Maxim Kochetkov [Mon, 19 Oct 2020 05:06:25 +0000 (08:06 +0300)]
net: dsa: seville: the packet buffer is 2 megabits, not megabytes
The VSC9953 Seville switch has 2 megabits of buffer split into 4360
words of 60 bytes each. 2048 * 1024 is 2 megabytes instead of 2 megabits.
2 megabits is (2048 / 8) * 1024 = 256 * 1024.
Heiner Kallweit [Sun, 18 Oct 2020 16:38:59 +0000 (18:38 +0200)]
r8169: fix operation under forced interrupt threading
For several network drivers it was reported that using
__napi_schedule_irqoff() is unsafe with forced threading. One way to
fix this is switching back to __napi_schedule, but then we lose the
benefit of the irqoff version in general. As stated by Eric it doesn't
make sense to make the minimal hard irq handlers in drivers using NAPI
a thread. Therefore ensure that the hard irq handler is never
thread-ified.
Martin KaFai Lau [Mon, 19 Oct 2020 19:42:12 +0000 (12:42 -0700)]
bpf: Enforce id generation for all may-be-null register type
The commit af7ec1383361 ("bpf: Add bpf_skc_to_tcp6_sock() helper")
introduces RET_PTR_TO_BTF_ID_OR_NULL and
the commit eaa6bcb71ef6 ("bpf: Introduce bpf_per_cpu_ptr()")
introduces RET_PTR_TO_MEM_OR_BTF_ID_OR_NULL.
Note that for RET_PTR_TO_MEM_OR_BTF_ID_OR_NULL, the reg0->type
could become PTR_TO_MEM_OR_NULL which is not covered by
BPF_PROBE_MEM.
The BPF_REG_0 will then hold a _OR_NULL pointer type. This _OR_NULL
pointer type requires the bpf program to explicitly do a NULL check first.
After NULL check, the verifier will mark all registers having
the same reg->id as safe to use. However, the reg->id
is not set for those new _OR_NULL return types. One of the ways
that may be wrong is, checking NULL for one btf_id typed pointer will
end up validating all other btf_id typed pointers because
all of them have id == 0. The later tests will exercise
this path.
To fix it and also avoid similar issue in the future, this patch
moves the id generation logic out of each individual RET type
test in check_helper_call(). Instead, it does one
reg_type_may_be_null() test and then do the id generation
if needed.
This patch also adds a WARN_ON_ONCE in mark_ptr_or_null_reg()
to catch future breakage.
The _OR_NULL pointer usage in the bpf_iter_reg.ctx_arg_info is
fine because it just happens that the existing id generation after
check_ctx_access() has covered it. It is also using the
reg_type_may_be_null() to decide if id generation is needed or not.
Linus Torvalds [Mon, 19 Oct 2020 21:38:46 +0000 (14:38 -0700)]
Merge tag 'xfs-5.10-merge-5' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Pull more xfs updates from Darrick Wong:
"The second large pile of new stuff for 5.10, with changes even more
monumental than last week!
We are formally announcing the deprecation of the V4 filesystem format
in 2030. All users must upgrade to the V5 format, which contains
design improvements that greatly strengthen metadata validation,
supports reflink and online fsck, and is the intended vehicle for
handling timestamps past 2038. We're also deprecating the old Irix
behavioral tweaks in September 2025.
Coming along for the ride are two design changes to the deferred
metadata ops subsystem. One of the improvements is to retain correct
logical ordering of tasks and subtasks, which is a more logical design
for upper layers of XFS and will become necessary when we add atomic
file range swaps and commits. The second improvement to deferred ops
improves the scalability of the log by helping the log tail to move
forward during long-running operations. This reduces log contention
when there are a large number of threads trying to run transactions.
In addition to that, this fixes numerous small bugs in log recovery;
refactors logical intent log item recovery to remove the last
remaining place in XFS where we could have nested transactions; fixes
a couple of ways that intent log item recovery could fail in ways that
wouldn't have happened in the regular commit paths; fixes a deadlock
vector in the GETFSMAP implementation (which improves its performance
by 20%); and fixes serious bugs in the realtime growfs, fallocate, and
bitmap handling code.
Summary:
- Deprecate the V4 filesystem format, some disused mount options, and
some legacy sysctl knobs now that we can support dates into the
25th century. Note that removal of V4 support will not happen until
the early 2030s.
- Fix some probles with inode realtime flag propagation.
- Fix some buffer handling issues when growing a rt filesystem.
- Fix a problem where a BMAP_REMAP unmap call would free rt extents
even though the purpose of BMAP_REMAP is to avoid freeing the
blocks.
- Strengthen the dabtree online scrubber to check hash values on
child dabtree blocks.
- Actually log new intent items created as part of recovering log
intent items.
- Fix a bug where quotas weren't attached to an inode undergoing bmap
intent item recovery.
- Fix a buffer overrun problem with specially crafted log buffer
headers.
- Various cleanups to type usage and slightly inaccurate comments.
- More cleanups to the xattr, log, and quota code.
- Don't run the (slower) shared-rmap operations on attr fork
mappings.
- Fix a bug where we failed to check the LSN of finobt blocks during
replay and could therefore overwrite newer data with older data.
- Clean up the ugly nested transaction mess that log recovery uses to
stage intent item recovery in the correct order by creating a
proper data structure to capture recovered chains.
- Use the capture structure to resume intent item chains with the
same log space and block reservations as when they were captured.
- Fix a UAF bug in bmap intent item recovery where we failed to
maintain our reference to the incore inode if the bmap operation
needed to relog itself to continue.
- Rearrange the defer ops mechanism to finish newly created subtasks
of a parent task before moving on to the next parent task.
- Automatically relog intent items in deferred ops chains if doing so
would help us avoid pinning the log tail. This will help fix some
log scaling problems now and will facilitate atomic file updates
later.
- Fix a deadlock in the GETFSMAP implementation by using an internal
memory buffer to reduce indirect calls and copies to userspace,
thereby improving its performance by ~20%.
- Fix various problems when calling growfs on a realtime volume would
not fully update the filesystem metadata.
- Fix broken Kconfig asking about deprecated XFS when XFS is
disabled"
* tag 'xfs-5.10-merge-5' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (48 commits)
xfs: fix Kconfig asking about XFS_SUPPORT_V4 when XFS_FS=n
xfs: fix high key handling in the rt allocator's query_range function
xfs: annotate grabbing the realtime bitmap/summary locks in growfs
xfs: make xfs_growfs_rt update secondary superblocks
xfs: fix realtime bitmap/summary file truncation when growing rt volume
xfs: fix the indent in xfs_trans_mod_dquot
xfs: do the ASSERT for the arguments O_{u,g,p}dqpp
xfs: fix deadlock and streamline xfs_getfsmap performance
xfs: limit entries returned when counting fsmap records
xfs: only relog deferred intent items if free space in the log gets low
xfs: expose the log push threshold
xfs: periodically relog deferred intent items
xfs: change the order in which child and parent defer ops are finished
xfs: fix an incore inode UAF in xfs_bui_recover
xfs: clean up xfs_bui_item_recover iget/trans_alloc/ilock ordering
xfs: clean up bmap intent item recovery checking
xfs: xfs_defer_capture should absorb remaining transaction reservation
xfs: xfs_defer_capture should absorb remaining block reservations
xfs: proper replay of deferred ops queued during log recovery
xfs: remove XFS_LI_RECOVERED
...
