
Reviewed-by: Ben Hutchings <ben@decadent.org.uk>
Fixes: 7e2893a16d ("nbd: Fix timeout detection")
Signed-off-by: Jens Axboe <axboe@fb.com>
Konrad writes:
Please git pull an update branch to your 'for-4.3/drivers' branch (which
oddly I don't see does not have the previous pull?)
git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/for-jens-4.3
which has two fixes - one where we use the Xen blockfront EFI driver and
don't release all the requests, the other if the allocation of resources
for a particular state failed - we would go back 'Closing' and assume
that an structure would be allocated while in fact it may not be - and
crash.
xen-blkfront will crash if the check to talk_to_blkback()
in blkback_changed()(XenbusStateInitWait) returns an error.
The driver data is freed and info is set to NULL. Later during
the close process via talk_to_blkback's call to xenbus_dev_fatal()
the null pointer is passed to and dereference in blkfront_closing.
CC: stable@vger.kernel.org
Signed-off-by: Cathy Avery <cathy.avery@oracle.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
blk_mq_complete_request may be a no-op if the request has already
been completed by others means (e.g. a timeout or cancellation), but
currently drivers have to set rq->errors before calling
blk_mq_complete_request, which might leave us with the wrong error value.
Add an error parameter to blk_mq_complete_request so that we can
defer setting rq->errors until we known we won the race to complete the
request.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
The asynchronous namespace scanning caused affinity hints to be set before
its tagset initialized, so there was no cpu mask to set the hint. This
patch moves the affinity hint setting to after namespaces are scanned.
Reported-by: 김경산 <ks0204.kim@samsung.com>
Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
There are at least 3 advantages to use direct I/O and AIO on
read/write loop's backing file:
1) double cache can be avoided, then memory usage gets
decreased a lot
2) not like user space direct I/O, there isn't cost of
pinning pages
3) avoid context switch for obtaining good throughput
- in buffered file read, random I/O top throughput is often obtained
only if they are submitted concurrently from lots of tasks; but for
sequential I/O, most of times they can be hit from page cache, so
concurrent submissions often introduce unnecessary context switch
and can't improve throughput much. There was such discussion[1]
to use non-blocking I/O to improve the problem for application.
- with direct I/O and AIO, concurrent submissions can be
avoided and random read throughput can't be affected meantime
xfstests(-g auto, ext4) is basically passed when running with
direct I/O(aio), one exception is generic/232, but it failed in
loop buffered I/O(4.2-rc6-next-20150814) too.
Follows the fio test result for performance purpose:
4 jobs fio test inside ext4 file system over loop block
1) How to run
- KVM: 4 VCPUs, 2G RAM
- linux kernel: 4.2-rc6-next-20150814(base) with the patchset
- the loop block is over one image on SSD.
- linux psync, 4 jobs, size 1500M, ext4 over loop block
- test result: IOPS from fio output
2) Throughput(IOPS) becomes a bit better with direct I/O(aio)
-------------------------------------------------------------
test cases |randread |read |randwrite |write |
-------------------------------------------------------------
base |8015 |113811 |67442 |106978
-------------------------------------------------------------
base+loop aio |8136 |125040 |67811 |111376
-------------------------------------------------------------
- somehow, it should be caused by more page cache avaiable for
application or one extra page copy is avoided in case of direct I/O
3) context switch
- context switch decreased by ~50% with loop direct I/O(aio)
compared with loop buffered I/O(4.2-rc6-next-20150814)
4) memory usage from /proc/meminfo
-------------------------------------------------------------
| Buffers | Cached
-------------------------------------------------------------
base | > 760MB | ~950MB
-------------------------------------------------------------
base+loop direct I/O(aio) | < 5MB | ~1.6GB
-------------------------------------------------------------
- so there are much more page caches available for application with
direct I/O
[1] https://lwn.net/Articles/612483/
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
If loop block is mounted via 'mount -o loop', it isn't easy
to pass file descriptor opened as O_DIRECT, so this patch
introduces a new command to support direct IO for this case.
Cc: linux-api@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
This patches provides one interface for enabling direct IO
from user space:
- userspace(such as losetup) can pass 'file' which is
opened/fcntl as O_DIRECT
Also __loop_update_dio() is introduced to check if direct I/O
can be used on current loop setting.
The last big change is to introduce LO_FLAGS_DIRECT_IO flag
for userspace to know if direct IO is used to access backing
file.
