Age | Commit message (Collapse) | Author |
|
under lock
|
|
correct current task under lock.
|
|
|
|
Test patch that throttles buffered writeback to make it a lot
more smooth, and has way less impact on other system activity.
Background writeback should be, by definition, background
activity. The fact that we flush huge bundles of it at the time
means that it potentially has heavy impacts on foreground workloads,
which isn't ideal. We can't easily limit the sizes of writes that
we do, since that would impact file system layout in the presence
of delayed allocation. So just throttle back buffered writeback,
unless someone is waiting for it.
The algorithm for when to throttle takes its inspiration in the
CoDel networking scheduling algorithm. Like CoDel, blk-wb monitors
the minimum latencies of requests over a window of time. In that
window of time, if the minimum latency of any request exceeds a
given target, then a scale count is incremented and the queue depth
is shrunk. The next monitoring window is shrunk accordingly. Unlike
CoDel, if we hit a window that exhibits good behavior, then we
simply increment the scale count and re-calculate the limits for that
scale value. This prevents us from oscillating between a
close-to-ideal value and max all the time, instead remaining in the
windows where we get good behavior.
Unlike CoDel, blk-wb allows the scale count to to negative. This
happens if we primarily have writes going on. Unlike positive
scale counts, this doesn't change the size of the monitoring window.
When the heavy writers finish, blk-bw quickly snaps back to it's
stable state of a zero scale count.
The patch registers two sysfs entries. The first one, 'wb_window_usec',
defines the window of monitoring. The second one, 'wb_lat_usec',
sets the latency target for the window. It defaults to 2 msec for
non-rotational storage, and 75 msec for rotational storage. Setting
this value to '0' disables blk-wb. Generally, a user would not have
to touch these settings.
We don't enable WBT on devices that are managed with CFQ, and have
a non-root block cgroup attached. If we have a proportional share setup
on this particular disk, then the wbt throttling will interfere with
that. We don't have a strong need for wbt for that case, since we will
rely on CFQ doing that for us.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
task, removing the need for racy setting of task_cpu of it in set_task_cpu. Check against wake_cpu when calling set_task_cpu.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
improvements and logic fixes, including the ability to change iso tasks back to
normal.
|
|
|
|
|
|
functions to match core.c
|
|
altered by another runqueue.
|
|
current cpu.
|
|
|
|
|
|
|
|
|
|
synchronise niffies to the highest value as idle ticks may have artificially
kept niffies low on one CPU and the truth can only be later.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
We can hook this up to the block layer, to help throttle buffered
writes. Or NFS can tap into it, to accomplish the same.
wbt registers a few trace points that can be used to track what is
happening in the system:
wbt_lat: 259:0: latency 2446318
wbt_stat: 259:0: rmean=2446318, rmin=2446318, rmax=2446318, rsamples=1,
wmean=518866, wmin=15522, wmax=5330353, wsamples=57
wbt_step: 259:0: step down: step=1, window=72727272, background=8, normal=16, max=32
This shows a sync issue event (wbt_lat) that exceeded it's time. wbt_stat
dumps the current read/write stats for that window, and wbt_step shows a
step down event where we now scale back writes. Each trace includes the
device, 259:0 in this case.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
For legacy block, we simply track them in the request queue. For
blk-mq, we track them on a per-sw queue basis, which we can then
sum up through the hardware queues and finally to a per device
state.
The stats are tracked in, roughly, 0.1s interval windows.
Add sysfs files to display the stats.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
For blk-mq, ->nr_requests does track queue depth, at least at init
time. But for the older queue paths, it's simply a soft setting.
On top of that, it's generally larger than the hardware setting
on purpose, to allow backup of requests for merging.
Fill a hole in struct request with a 'queue_depth' member, that
drivers can call to more closely inform the block layer of the
real queue depth.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Note in the bdi_writeback structure whenever a task ends up sleeping
waiting for progress. We can use that information in the lower layers
to increase the priority of writes.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
If we're doing background type writes, then use the appropriate
write command for that.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
Add wbc_to_write_flags(), which returns the write modifier flags to use,
based on a struct writeback_control. No functional changes in this
patch, but it prepares us for factoring other wbc fields for write type.
Signed-off-by: Jens Axboe <axboe@fb.com>
Reviewed-by: Jan Kara <jack@suse.cz>
|
|
This adds a new request flag, REQ_BG, that callers can use to tell
the block layer that this is background (non-urgent) IO.