Linus Torvalds [Mon, 19 Oct 2020 21:28:30 +0000 (14:28 -0700)]
Merge tag 'fuse-update-5.10' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse
Pull fuse updates from Miklos Szeredi:
- Support directly accessing host page cache from virtiofs. This can
improve I/O performance for various workloads, as well as reducing
the memory requirement by eliminating double caching. Thanks to Vivek
Goyal for doing most of the work on this.
- Allow automatic submounting inside virtiofs. This allows unique
st_dev/ st_ino values to be assigned inside the guest to files
residing on different filesystems on the host. Thanks to Max Reitz
for the patches.
- Fix an old use after free bug found by Pradeep P V K.
* tag 'fuse-update-5.10' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse: (25 commits)
virtiofs: calculate number of scatter-gather elements accurately
fuse: connection remove fix
fuse: implement crossmounts
fuse: Allow fuse_fill_super_common() for submounts
fuse: split fuse_mount off of fuse_conn
fuse: drop fuse_conn parameter where possible
fuse: store fuse_conn in fuse_req
fuse: add submount support to <uapi/linux/fuse.h>
fuse: fix page dereference after free
virtiofs: add logic to free up a memory range
virtiofs: maintain a list of busy elements
virtiofs: serialize truncate/punch_hole and dax fault path
virtiofs: define dax address space operations
virtiofs: add DAX mmap support
virtiofs: implement dax read/write operations
virtiofs: introduce setupmapping/removemapping commands
virtiofs: implement FUSE_INIT map_alignment field
virtiofs: keep a list of free dax memory ranges
virtiofs: add a mount option to enable dax
virtiofs: set up virtio_fs dax_device
...
Michael Neuling [Tue, 13 Oct 2020 04:37:41 +0000 (15:37 +1100)]
selftests/powerpc: Make alignment handler test P9N DD2.1 vector CI load workaround
alignment_handler currently only tests the unaligned cases but it can
also be useful for testing the workaround for the P9N DD2.1 vector CI
load issue fixed by p9_hmi_special_emu(). This workaround was
introduced in 5080332c2c89 ("powerpc/64s: Add workaround for P9 vector
CI load issue").
This changes the loop to start from offset 0 rather than 1 so that we
test the kernel emulation in p9_hmi_special_emu().
Michael Neuling [Tue, 13 Oct 2020 04:37:40 +0000 (15:37 +1100)]
powerpc: Fix undetected data corruption with P9N DD2.1 VSX CI load emulation
__get_user_atomic_128_aligned() stores to kaddr using stvx which is a
VMX store instruction, hence kaddr must be 16 byte aligned otherwise
the store won't occur as expected.
Unfortunately when we call __get_user_atomic_128_aligned() in
p9_hmi_special_emu(), the buffer we pass as kaddr (ie. vbuf) isn't
guaranteed to be 16B aligned. This means that the write to vbuf in
__get_user_atomic_128_aligned() has the bottom bits of the address
truncated. This results in other local variables being
overwritten. Also vbuf will not contain the correct data which results
in the userspace emulation being wrong and hence undetected user data
corruption.
In the past we've been mostly lucky as vbuf has ended up aligned but
this is fragile and isn't always true. CONFIG_STACKPROTECTOR in
particular can change the stack arrangement enough that our luck runs
out.
This issue only occurs on POWER9 Nimbus <= DD2.1 bare metal.
Linus Torvalds [Mon, 19 Oct 2020 20:52:01 +0000 (13:52 -0700)]
Merge tag 'zonefs-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs
Pull zonefs updates from Damien Le Moal:
"Add an 'explicit-open' mount option to automatically issue a
REQ_OP_ZONE_OPEN command to the device whenever a sequential zone file
is open for writing for the first time.
This avoids 'insufficient zone resources' errors for write operations
on some drives with limited zone resources or on ZNS drives with a
limited number of active zones. From Johannes"
* tag 'zonefs-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs:
zonefs: document the explicit-open mount option
zonefs: open/close zone on file open/close
zonefs: provide no-lock zonefs_io_error variant
zonefs: introduce helper for zone management
rtc: rv3032: Add a driver for Microcrystal RV-3032
New driver for the Microcrystal RV-3032, including support for:
- Date/time
- Alarms
- Low voltage detection
- Trickle charge
- Trimming
- Clkout
- RAM
- EEPROM
- Temperature sensor
Some RTCs have a trickle charge that is able to output different voltages
depending on the type of the connected auxiliary power (battery, supercap,
...). Add a property allowing to specify the necessary voltage.
Shyam Prasad N [Thu, 15 Oct 2020 17:41:31 +0000 (10:41 -0700)]
cifs: Return the error from crypt_message when enc/dec key not found.
In crypt_message, when smb2_get_enc_key returns error, we need to
return the error back to the caller. If not, we end up processing
the message further, causing a kernel oops due to unwarranted access
of memory.
Steve French [Thu, 15 Oct 2020 05:25:02 +0000 (00:25 -0500)]
smb3.1.1: rename nonces used for GCM and CCM encryption
Now that 256 bit encryption can be negotiated, update
names of the nonces to match the updated official protocol
documentation (e.g. AES_GCM_NONCE instead of AES_128GCM_NONCE)
since they apply to both 128 bit and 256 bit encryption.
Steve French [Thu, 15 Oct 2020 05:14:47 +0000 (00:14 -0500)]
smb3.1.1: print warning if server does not support requested encryption type
If server does not support AES-256-GCM and it was required on mount, print
warning message. Also log and return a different error message (EOPNOTSUPP)
when encryption mechanism is not supported vs the case when an unknown
unrequested encryption mechanism could be returned (EINVAL).
Pavel Begunkov [Mon, 19 Oct 2020 15:39:16 +0000 (16:39 +0100)]
io_uring: fix racy REQ_F_LINK_TIMEOUT clearing
io_link_timeout_fn() removes REQ_F_LINK_TIMEOUT from the link head's
flags, it's not atomic and may race with what the head is doing.
If io_link_timeout_fn() doesn't clear the flag, as forced by this patch,
then it may happen that for "req -> link_timeout1 -> link_timeout2",
__io_kill_linked_timeout() would find link_timeout2 and try to cancel
it, so miscounting references. Teach it to ignore such double timeouts
by marking the active one with a new flag in io_prep_linked_timeout().
Pavel Begunkov [Sun, 18 Oct 2020 09:17:40 +0000 (10:17 +0100)]
io_uring: make cached_cq_overflow non atomic_t
ctx->cached_cq_overflow is changed only under completion_lock. Convert
it from atomic_t to just int, and mark all places when it's read without
lock with READ_ONCE, which guarantees atomicity (relaxed ordering).
Pavel Begunkov [Sun, 18 Oct 2020 09:17:38 +0000 (10:17 +0100)]
io_uring: kill ref get/drop in personality init
Don't take an identity on personality/creds init only to drop it a few
lines after. Extract a function which prepares req->work but leaves it
without identity.