Cc: linux-api@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
The following patch will use dio/aio to submit IO to backing file,
then it needn't to schedule IO concurrently from work, so
use kthread_work for decreasing context switch cost a lot.
For non-AIO case, single thread has been used for long long time,
and it was just converted to work in v4.0, which has caused performance
regression for fedora live booting already. In discussion[1], even
though submitting I/O via work concurrently can improve random read IO
throughput, meantime it might hurt sequential read IO performance, so
better to restore to single thread behaviour.
For the following AIO support, it is better to use multi hw-queue
with per-hwq kthread than current work approach suppose there is so
high performance requirement for loop.
[1] http://marc.info/?t=143082678400002&r=1&w=2
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
It doesn't make sense to enable merge because the I/O
submitted to backing file is handled page by page.
Signed-off-by: Ming Lei <ming.lei@canonical.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@fb.com>
Konrad writes:
It has one fix that should go in and also be put in stable tree (I've
added the CC already).
It is a fix for a memory leak that can exposed via using UEFI
xen-blkfront driver.
This is due to commit 86839c56de
"xen/block: add multi-page ring support"
When using an guest under UEFI - after the domain is destroyed
the following warning comes from blkback.
------------[ cut here ]------------
WARNING: CPU: 2 PID: 95 at
/home/julien/works/linux/drivers/block/xen-blkback/xenbus.c:274
xen_blkif_deferred_free+0x1f4/0x1f8()
Modules linked in:
CPU: 2 PID: 95 Comm: kworker/2:1 Tainted: G W 4.2.0 #85
Hardware name: APM X-Gene Mustang board (DT)
Workqueue: events xen_blkif_deferred_free
Call trace:
[<ffff8000000890a8>] dump_backtrace+0x0/0x124
[<ffff8000000891dc>] show_stack+0x10/0x1c
[<ffff8000007653bc>] dump_stack+0x78/0x98
[<ffff800000097e88>] warn_slowpath_common+0x9c/0xd4
[<ffff800000097f80>] warn_slowpath_null+0x14/0x20
[<ffff800000557a0c>] xen_blkif_deferred_free+0x1f0/0x1f8
[<ffff8000000ad020>] process_one_work+0x160/0x3b4
[<ffff8000000ad3b4>] worker_thread+0x140/0x494
[<ffff8000000b2e34>] kthread+0xd8/0xf0
---[ end trace 6f859b7883c88cdd ]---
Request allocation has been moved to connect_ring, which is called every
time blkback connects to the frontend (this can happen multiple times during
a blkback instance life cycle). On the other hand, request freeing has not
been moved, so it's only called when destroying the backend instance. Due to
this mismatch, blkback can allocate the request pool multiple times, without
freeing it.
In order to fix it, move the freeing of requests to xen_blkif_disconnect to
restore the symmetry between request allocation and freeing.
Reported-by: Julien Grall <julien.grall@citrix.com>
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Tested-by: Julien Grall <julien.grall@citrix.com>
Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: David Vrabel <david.vrabel@citrix.com>
Cc: xen-devel@lists.xenproject.org
CC: stable@vger.kernel.org # 4.2
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Pull block updates from Jens Axboe:
"This is a bit bigger than it should be, but I could (did) not want to
send it off last week due to both wanting extra testing, and expecting
a fix for the bounce regression as well. In any case, this contains:
- Fix for the blk-merge.c compilation warning on gcc 5.x from me.
- A set of back/front SG gap merge fixes, from me and from Sagi.
This ensures that we honor SG gapping for integrity payloads as
well.
- Two small fixes for null_blk from Matias, fixing a leak and a
capacity propagation issue.
- A blkcg fix from Tejun, fixing a NULL dereference.
- A fast clone optimization from Ming, fixing a performance
regression since the arbitrarily sized bio's were introduced.
- Also from Ming, a regression fix for bouncing IOs"
* 'for-linus' of git://git.kernel.dk/linux-block:
block: fix bounce_end_io
block: blk-merge: fast-clone bio when splitting rw bios
block: blkg_destroy_all() should clear q->root_blkg and ->root_rl.blkg
block: Copy a user iovec if it includes gaps
block: Refuse adding appending a gapped integrity page to a bio
block: Refuse request/bio merges with gaps in the integrity payload
block: Check for gaps on front and back merges
null_blk: fix wrong capacity when bs is not 512 bytes
null_blk: fix memory leak on cleanup
block: fix bogus compiler warnings in blk-merge.c
zcomp_create() verifies the success of zcomp_strm_{multi,single}_create()
through comp->stream, which can potentially be pointing to memory that
was freed if these functions returned an error.