Signed-off-by: Jens Axboe <axboe@fb.com>
|
|
|
|
Pull ARM fixes from Russell King:
"Three relatively small fixes for ARM:
- Roger noticed that dma_max_pfn() was calculating the upper limit
wrongly, by adding the PFN offset of memory twice.
- A fix from Robin to correct parsing of MPIDR values when the
address size is larger than one BE32 unit.
- A fix from Srinivas to ensure that we do not rely on the boot
loader (or previous Linux kernel) setting the translation table
base register a certain way in the decompressor, which can lead to
crashes"
* 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm:
ARM: 8618/1: decompressor: reset ttbcr fields to use TTBR0 on ARMv7
ARM: 8617/1: dma: fix dma_max_pfn()
ARM: 8616/1: dt: Respect property size when parsing CPUs
|
|
If the bootloader uses the long descriptor format and jumps to
kernel decompressor code, TTBCR may not be in a right state.
Before enabling the MMU, it is required to clear the TTBCR.PD0
field to use TTBR0 for translation table walks.
The commit dbece45894d3a ("ARM: 7501/1: decompressor:
reset ttbcr for VMSA ARMv7 cores") does the reset of TTBCR.N, but
doesn't consider all the bits for the size of TTBCR.N.
Clear TTBCR.PD0 field and reset all the three bits of TTBCR.N to
indicate the use of TTBR0 and the correct base address width.
Fixes: dbece45894d3 ("ARM: 7501/1: decompressor: reset ttbcr for VMSA ARMv7 cores")
Acked-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Srinivas Ramana <sramana@codeaurora.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Thomas Gleixner:
"The last regression fixes for 4.8 final:
- Two patches addressing the fallout of the CR4 optimizations which
caused CR4-less machines to fail.
- Fix the VDSO build on big endian machines
- Take care of FPU initialization if no CPUID is available otherwise
task struct size ends up being zero
- Fix up context tracking in case load_gs_index fails"
* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/entry/64: Fix context tracking state warning when load_gs_index fails
x86/boot: Initialize FPU and X86_FEATURE_ALWAYS even if we don't have CPUID
x86/vdso: Fix building on big endian host
x86/boot: Fix another __read_cr4() case on 486
x86/init: Fix cr4_init_shadow() on CR4-less machines
|
|
Pull MIPS fixes from Ralf Baechle:
"Another round of fixes:
- CM: Fix mips_cm_max_vp_width for non-MT kernels on MT systems
- CPS: Avoid BUG() when offlining pre-r6 CPUs
- DEC: Avoid gas warnings due to suspicious instruction scheduling by
manually expanding assembler macros.
- FTLB: Fix configuration by moving confiuguratoin after probing
- FTLB: clear execution hazard after changing FTLB enable
- Highmem: Fix detection of unsupported highmem with cache aliases
- I6400: Don't touch FTLBP chicken bits
- microMIPS: Fix BUILD_ROLLBACK_PROLOGUE
- Malta: Fix IOCU disable switch read for MIPS64
- Octeon: Fix probing of devices attached to GPIO lines
- uprobes: Misc small fixes"
* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus:
MIPS: CM: Fix mips_cm_max_vp_width for non-MT kernels on MT systems
MIPS: Fix detection of unsupported highmem with cache aliases
MIPS: Malta: Fix IOCU disable switch read for MIPS64
MIPS: Fix BUILD_ROLLBACK_PROLOGUE for microMIPS
MIPS: clear execution hazard after changing FTLB enable
MIPS: Configure FTLB after probing TLB sizes from config4
MIPS: Stop setting I6400 FTLBP
MIPS: DEC: Avoid la pseudo-instruction in delay slots
MIPS: Octeon: mark GPIO controller node not populated after IRQ init.
MIPS: uprobes: fix use of uninitialised variable
MIPS: uprobes: remove incorrect set_orig_insn
MIPS: fix uretprobe implementation
MIPS: smp-cps: Avoid BUG() when offlining pre-r6 CPUs
|
|
Pull sparc fixes from David Miller:
1) Fix section mismatches in some builds, from Paul Gortmaker.
2) Need to count huge zero page mappings when doing TSB sizing, from
Mike Kravetz.
3) Fix handing of cpu_possible_mask when nr_cpus module option is
specified, from Atish Patra.
4) Don't allocate irq stacks until nr_irqs has been processed, also
from Atish Patra.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc:
sparc64: Fix non-SMP build.
sparc64: Fix irq stack bootmem allocation.
sparc64: Fix cpu_possible_mask if nr_cpus is set
sparc64 mm: Fix more TSB sizing issues
sparc64: fix section mismatch in find_numa_latencies_for_group
|
|
Pull networking fixes from David Miller:
1) Fix wrong TCP checksums on MTU probing when checksum offloading is
disabled, from Douglas Caetano dos Santos.