Note: it's safe to not check REQ_F_WORK_INITIALIZED there because it's
nobody had a chance to init it before io_init_req().
Chris Wilson [Tue, 15 Sep 2020 13:49:21 +0000 (14:49 +0100)]
drm/i915/gt: Wait for CSB entries on Tigerlake
On Tigerlake, we are seeing a repeat of commit d8f505311717 ("drm/i915/icl:
Forcibly evict stale csb entries") where, presumably, due to a missing
Global Observation Point synchronisation, the write pointer of the CSB
ringbuffer is updated _prior_ to the contents of the ringbuffer. That is
we see the GPU report more context-switch entries for us to parse, but
those entries have not been written, leading us to process stale events,
and eventually report a hung GPU.
However, this effect appears to be much more severe than we previously
saw on Icelake (though it might be best if we try the same approach
there as well and measure), and Bruce suggested the good idea of resetting
the CSB entry after use so that we can detect when it has been updated by
the GPU. By instrumenting how long that may be, we can set a reliable
upper bound for how long we should wait for:
513 late, avg of 61 retries (590 ns), max of 1061 retries (10099 ns)
Arvind Sankar [Fri, 16 Oct 2020 20:04:01 +0000 (16:04 -0400)]
x86/boot/64: Explicitly map boot_params and command line
Commits
ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table") 8570978ea030 ("x86/boot/compressed/64: Don't pre-map memory in KASLR code")
set up a new page table in the decompressor stub, but without explicit
mappings for boot_params and the kernel command line, relying on the #PF
handler instead.
This is fragile, as boot_params and the command line mappings are
required for the main kernel. If EARLY_PRINTK and RANDOMIZE_BASE are
disabled, a QEMU/OVMF boot never accesses the command line in the
decompressor stub, and so it never gets mapped. The main kernel accesses
it from the identity mapping if AMD_MEM_ENCRYPT is enabled, and will
crash.
Fix this by adding back the explicit mapping of boot_params and the
command line.
Note: the changes also removed the explicit mapping of the main kernel,
with the result that .bss and .brk may not be in the identity mapping,
but those don't get accessed by the main kernel before it switches to
its own page tables.
[ bp: Pass boot_params with a MOV %rsp... instead of PUSH/POP. Use
block formatting for the comment. ]
Al Grant [Mon, 21 Sep 2020 20:46:37 +0000 (21:46 +0100)]
perf: correct SNOOPX field offset
perf_event.h has macros that define the field offsets in the
data_src bitmask in perf records. The SNOOPX and REMOTE offsets
were both 37. These are distinct fields, and the bitfield layout
in perf_mem_data_src confirms that SNOOPX should be at offset 38.
Chris Wilson [Fri, 16 Oct 2020 09:25:27 +0000 (10:25 +0100)]
drm/i915: Use the active reference on the vma while capturing
During error capture, we need to take a reference to the vma from before
the reset in order to catpure the contents of the vma later. Currently
we are using both an active reference and a kref, but due to nature of
the i915_vma reference handling, that kref is on the vma->obj and not
the vma itself. This means the vma may be destroyed as soon as it is
idle, that is in between the i915_active_release(&vma->active) and the
i915_vma_put(vma):
Chris Wilson [Fri, 2 Oct 2020 08:34:25 +0000 (09:34 +0100)]
drm/i915/gt: Undo forced context restores after trivial preemptions
We may try to preempt the currently executing request, only to find that
after unravelling all the dependencies that the original executing
context is still the earliest in the topological sort and re-submitted
back to HW (if we do detect some change in the ELSP that requires
re-submission). However, due to the way we check for wrap-around during
the unravelling, we mark any context that has been submitted just once
(i.e. with the rq->wa_tail set, but the ring->tail earlier) as
potentially wrapping and requiring a forced restore on resubmission.
This was expected to be not a problem, as it was anticipated that most
unwinding for preemption would result in a context switch and the few
that did not would be lost in the noise. It did not take long for
someone to find one particular workload where the cost of those extra
context restores was measurable.
However, since we know the wa_tail is of fixed size, and we know that a
request must be larger than the wa_tail itself, we can safely maintain
the check for request wrapping and check against a slightly future point
in the ring that includes an expected wa_tail. (That is if the
ring->tail is already set to rq->wa_tail, including another 8 bytes in
the check does not invalidate the incremental wrap detection.)
Chris Wilson [Thu, 15 Oct 2020 19:50:23 +0000 (20:50 +0100)]
drm/i915/gt: Delay execlist processing for tgl
When running gem_exec_nop, it floods the system with many requests (with
the goal of userspace submitting faster than the HW can process a single
empty batch). This causes the driver to continually resubmit new
requests onto the end of an active context, a flood of lite-restore
preemptions. If we time this just right, Tigerlake hangs.
Inserting a small delay between the processing of CS events and
submitting the next context, prevents the hang. Naturally it does not
occur with debugging enabled. The suspicion then is that this is related
to the issues with the CS event buffer, and inserting an mmio read of
the CS pointer status appears to be very successful in preventing the
hang. Other registers, or uncached reads, or plain mb, do not prevent
the hang, suggesting that register is key -- but that the hang can be
prevented by a simple udelay, suggests it is just a timing issue like
that encountered by commit 233c1ae3c83f ("drm/i915/gt: Wait for CSB
entries on Tigerlake"). Also note that the hang is not prevented by
applying CTX_DESC_FORCE_RESTORE, or by inserting a delay on the GPU
between requests.
Chris Wilson [Thu, 15 Oct 2020 11:59:54 +0000 (12:59 +0100)]
drm/i915/gem: Support parsing of oversize batches
Matthew Auld noted that on more recent systems (such as the parser for
gen9) we may have objects that are larger than expected by the GEM uAPI
(i.e. greater than u32). These objects would have incorrect implicit
batch lengths, causing the parser to reject them for being incomplete,
or worse.
Ville Syrjälä [Thu, 15 Oct 2020 12:21:35 +0000 (13:21 +0100)]
drm/i915: Mark ininitial fb obj as WT on eLLC machines to avoid rcu lockup during fbdev init
Currently we leave the cache_level of the initial fb obj
set to NONE. This means on eLLC machines the first pin_to_display()
will try to switch it to WT which requires a vma unbind+bind.
If that happens during the fbdev initialization rcu does not
seem operational which causes the unbind to get stuck. To
most appearances this looks like a dead machine on boot.
Avoid the unbind by already marking the object cache_level
as WT when creating it. We still do an excplicit ggtt pin
which will rewrite the PTEs anyway, so they will match whatever
cache level we set.
drm/i915/gt: Initialize reserved and unspecified MOCS indices
In order to avoid functional breakage of mis-programmed applications that
have grown to depend on unused MOCS entries, we are programming
those entries to be equal to fully cached ("L3 + LLC") entry.
These reserved and unspecified entries should not be used as they may be
changed to less performant variants with better coherency in the future
if more entries are needed.
v2: As suggested by Lucas De Marchi to utilise __init_mocs_table for
programming default value, setting I915_MOCS_PTE index of tgl_mocs_table
with desired value.