While at it, replace a 'ERR_PTR(-ENOMEM)' by a more generic
'ERR_PTR(error)' as in the future zcomp_strm_{multi,siggle}_create()
could return other error codes. Function documentation updated
accordingly.
Fixes: beca3ec71f ("zram: add multi stream functionality")
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull Ceph update from Sage Weil:
"There are a few fixes for snapshot behavior with CephFS and support
for the new keepalive protocol from Zheng, a libceph fix that affects
both RBD and CephFS, a few bug fixes and cleanups for RBD from Ilya,
and several small fixes and cleanups from Jianpeng and others"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client:
ceph: improve readahead for file holes
ceph: get inode size for each append write
libceph: check data_len in ->alloc_msg()
libceph: use keepalive2 to verify the mon session is alive
rbd: plug rbd_dev->header.object_prefix memory leak
rbd: fix double free on rbd_dev->header_name
libceph: set 'exists' flag for newly up osd
ceph: cleanup use of ceph_msg_get
ceph: no need to get parent inode in ceph_open
ceph: remove the useless judgement
ceph: remove redundant test of head->safe and silence static analysis warnings
ceph: fix queuing inode to mdsdir's snaprealm
libceph: rename con_work() to ceph_con_workfn()
libceph: Avoid holding the zero page on ceph_msgr_slab_init errors
libceph: remove the unused macro AES_KEY_SIZE
ceph: invalidate dirty pages after forced umount
ceph: EIO all operations after forced umount
- Use the correct GFN/BFN terms more consistently.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJV8VRMAAoJEFxbo/MsZsTRiGQH/i/jrAJUJfrFC2PINaA2gDwe
O0dlrkCiSgAYChGmxxxXZQSPM5Po5+EbT/dLjZ/uvSooeorM9RYY/mFo7ut/qLep
4pyQUuwGtebWGBZTrj9sygUVXVhgJnyoZxskNUbhj9zvP7hb9++IiI78mzne6cpj
lCh/7Z2dgpfRcKlNRu+qpzP79Uc7OqIfDK+IZLrQKlXa7IQDJTQYoRjbKpfCtmMV
BEG3kN9ESx5tLzYiAfxvaxVXl9WQFEoktqe9V8IgOQlVRLgJ2DQWS6vmraGrokWM
3HDOCHtRCXlPhu1Vnrp0R9OgqWbz8FJnmVAndXT8r3Nsjjmd0aLwhJx7YAReO/4=
=JDia
-----END PGP SIGNATURE-----
Merge tag 'for-linus-4.3-rc0b-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen terminology fixes from David Vrabel:
"Use the correct GFN/BFN terms more consistently"
* tag 'for-linus-4.3-rc0b-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/xenbus: Rename the variable xen_store_mfn to xen_store_gfn
xen/privcmd: Further s/MFN/GFN/ clean-up
hvc/xen: Further s/MFN/GFN clean-up
video/xen-fbfront: Further s/MFN/GFN clean-up
xen/tmem: Use xen_page_to_gfn rather than pfn_to_gfn
xen: Use correctly the Xen memory terminologies
arm/xen: implement correctly pfn_to_mfn
xen: Make clear that swiotlb and biomerge are dealing with DMA address
virtio-mmio can now be auto-loaded through acpi.
virtio blk supports extended partitions.
total memory is better reported when using virtio balloon with auto-deflate.
cache control is re-enabled when using virtio-blk in modern mode.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJV7/iUAAoJECgfDbjSjVRpzRIIAMJz5enk2FbyiOsky5THASqK
rWAEx+Jz5Av3FVM99W+Fs8dmXrBGkCTyRgmX3mZx+pGNhOsLjc6xHElTekzwQBas
18oFsQDTWk0rPZkO7X6/OY3GgHyq8VSI/2kW+yKMYj4k/gAaGV+Azf0m5DGMp4wn
rngevvcDSkFEN2Gjxbi3MWT8sJ9Br/EFg6KzKUOiOycHh/rQZsaQK6tq4kpUZkil
2wvXspQY6ix2fD558gsiqUv6l8yZ9mXVgxtyRcnxjfM+8jr+qtqai8yVGCAC5krh
50RXrR22tN/6aLYuvtPj46+Y0dleCMEwW2rSmcw5h5yr3B/0ylnZ1YNp6vlFyv0=
=8xQ2
-----END PGP SIGNATURE-----
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio updates from Michael Tsirkin:
"Virtio fixes and features for 4.3:
- virtio-mmio can now be auto-loaded through acpi.