2) Fix qdisc backlog updates in qfq and sfb schedulers, from Cong Wang.
3) Route lookup flow key protocol value is wrong in ip6gre_xmit_other(),
fix from Lance Richardson.
4) Scheduling while atomic in multicast routing code of ipv4 and ipv6,
fix from Nikolay Aleksandrov.
5) Fix packet alignment in fec driver, from Eric Nelson.
6) Fix perf regression in sctp due to struct layout and cache misses,
from Xin Long.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
sctp: fix the issue sctp_diag uses lock_sock in rcu_read_lock
sctp: change to check peer prsctp_capable when using prsctp polices
sctp: remove prsctp_param from sctp_chunk
sctp: move sent_count to the memory hole in sctp_chunk
tg3: Avoid NULL pointer dereference in tg3_io_error_detected()
act_ife: Fix false encoding
act_ife: Fix external mac header on encode
VSOCK: Don't dec ack backlog twice for rejected connections
Revert "net: ethernet: bcmgenet: use phydev from struct net_device"
net: fec: align IP header in hardware
net: fec: remove QUIRK_HAS_RACC from i.mx27
net: fec: remove QUIRK_HAS_RACC from i.mx25
ipmr, ip6mr: fix scheduling while atomic and a deadlock with ipmr_get_route
ip6_gre: fix flowi6_proto value in ip6gre_xmit_other()
tcp: fix a compile error in DBGUNDO()
tcp: fix wrong checksum calculation on MTU probing
sch_sfb: keep backlog updated with qlen
sch_qfq: keep backlog updated with qlen
can: dev: fix deadlock reported after bus-off
|
|
When discovering the number of VPEs per core, smp_num_siblings will be
incorrect for kernels built without support for the MIPS MultiThreading
(MT) ASE running on systems which implement said ASE. This leads to
accesses to VPEs in secondary cores being performed incorrectly since
mips_cm_vp_id calculates the wrong ID to write to the local "other"
registers. Fix this by examining the number of VPEs in the core as
reported by the CM.
This patch presumes that the number of VPEs will be the same in each
core of the system. As this path only applies to systems with CM version
2.5 or lower, and this property is true of all such known systems, this
is likely to be fine but is described in a comment for good measure.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/14338/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fix from James Bottomley:
"One final fix before 4.8.
There was a memory leak triggered by turning scsi mq off due to the
fact that we assume on host release that the already running hosts
weren't mq based because that's the state of the global flag (even
though they were).
Fix it by tracking this on a per host host basis"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: Avoid that toggling use_blk_mq triggers a memory leak
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Pull input fix from Dmitry Torokhov:
"One small change to make joydev (which is used by older games) to bind
to devices that export Z axis but not X or Y (such as TRC rudder)"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: joydev - recognize devices with Z axis as joysticks
|
|
Merge more fixes from Andrew Morton:
"Three fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
include/linux/property.h: fix typo/compile error
ocfs2: fix deadlock on mmapped page in ocfs2_write_begin_nolock()
mm: workingset: fix crash in shadow node shrinker caused by replace_page_cache_page()
|
|
This fixes commit d76eebfa175e ("include/linux/property.h: fix build
issues with gcc-4.4.4").
With that commit we get the following compile error when using the
PROPERTY_ENTRY_INTEGER_ARRAY macro.
include/linux/property.h:201:39: error: `u32_data' undeclared (first
use in this function)
PROPERTY_ENTRY_INTEGER_ARRAY(_name_, u32, _val_)
^
include/linux/property.h:193:17: note: in definition of macro
`PROPERTY_ENTRY_INTEGER_ARRAY'
{ .pointer = { _type_##_data = _val_ } }, \
^
This needs a '.' to reference the union member. It seems this was just
overlooked here since it is done correctly in similar constructs in
other parts of the original commit.
This fix is in preparation of upcoming commits that will use this macro.
Fixes: commit d76eebfa175e ("include/linux/property.h: fix build issues with gcc-4.4.4")
Link: http://lkml.kernel.org/r/2de3b929290d88a723ed829a3e3cbd02044714df.1475114627.git.johnyoun@synopsys.com
Signed-off-by: John Youn <johnyoun@synopsys.com>
Cc: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The testcase "mmaptruncate" of ocfs2-test deadlocks occasionally.