Sean Paul [Fri, 18 Sep 2020 00:28:42 +0000 (20:28 -0400)]
drm/i915/dp: Tweak initial dpcd backlight.enabled value
In commit 79946723092b ("drm/i915: Assume 100% brightness when not in
DPCD control mode"), we fixed the brightness level when DPCD control was
not active to max brightness. This is as good as we can guess since most
backlights go on full when uncontrolled.
However in doing so we changed the semantics of the initial
'backlight.enabled' value. At least on Pixelbooks, they were relying
on the brightness level in DP_EDP_BACKLIGHT_BRIGHTNESS_MSB to be 0 on
boot such that enabled would be false. This causes the device to be
enabled when the brightness is set. Without this, brightness control
doesn't work. So by changing brightness to max, we also flipped enabled
to be true on boot.
To fix this, make enabled a function of brightness and backlight control
mechanism.
Peilin Ye [Sun, 18 Oct 2020 18:12:04 +0000 (14:12 -0400)]
Fonts: Support FONT_EXTRA_WORDS macros for font_6x8
Recently, in commit 6735b4632def ("Fonts: Support FONT_EXTRA_WORDS macros
for built-in fonts"), we wrapped each of our built-in data buffers in a
`font_data` structure, in order to use the following macros on them, see
include/linux/font.h:
Do the same thing to our new 6x8 font. For built-in fonts, currently we
only use FNTSIZE(). Since this is only a temporary solution for an
out-of-bounds issue in the framebuffer layer (see commit 5af08640795b
("fbcon: Fix global-out-of-bounds read in fbcon_get_font()")), all the
three other fields are intentionally set to zero in order to discourage
using these negative-indexing macros.
Wei Wang [Fri, 16 Oct 2020 18:17:22 +0000 (11:17 -0700)]
cpufreq: schedutil: restore cached freq when next_f is not changed
We have the raw cached freq to reduce the chance in calling cpufreq
driver where it could be costly in some arch/SoC.
Currently, the raw cached freq is reset in sugov_update_single() when
it avoids frequency reduction (which is not desirable sometimes), but
it is better to restore the previous value of it in that case,
because it may not change in the next cycle and it is not necessary
to change the CPU frequency then.
Adapted from https://android-review.googlesource.com/1352810/
Wei Huang [Mon, 19 Oct 2020 03:57:41 +0000 (22:57 -0500)]
acpi-cpufreq: Honor _PSD table setting on new AMD CPUs
acpi-cpufreq has a old quirk that overrides the _PSD table supplied by
BIOS on AMD CPUs. However the _PSD table of new AMD CPUs (Family 19h+)
now accurately reports the P-state dependency of CPU cores. Hence this
quirk needs to be fixed in order to support new CPUs' frequency control.
Fixes: acd316248205 ("acpi-cpufreq: Add quirk to disable _PSD usage on all AMD CPUs") Signed-off-by: Wei Huang <[email protected]>
[ rjw: Subject edit ] Cc: 3.10+ <[email protected]> # 3.10+ Signed-off-by: Rafael J. Wysocki <[email protected]>
Jeffle Xu [Mon, 19 Oct 2020 08:59:42 +0000 (16:59 +0800)]
io_uring: use blk_queue_nowait() to check if NOWAIT supported
commit 021a24460dc2 ("block: add QUEUE_FLAG_NOWAIT") adds a new helper
function blk_queue_nowait() to check if the bdev supports handling of
REQ_NOWAIT or not. Since then bio-based dm device can also support
REQ_NOWAIT, and currently only dm-linear supports that since
commit 6abc49468eea ("dm: add support for REQ_NOWAIT and enable it for
linear target").
Alex Williamson [Mon, 19 Oct 2020 13:13:55 +0000 (07:13 -0600)]
vfio/pci: Clear token on bypass registration failure
The eventfd context is used as our irqbypass token, therefore if an
eventfd is re-used, our token is the same. The irqbypass code will
return an -EBUSY in this case, but we'll still attempt to unregister
the producer, where if that duplicate token still exists, results in
removing the wrong object. Clear the token of failed producers so
that they harmlessly fall out when unregistered.
Diana Craciun [Fri, 16 Oct 2020 09:32:32 +0000 (12:32 +0300)]
vfio/fsl-mc: fix the return of the uninitialized variable ret
The vfio_fsl_mc_reflck_attach function may return, on success path,
an uninitialized variable. Fix the problem by initializing the return
variable to 0.
Addresses-Coverity: ("Uninitialized scalar variable") Fixes: f2ba7e8c947b ("vfio/fsl-mc: Added lock support in preparation for interrupt handling") Reported-by: Colin Ian King <[email protected]> Signed-off-by: Diana Craciun <[email protected]> Signed-off-by: Alex Williamson <[email protected]>
ACPICA: Add missing type casts in GPE register access code
Type casts needed on 32-bit systems are missing in two places in the
GPE register access code, so add them.
Fixes: 7a8379eb41a4 ("ACPICA: Add support for using logical addresses of GPE blocks") Reported-and-tested-by: Matthieu Baerts <[email protected]> Signed-off-by: Rafael J. Wysocki <[email protected]>
iommu/vt-d: Don't dereference iommu_device if IOMMU_API is not built
Since commit c40aaaac1018 ("iommu/vt-d: Gracefully handle DMAR units
with no supported address widths") dmar.c needs struct iommu_device to
be selected. We can drop this dependency by not dereferencing struct
iommu_device if IOMMU_API is not selected and by reusing the information
stored in iommu->drhd->ignored instead.
This fixes the following build error when IOMMU_API is not selected:
drivers/iommu/intel/dmar.c: In function ‘free_iommu’:
drivers/iommu/intel/dmar.c:1139:41: error: ‘struct iommu_device’ has no member named ‘ops’
1139 | if (intel_iommu_enabled && iommu->iommu.ops) {
^
Vasant Hegde [Sat, 17 Oct 2020 16:42:36 +0000 (22:12 +0530)]
powerpc/powernv/dump: Handle multiple writes to ack attribute
Even though we use self removing sysfs helper, we still need
to make sure we do the final kobject delete conditionally.
sysfs_remove_file_self() will handle parallel calls to remove
the sysfs attribute file and returns true only in the caller
that removed the attribute file. The other parallel callers
are returned false. Do the final kobject delete checking
the return value of sysfs_remove_file_self().
Vasant Hegde [Sat, 17 Oct 2020 16:42:10 +0000 (22:12 +0530)]
powerpc/powernv/dump: Fix race while processing OPAL dump
Every dump reported by OPAL is exported to userspace through a sysfs
interface and notified using kobject_uevent(). The userspace daemon
(opal_errd) then reads the dump and acknowledges that the dump is
saved safely to disk. Once acknowledged the kernel removes the
respective sysfs file entry causing respective resources to be
released including kobject.
However it's possible the userspace daemon may already be scanning
dump entries when a new sysfs dump entry is created by the kernel.
User daemon may read this new entry and ack it even before kernel can
notify userspace about it through kobject_uevent() call. If that
happens then we have a potential race between
dump_ack_store->kobject_put() and kobject_uevent which can lead to
use-after-free of a kernfs object resulting in a kernel crash.