- virtio blk supports extended partitions.
- total memory is better reported when using virtio balloon with
auto-deflate.
- cache control is re-enabled when using virtio-blk in modern mode"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
virtio_balloon: do not change memory amount visible via /proc/meminfo
virtio_ballon: change stub of release_pages_by_pfn
virtio-blk: Allow extended partitions
virtio_mmio: add ACPI probing
virtio-blk: use VIRTIO_BLK_F_WCE and VIRTIO_BLK_F_CONFIG_WCE in virtio1
Merge second patch-bomb from Andrew Morton:
"Almost all of the rest of MM. There was an unusually large amount of
MM material this time"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (141 commits)
zpool: remove no-op module init/exit
mm: zbud: constify the zbud_ops
mm: zpool: constify the zpool_ops
mm: swap: zswap: maybe_preload & refactoring
zram: unify error reporting
zsmalloc: remove null check from destroy_handle_cache()
zsmalloc: do not take class lock in zs_shrinker_count()
zsmalloc: use class->pages_per_zspage
zsmalloc: consider ZS_ALMOST_FULL as migrate source
zsmalloc: partial page ordering within a fullness_list
zsmalloc: use shrinker to trigger auto-compaction
zsmalloc: account the number of compacted pages
zsmalloc/zram: introduce zs_pool_stats api
zsmalloc: cosmetic compaction code adjustments
zsmalloc: introduce zs_can_compact() function
zsmalloc: always keep per-class stats
zsmalloc: drop unused variable `nr_to_migrate'
mm/memblock.c: fix comment in __next_mem_range()
mm/page_alloc.c: fix type information of memoryless node
memory-hotplug: fix comments in zone_spanned_pages_in_node() and zone_spanned_pages_in_node()
...
Make zram syslog error reporting more consistent. We have random
error levels in some places. For example, critical errors like
"Error allocating memory for compressed page"
and
"Unable to allocate temp memory"
are reported as KERN_INFO messages.
a) Reassign error levels
Error messages that directly affect zram
functionality -- pr_err():
Error allocating zram address table
Error creating memory pool
Decompression failed! err=%d, page=%u
Unable to allocate temp memory
Compression failed! err=%d
Error allocating memory for compressed page: %u, size=%zu
Cannot initialise %s compressing backend
Error allocating disk queue for device %d
Error allocating disk structure for device %d
Error creating sysfs group for device %d
Unable to register zram-control class
Unable to get major number
Messages that do not affect functionality, but user
must be warned (because sysfs attrs will be removed in
this particular case) -- pr_warn():
%d (%s) Attribute %s (and others) will be removed. %s
Messages that do not affect functionality and mostly are
informative -- pr_info():
Cannot change max compression streams
Can't change algorithm for initialized device
Cannot change disksize for initialized device
Added device: %s
Removed device: %s
b) Update sysfs_create_group() error message
First, it lacks a trailing new line; add it. Second, every error message
in zram_add() has a "for device %d" part, which makes errors more
informative. Add missing part to "Error creating sysfs group" message.
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Compaction returns back to zram the number of migrated objects, which is
quite uninformative -- we have objects of different sizes so user space
cannot obtain any valuable data from that number. Change compaction to
operate in terms of pages and return back to compaction issuer the
number of pages that were freed during compaction. So from now on we
will export more meaningful value in zram<id>/mm_stat -- the number of
freed (compacted) pages.
This requires:
(a) a rename of `num_migrated' to 'pages_compacted'
(b) a internal API change -- return first_page's fullness_group from
putback_zspage(), so we know when putback_zspage() did
free_zspage(). It helps us to account compaction stats correctly.
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
`zs_compact_control' accounts the number of migrated objects but it has
a limited lifespan -- we lose it as soon as zs_compaction() returns back
to zram. It worked fine, because (a) zram had it's own counter of
migrated objects and (b) only zram could trigger compaction. However,
this does not work for automatic pool compaction (not issued by zram).
To account objects migrated during auto-compaction (issued by the
shrinker) we need to store this number in zs_pool.
Define a new `struct zs_pool_stats' structure to keep zs_pool's stats
there. It provides only `num_migrated', as of this writing, but it
surely can be extended.
A new zsmalloc zs_pool_stats() symbol exports zs_pool's stats back to
caller.
Use zs_pool_stats() in zram and remove `num_migrated' from zram_stats.
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Suggested-by: Minchan Kim <minchan@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