In this testcase, we create a 2*CLUSTER_SIZE file and mmap() on it;
there are 2 process repeatedly performing the following operations
respectively: one is doing memset(mmaped_addr + 2*CLUSTER_SIZE - 1, 'a',
1), while the another is playing ftruncate(fd, 2*CLUSTER_SIZE) and then
ftruncate(fd, CLUSTER_SIZE) again and again.
This is the backtrace when the deadlock happens:
__wait_on_bit_lock+0x50/0xa0
__lock_page+0xb7/0xc0
ocfs2_write_begin_nolock+0x163f/0x1790 [ocfs2]
ocfs2_page_mkwrite+0x1c7/0x2a0 [ocfs2]
do_page_mkwrite+0x66/0xc0
handle_mm_fault+0x685/0x1350
__do_page_fault+0x1d8/0x4d0
trace_do_page_fault+0x37/0xf0
do_async_page_fault+0x19/0x70
async_page_fault+0x28/0x30
In ocfs2_write_begin_nolock(), we first grab the pages and then allocate
disk space for this write; ocfs2_try_to_free_truncate_log() will be
called if -ENOSPC is returned; if we're lucky to get enough clusters,
which is usually the case, we start over again.
But in ocfs2_free_write_ctxt() the target page isn't unlocked, so we
will deadlock when trying to grab the target page again.
Also, -ENOMEM might be returned in ocfs2_grab_pages_for_write().
Another deadlock will happen in __do_page_mkwrite() if
ocfs2_page_mkwrite() returns non-VM_FAULT_LOCKED, and along with a
locked target page.
These two errors fail on the same path, so fix them by unlocking the
target page manually before ocfs2_free_write_ctxt().
Jan Kara helps me clear out the JBD2 part, and suggest the hint for root
cause.
Changes since v1:
1. Also put ENOMEM error case into consideration.
Link: http://lkml.kernel.org/r/1474173902-32075-1-git-send-email-zren@suse.com
Signed-off-by: Eric Ren <zren@suse.com>
Reviewed-by: He Gang <ghe@suse.com>
Acked-by: Joseph Qi <joseph.qi@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
replace_page_cache_page()
Antonio reports the following crash when using fuse under memory pressure:
kernel BUG at /build/linux-a2WvEb/linux-4.4.0/mm/workingset.c:346!
invalid opcode: 0000 [#1] SMP
Modules linked in: all of them
CPU: 2 PID: 63 Comm: kswapd0 Not tainted 4.4.0-36-generic #55-Ubuntu
Hardware name: System manufacturer System Product Name/P8H67-M PRO, BIOS 3904 04/27/2013
task: ffff88040cae6040 ti: ffff880407488000 task.ti: ffff880407488000
RIP: shadow_lru_isolate+0x181/0x190
Call Trace:
__list_lru_walk_one.isra.3+0x8f/0x130
list_lru_walk_one+0x23/0x30
scan_shadow_nodes+0x34/0x50
shrink_slab.part.40+0x1ed/0x3d0
shrink_zone+0x2ca/0x2e0
kswapd+0x51e/0x990
kthread+0xd8/0xf0
ret_from_fork+0x3f/0x70
which corresponds to the following sanity check in the shadow node
tracking:
BUG_ON(node->count & RADIX_TREE_COUNT_MASK);
The workingset code tracks radix tree nodes that exclusively contain
shadow entries of evicted pages in them, and this (somewhat obscure)
line checks whether there are real pages left that would interfere with
reclaim of the radix tree node under memory pressure.
While discussing ways how fuse might sneak pages into the radix tree
past the workingset code, Miklos pointed to replace_page_cache_page(),
and indeed there is a problem there: it properly accounts for the old
page being removed - __delete_from_page_cache() does that - but then
does a raw raw radix_tree_insert(), not accounting for the replacement
page. Eventually the page count bits in node->count underflow while
leaving the node incorrectly linked to the shadow node LRU.
To address this, make sure replace_page_cache_page() uses the tracked
page insertion code, page_cache_tree_insert(). This fixes the page
accounting and makes sure page-containing nodes are properly unlinked
from the shadow node LRU again.
Also, make the sanity checks a bit less obscure by using the helpers for
checking the number of pages and shadows in a radix tree node.
Fixes: 449dd6984d0e ("mm: keep page cache radix tree nodes in check")
Link: http://lkml.kernel.org/r/20160919155822.29498-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Antonio SJ Musumeci <trapexit@spawn.link>
Debugged-by: Miklos Szeredi <miklos@szeredi.hu>
Cc: <stable@vger.kernel.org> [3.15+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Change my email address to my kernel.org account instead of the ARM one.
Signed-off-by: Javi Merino <javi.merino@arm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|