This patch fixes this race by protecting the sysfs file
creation/notification by holding a reference count on kobject until we
safely send kobject_uevent().
The function create_dump_obj() returns the dump object which if used
by caller function will end up in use-after-free problem again.
However, the return value of create_dump_obj() function isn't being
used today and there is no need as well. Hence change it to return
void to make this fix complete.
has stack protection enabled because of set_bringup_idt_handler().
This happens when CONFIG_STACKPROTECTOR_STRONG is enabled. It
also currently needs CONFIG_AMD_MEM_ENCRYPT enabled because then
set_bringup_idt_handler() is not an empty stub but that might change in
the future, when the other vendor adds their similar technology.
At this point, %gs is not yet initialized, and this doesn't cause a
crash only because the #PF handler from the decompressor stub is still
installed and handles the page fault.
Disable stack protection for the whole file, and do it on 32-bit as
well to avoid surprises.
[ bp: Extend commit message with the exact explanation how it happens. ]
ca0e22d4f011 ("x86/boot/compressed/64: Always switch to own page table")
started using a new set of pagetables even without KASLR.
After that commit, initialize_identity_maps() is called before the
5-level paging variables are setup in choose_random_location(), which
will not work if 5-level paging is actually enabled.
Fix this by moving the initialization of __pgtable_l5_enabled,
pgdir_shift and ptrs_per_p4d into cleanup_trampoline(), which is called
immediately after the finalization of whether the kernel is executing
with 4- or 5-level paging. This will be earlier than anything that might
require those variables, and keeps the 4- vs 5-level paging code all in
one place.
powerpc/smp: Use GFP_ATOMIC while allocating tmp mask
Qian Cai reported a regression where CPU Hotplug fails with the latest
powerpc/next
BUG: sleeping function called from invalid context at mm/slab.h:494
in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 0, name: swapper/88
no locks held by swapper/88/0.
irq event stamp: 18074448
hardirqs last enabled at (18074447): [<c0000000001a2a7c>] tick_nohz_idle_enter+0x9c/0x110
hardirqs last disabled at (18074448): [<c000000000106798>] do_idle+0x138/0x3b0
do_idle at kernel/sched/idle.c:253 (discriminator 1)
softirqs last enabled at (18074440): [<c0000000000bbec4>] irq_enter_rcu+0x94/0xa0
softirqs last disabled at (18074439): [<c0000000000bbea0>] irq_enter_rcu+0x70/0xa0
CPU: 88 PID: 0 Comm: swapper/88 Tainted: G W 5.9.0-rc8-next-20201007 #1
Call Trace:
[c00020000a4bfcf0] [c000000000649e98] dump_stack+0xec/0x144 (unreliable)
[c00020000a4bfd30] [c0000000000f6c34] ___might_sleep+0x2f4/0x310
[c00020000a4bfdb0] [c000000000354f94] slab_pre_alloc_hook.constprop.82+0x124/0x190
[c00020000a4bfe00] [c00000000035e9e8] __kmalloc_node+0x88/0x3a0
slab_alloc_node at mm/slub.c:2817
(inlined by) __kmalloc_node at mm/slub.c:4013
[c00020000a4bfe80] [c0000000006494d8] alloc_cpumask_var_node+0x38/0x80
kmalloc_node at include/linux/slab.h:577
(inlined by) alloc_cpumask_var_node at lib/cpumask.c:116
[c00020000a4bfef0] [c00000000003eedc] start_secondary+0x27c/0x800
update_mask_by_l2 at arch/powerpc/kernel/smp.c:1267
(inlined by) add_cpu_to_masks at arch/powerpc/kernel/smp.c:1387
(inlined by) start_secondary at arch/powerpc/kernel/smp.c:1420
[c00020000a4bff90] [c00000000000c468] start_secondary_resume+0x10/0x14
Allocating a temporary mask while performing a CPU Hotplug operation
with CONFIG_CPUMASK_OFFSTACK enabled, leads to calling a sleepable
function from a atomic context. Fix this by allocating the temporary
mask with GFP_ATOMIC flag. Also instead of having to allocate twice,
allocate the mask in the caller so that we only have to allocate once.
If the allocation fails, assume the mask to be same as sibling mask, which
will make the scheduler to drop this domain for this CPU.
Commit 3ab33d6dc3e9 ("powerpc/smp: Optimize update_mask_by_l2")
introduced submask_fn in update_mask_by_l2 to track the right submask.
However commit f6606cfdfbcd ("powerpc/smp: Dont assume l2-cache to be
superset of sibling") introduced sibling_mask in update_mask_by_l2 to
track the same submask. Remove sibling_mask in favour of submask_fn.
Taehee Yoo [Thu, 15 Oct 2020 16:26:06 +0000 (16:26 +0000)]
net: core: use list_del_init() instead of list_del() in netdev_run_todo()
dev->unlink_list is reused unless dev is deleted.
So, list_del() should not be used.
Due to using list_del(), dev->unlink_list can't be reused so that
dev->nested_level update logic doesn't work.
In order to fix this bug, list_del_init() should be used instead
of list_del().
Test commands:
ip link add bond0 type bond
ip link add bond1 type bond
ip link set bond0 master bond1
ip link set bond0 nomaster
ip link set bond1 master bond0
ip link set bond1 nomaster
Linus Torvalds [Sun, 18 Oct 2020 21:45:59 +0000 (14:45 -0700)]
Merge tag 'linux-kselftest-kunit-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest
Pull more Kunit updates from Shuah Khan:
- add Kunit to kernel_init() and remove KUnit from init calls entirely.
This addresses the concern that Kunit would not work correctly during
late init phase.
- add a linker section where KUnit can put references to its test
suites.
This is the first step in transitioning to dispatching all KUnit
tests from a centralized executor rather than having each as its own
separate late_initcall.
- add a centralized executor to dispatch tests rather than relying on
late_initcall to schedule each test suite separately. Centralized
execution is for built-in tests only; modules will execute tests when
loaded.
- convert bitfield test to use KUnit framework
- Documentation updates for naming guidelines and how
kunit_test_suite() works.
- add test plan to KUnit TAP format
* tag 'linux-kselftest-kunit-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
lib: kunit: Fix compilation test when using TEST_BIT_FIELD_COMPILE
lib: kunit: add bitfield test conversion to KUnit
Documentation: kunit: add a brief blurb about kunit_test_suite
kunit: test: add test plan to KUnit TAP format
init: main: add KUnit to kernel init
kunit: test: create a single centralized executor for all tests
vmlinux.lds.h: add linker section for KUnit test suites
Documentation: kunit: Add naming guidelines
Linus Torvalds [Sun, 18 Oct 2020 21:34:50 +0000 (14:34 -0700)]
Merge tag 'core-rcu-2020-10-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull RCU changes from Ingo Molnar:
- Debugging for smp_call_function()
- RT raw/non-raw lock ordering fixes
- Strict grace periods for KASAN
- New smp_call_function() torture test
- Torture-test updates
- Documentation updates
- Miscellaneous fixes
[ This doesn't actually pull the tag - I've dropped the last merge from
the RCU branch due to questions about the series. - Linus ]
* tag 'core-rcu-2020-10-12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (77 commits)
smp: Make symbol 'csd_bug_count' static
kernel/smp: Provide CSD lock timeout diagnostics
smp: Add source and destination CPUs to __call_single_data
rcu: Shrink each possible cpu krcp
rcu/segcblist: Prevent useless GP start if no CBs to accelerate
torture: Add gdb support
rcutorture: Allow pointer leaks to test diagnostic code
rcutorture: Hoist OOM registry up one level
refperf: Avoid null pointer dereference when buf fails to allocate
rcutorture: Properly synchronize with OOM notifier
rcutorture: Properly set rcu_fwds for OOM handling
torture: Add kvm.sh --help and update help message
rcutorture: Add CONFIG_PROVE_RCU_LIST to TREE05
torture: Update initrd documentation
rcutorture: Replace HTTP links with HTTPS ones
locktorture: Make function torture_percpu_rwsem_init() static
torture: document --allcpus argument added to the kvm.sh script
rcutorture: Output number of elapsed grace periods
rcutorture: Remove KCSAN stubs
rcu: Remove unused "cpu" parameter from rcu_report_qs_rdp()
...
Linus Torvalds [Sun, 18 Oct 2020 21:29:19 +0000 (14:29 -0700)]
Merge tag 'mailbox-v5.10' of git://git.linaro.org/landing-teams/working/fujitsu/integration
Pull mailbox updates from Jassi Brar:
- arm: implementation of mhu as a doorbell driver and conversion of
dt-bindings to json-schema
- mediatek: fix platform_get_irq error handling
- bcm: convert tasklets to use new tasklet_setup api
- core: fix race cause by hrtimer starting inappropriately
* tag 'mailbox-v5.10' of git://git.linaro.org/landing-teams/working/fujitsu/integration:
mailbox: avoid timer start from callback
maiblox: mediatek: Fix handling of platform_get_irq() error
mailbox: arm_mhu: Add ARM MHU doorbell driver
mailbox: arm_mhu: Match only if compatible is "arm,mhu"
dt-bindings: mailbox: add doorbell support to ARM MHU
dt-bindings: mailbox : arm,mhu: Convert to Json-schema
mailbox: bcm: convert tasklets to use new tasklet_setup() API
Jakub Kicinski [Fri, 16 Oct 2020 23:20:06 +0000 (16:20 -0700)]
ixgbe: fix probing of multi-port devices with one MDIO
Ian reports that after upgrade from v5.8.14 to v5.9 only one
of his 4 ixgbe netdevs appear in the system.
Quoting the comment on ixgbe_x550em_a_has_mii():
* Returns true if hw points to lowest numbered PCI B:D.F x550_em_a device in
* the SoC. There are up to 4 MACs sharing a single MDIO bus on the x550em_a,
* but we only want to register one MDIO bus.
This matches the symptoms, since the return value from
ixgbe_mii_bus_init() is no longer ignored we need to handle
the higher ports of x550em without an error.
net: usb: rtl8150: don't incorrectly assign random MAC addresses
In set_ethernet_addr(), if get_registers() succeeds, the ethernet address
that was read must be copied over. Otherwise, a random ethernet address
must be assigned.
get_registers() returns 0 if successful, and negative error number
otherwise. However, in set_ethernet_addr(), this return value is
incorrectly checked.
Since this return value will never be equal to sizeof(node_id), a
random MAC address will always be generated and assigned to the
device; even in cases when get_registers() is successful.
Correctly modifying the condition that checks if get_registers() was
successful or not fixes this problem, and copies the ethernet address
appropriately.
When 'rp_filter' is configured in strict mode (1) the tests fail because
packets received from the macvlan netdevs would not be forwarded through
them on the reverse path.
Fix this by disabling the 'rp_filter', meaning no source validation is
performed.
Eelco Chaudron [Sat, 17 Oct 2020 18:24:51 +0000 (20:24 +0200)]
net: openvswitch: fix to make sure flow_lookup() is not preempted
The flow_lookup() function uses per CPU variables, which must be called
with BH disabled. However, this is fine in the general NAPI use case
where the local BH is disabled. But, it's also called from the netlink
context. The below patch makes sure that even in the netlink path, the
BH is disabled.
In addition, u64_stats_update_begin() requires a lock to ensure one writer
which is not ensured here. Making it per-CPU and disabling NAPI (softirq)
ensures that there is always only one writer.
Linus Torvalds [Sun, 18 Oct 2020 19:25:25 +0000 (12:25 -0700)]
Merge branch 'akpm' (patches from Andrew)
Merge yet more updates from Andrew Morton:
"Subsystems affected by this patch series: mm (memcg, migration,
pagemap, gup, madvise, vmalloc), ia64, and misc"
* emailed patches from Andrew Morton <[email protected]>: (31 commits)
mm: remove duplicate include statement in mmu.c
mm: remove the filename in the top of file comment in vmalloc.c
mm: cleanup the gfp_mask handling in __vmalloc_area_node
mm: remove alloc_vm_area
x86/xen: open code alloc_vm_area in arch_gnttab_valloc
xen/xenbus: use apply_to_page_range directly in xenbus_map_ring_pv
drm/i915: use vmap in i915_gem_object_map
drm/i915: stop using kmap in i915_gem_object_map
drm/i915: use vmap in shmem_pin_map
zsmalloc: switch from alloc_vm_area to get_vm_area
mm: allow a NULL fn callback in apply_to_page_range
mm: add a vmap_pfn function
mm: add a VM_MAP_PUT_PAGES flag for vmap
mm: update the documentation for vfree
mm/madvise: introduce process_madvise() syscall: an external memory hinting API
pid: move pidfd_get_pid() to pid.c
mm/madvise: pass mm to do_madvise
selftests/vm: 10x speedup for hmm-tests
binfmt_elf: take the mmap lock around find_extend_vma()
mm/gup_benchmark: take the mmap lock around GUP
...
Linus Torvalds [Sun, 18 Oct 2020 17:03:23 +0000 (10:03 -0700)]
Merge tag 'for-linus-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml
Pull UML updates from Richard Weinberger:
- Improve support for non-glibc systems
- Vector: Add support for scripting and dynamic tap devices
- Various fixes for the vector networking driver
- Various fixes for time travel mode
* tag 'for-linus-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml:
um: vector: Add dynamic tap interfaces and scripting
um: Clean up stacktrace dump
um: Fix incorrect assumptions about max pid length
um: Remove dead usage of TIF_IA32
um: Remove redundant NULL check
um: change sigio_spinlock to a mutex
um: time-travel: Return the sequence number in ACK messages
um: time-travel: Fix IRQ handling in time_travel_handle_message()
um: Allow static linking for non-glibc implementations
um: Some fixes to build UML with musl
um: vector: Use GFP_ATOMIC under spin lock
um: Fix null pointer dereference in vector_user_bpf
Linus Torvalds [Sun, 18 Oct 2020 16:56:50 +0000 (09:56 -0700)]
Merge tag 'for-linus-5.10-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs
Pull more ubi and ubifs updates from Richard Weinberger:
"UBI:
- Correctly use kthread_should_stop in ubi worker
UBIFS:
- Fixes for memory leaks while iterating directory entries
- Fix for a user triggerable error message
- Fix for a space accounting bug in authenticated mode"
* tag 'for-linus-5.10-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
ubifs: journal: Make sure to not dirty twice for auth nodes
ubifs: setflags: Don't show error message when vfs_ioc_setflags_prepare() fails
ubifs: ubifs_jnl_change_xattr: Remove assertion 'nlink > 0' for host inode
ubi: check kthread_should_stop() after the setting of task state
ubifs: dent: Fix some potential memory leaks while iterating entries
ubifs: xattr: Fix some potential memory leaks while iterating entries
Linus Torvalds [Sun, 18 Oct 2020 16:51:10 +0000 (09:51 -0700)]
Merge tag 'for-linus-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs
Pull ubifs updates from Richard Weinberger:
- Kernel-doc fixes
- Fixes for memory leaks in authentication option parsing
* tag 'for-linus-5.10-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/ubifs:
ubifs: mount_ubifs: Release authentication resource in error handling path
ubifs: Don't parse authentication mount options in remount process
ubifs: Fix a memleak after dumping authentication mount options
ubifs: Fix some kernel-doc warnings in tnc.c
ubifs: Fix some kernel-doc warnings in replay.c
ubifs: Fix some kernel-doc warnings in gc.c
ubifs: Fix 'hash' kernel-doc warning in auth.c
mm: cleanup the gfp_mask handling in __vmalloc_area_node
Patch series "two small vmalloc cleanups".
This patch (of 2):
__vmalloc_area_node currently has four different gfp_t variables to
just express this simple logic:
- use the passed in mask, plus __GFP_NOWARN and __GFP_HIGHMEM (if
suitable) for the underlying page allocation
- use just the reclaim flags from the passed in mask plus __GFP_ZERO
for allocating the page array
Simplify this down to just use the pre-existing nested_gfp as-is for
the page array allocation, and just the passed in gfp_mask for the
page allocation, after conditionally ORing __GFP_HIGHMEM into it. This
also makes the allocation warning a little more correct.
Also initialize two variables at the time of declaration while touching
this area.
x86/xen: open code alloc_vm_area in arch_gnttab_valloc
Replace the last call to alloc_vm_area with an open coded version using an
iterator in struct gnttab_vm_area instead of the triple indirection magic
in alloc_vm_area.
xen/xenbus: use apply_to_page_range directly in xenbus_map_ring_pv
Replacing alloc_vm_area with get_vm_area_caller + apply_page_range allows
to fill put the phys_addr values directly instead of doing another loop
over all addresses.
i915_gem_object_map implements fairly low-level vmap functionality in a
driver. Split it into two helpers, one for remapping kernel memory which
can use vmap, and one for I/O memory that uses vmap_pfn.
The only practical difference is that alloc_vm_area prefeaults the vmalloc
area PTEs, which doesn't seem to be required here for the kernel memory
case (and could be added to vmap using a flag if actually required).
shmem_pin_map somewhat awkwardly reimplements vmap using alloc_vm_area and
manual pte setup. The only practical difference is that alloc_vm_area
prefeaults the vmalloc area PTEs, which doesn't seem to be required here
(and could be added to vmap using a flag if actually required). Switch to
use vmap, and use vfree to free both the vmalloc mapping and the page
array, as well as dropping the references to each page.
mm: allow a NULL fn callback in apply_to_page_range
Besides calling the callback on each page, apply_to_page_range also has
the effect of pre-faulting all PTEs for the range. To support callers
that only need the pre-faulting, make the callback optional.
Add a flag so that vmap takes ownership of the passed in page array. When
vfree is called on such an allocation it will put one reference on each
page, and free the page array itself.
This series removes alloc_vm_area, which was left over from the big
vmalloc interface rework. It is a rather arkane interface, basicaly the
equivalent of get_vm_area + actually faulting in all PTEs in the allocated
area. It was originally addeds for Xen (which isn't modular to start
with), and then grew users in zsmalloc and i915 which seems to mostly
qualify as abuses of the interface, especially for i915 as a random driver
should not set up PTE bits directly.
This patch (of 11):
* Document that you can call vfree() on an address returned from vmap()
* Remove the note about the minimum size -- the minimum size of a vmalloc
allocation is one page
* Add a Context: section
* Fix capitalisation
* Reword the prohibition on calling from NMI context to avoid a double
negative
Minchan Kim [Sat, 17 Oct 2020 23:14:59 +0000 (16:14 -0700)]
mm/madvise: introduce process_madvise() syscall: an external memory hinting API
There is usecase that System Management Software(SMS) want to give a
memory hint like MADV_[COLD|PAGEEOUT] to other processes and in the
case of Android, it is the ActivityManagerService.
The information required to make the reclaim decision is not known to the
app. Instead, it is known to the centralized userspace
daemon(ActivityManagerService), and that daemon must be able to initiate
reclaim on its own without any app involvement.
To solve the issue, this patch introduces a new syscall
process_madvise(2). It uses pidfd of an external process to give the
hint. It also supports vector address range because Android app has
thousands of vmas due to zygote so it's totally waste of CPU and power if
we should call the syscall one by one for each vma.(With testing 2000-vma
syscall vs 1-vector syscall, it showed 15% performance improvement. I
think it would be bigger in real practice because the testing ran very
cache friendly environment).
Another potential use case for the vector range is to amortize the cost
ofTLB shootdowns for multiple ranges when using MADV_DONTNEED; this could
benefit users like TCP receive zerocopy and malloc implementations. In
future, we could find more usecases for other advises so let's make it
happens as API since we introduce a new syscall at this moment. With
that, existing madvise(2) user could replace it with process_madvise(2)
with their own pid if they want to have batch address ranges support
feature.
ince it could affect other process's address range, only privileged
process(PTRACE_MODE_ATTACH_FSCREDS) or something else(e.g., being the same
UID) gives it the right to ptrace the process could use it successfully.
The flag argument is reserved for future use if we need to extend the API.
I think supporting all hints madvise has/will supported/support to
process_madvise is rather risky. Because we are not sure all hints make
sense from external process and implementation for the hint may rely on
the caller being in the current context so it could be error-prone. Thus,
I just limited hints as MADV_[COLD|PAGEOUT] in this patch.
If someone want to add other hints, we could hear the usecase and review
it for each hint. It's safer for maintenance rather than introducing a
buggy syscall but hard to fix it later.
So finally, the API is as follows,
ssize_t process_madvise(int pidfd, const struct iovec *iovec,
unsigned long vlen, int advice, unsigned int flags);
DESCRIPTION
The process_madvise() system call is used to give advice or directions
to the kernel about the address ranges from external process as well as
local process. It provides the advice to address ranges of process
described by iovec and vlen. The goal of such advice is to improve
system or application performance.
The pidfd selects the process referred to by the PID file descriptor
specified in pidfd. (See pidofd_open(2) for further information)
The pointer iovec points to an array of iovec structures, defined in
<sys/uio.h> as:
struct iovec {
void *iov_base; /* starting address */
size_t iov_len; /* number of bytes to be advised */
};
The iovec describes address ranges beginning at address(iov_base)
and with size length of bytes(iov_len).
The vlen represents the number of elements in iovec.
The advice is indicated in the advice argument, which is one of the
following at this moment if the target process specified by pidfd is
external.
MADV_COLD
MADV_PAGEOUT
Permission to provide a hint to external process is governed by a
ptrace access mode PTRACE_MODE_ATTACH_FSCREDS check; see ptrace(2).
The process_madvise supports every advice madvise(2) has if target
process is in same thread group with calling process so user could
use process_madvise(2) to extend existing madvise(2) to support
vector address ranges.
RETURN VALUE
On success, process_madvise() returns the number of bytes advised.
This return value may be less than the total number of requested
bytes, if an error occurred. The caller should check return value
to determine whether a partial advice occurred.
FAQ:
Q.1 - Why does any external entity have better knowledge?
Quote from Sandeep
"For Android, every application (including the special SystemServer)
are forked from Zygote. The reason of course is to share as many
libraries and classes between the two as possible to benefit from the
preloading during boot.
After applications start, (almost) all of the APIs end up calling into
this SystemServer process over IPC (binder) and back to the
application.
In a fully running system, the SystemServer monitors every single
process periodically to calculate their PSS / RSS and also decides
which process is "important" to the user for interactivity.
So, because of how these processes start _and_ the fact that the
SystemServer is looping to monitor each process, it does tend to *know*
which address range of the application is not used / useful.
Besides, we can never rely on applications to clean things up
themselves. We've had the "hey app1, the system is low on memory,
please trim your memory usage down" notifications for a long time[1].
They rely on applications honoring the broadcasts and very few do.
So, if we want to avoid the inevitable killing of the application and
restarting it, some way to be able to tell the OS about unimportant
memory in these applications will be useful.
- ssp
Q.2 - How to guarantee the race(i.e., object validation) between when
giving a hint from an external process and get the hint from the target
process?
process_madvise operates on the target process's address space as it
exists at the instant that process_madvise is called. If the space
target process can run between the time the process_madvise process
inspects the target process address space and the time that
process_madvise is actually called, process_madvise may operate on
memory regions that the calling process does not expect. It's the
responsibility of the process calling process_madvise to close this
race condition. For example, the calling process can suspend the
target process with ptrace, SIGSTOP, or the freezer cgroup so that it
doesn't have an opportunity to change its own address space before
process_madvise is called. Another option is to operate on memory
regions that the caller knows a priori will be unchanged in the target
process. Yet another option is to accept the race for certain
process_madvise calls after reasoning that mistargeting will do no
harm. The suggested API itself does not provide synchronization. It
also apply other APIs like move_pages, process_vm_write.
The race isn't really a problem though. Why is it so wrong to require
that callers do their own synchronization in some manner? Nobody
objects to write(2) merely because it's possible for two processes to
open the same file and clobber each other's writes --- instead, we tell
people to use flock or something. Think about mmap. It never
guarantees newly allocated address space is still valid when the user
tries to access it because other threads could unmap the memory right
before. That's where we need synchronization by using other API or
design from userside. It shouldn't be part of API itself. If someone
needs more fine-grained synchronization rather than process level,
there were two ideas suggested - cookie[2] and anon-fd[3]. Both are
applicable via using last reserved argument of the API but I don't
think it's necessary right now since we have already ways to prevent
the race so don't want to add additional complexity with more
fine-grained optimization model.
To make the API extend, it reserved an unsigned long as last argument
so we could support it in future if someone really needs it.
Q.3 - Why doesn't ptrace work?
Injecting an madvise in the target process using ptrace would not work
for us because such injected madvise would have to be executed by the
target process, which means that process would have to be runnable and
that creates the risk of the abovementioned race and hinting a wrong
VMA. Furthermore, we want to act the hint in caller's context, not the
callee's, because the callee is usually limited in cpuset/cgroups or
even freezed state so they can't act by themselves quick enough, which
causes more thrashing/kill. It doesn't work if the target process are
ptraced(e.g., strace, debugger, minidump) because a process can have at
most one ptracer.
[2] process_getinfo for getting the cookie which is updated whenever
vma of process address layout are changed - Daniel Colascione -
https://lore.kernel.org/lkml/20190520035254[email protected]/T/#m7694416fd179b2066a2c62b5b139b14e3894e224
[3] anonymous fd which is used for the object(i.e., address range)
validation - Michal Hocko -
https://lore.kernel.org/lkml/20200120112722[email protected]/
Minchan Kim [Sat, 17 Oct 2020 23:14:50 +0000 (16:14 -0700)]
mm/madvise: pass mm to do_madvise
Patch series "introduce memory hinting API for external process", v9.
Now, we have MADV_PAGEOUT and MADV_COLD as madvise hinting API. With
that, application could give hints to kernel what memory range are
preferred to be reclaimed. However, in some platform(e.g., Android), the
information required to make the hinting decision is not known to the app.
Instead, it is known to a centralized userspace daemon(e.g.,
ActivityManagerService), and that daemon must be able to initiate reclaim
on its own without any app involvement.
To solve the concern, this patch introduces new syscall -
process_madvise(2). Bascially, it's same with madvise(2) syscall but it
has some differences.
1. It needs pidfd of target process to provide the hint
2. It supports only MADV_{COLD|PAGEOUT|MERGEABLE|UNMEREABLE} at this
moment. Other hints in madvise will be opened when there are explicit
requests from community to prevent unexpected bugs we couldn't support.
3. Only privileged processes can do something for other process's
address space.
For more detail of the new API, please see "mm: introduce external memory
hinting API" description in this patchset.
This patch (of 3):
In upcoming patches, do_madvise will be called from external process
context so we shouldn't asssume "current" is always hinted process's
task_struct.
Furthermore, we must not access mm_struct via task->mm, but obtain it via
access_mm() once (in the following patch) and only use that pointer [1],
so pass it to do_madvise() as well. Note the vma->vm_mm pointers are
safe, so we can use them further down the call stack.
And let's pass current->mm as arguments of do_madvise so it shouldn't
change existing behavior but prepare next patch to make review easy.
John Hubbard [Sat, 17 Oct 2020 23:14:47 +0000 (16:14 -0700)]
selftests/vm: 10x speedup for hmm-tests
This patch reduces the running time for hmm-tests from about 10+ seconds,
to just under 1.0 second, for an approximately 10x speedup. That brings
it in line with most of the other tests in selftests/vm, which mostly run
in < 1 sec.
This is done with a one-line change that simply reduces the number of
iterations of several tests, from 256, to 10. Thanks to Ralph Campbell
for suggesting changing NTIMES as a way to get the speedup.
Jann Horn [Sat, 17 Oct 2020 23:14:15 +0000 (16:14 -0700)]
binfmt_elf: take the mmap lock around find_extend_vma()
create_elf_tables() runs after setup_new_exec(), so other tasks can
already access our new mm and do things like process_madvise() on it. (At
the time I'm writing this commit, process_madvise() is not in mainline
yet, but has been in akpm's tree for some time.)
While I believe that there are currently no APIs that would actually allow
another process to mess up our VMA tree (process_madvise() is limited to
MADV_COLD and MADV_PAGEOUT, and uring and userfaultfd cannot reach an mm
under which no syscalls have been executed yet), this seems like an
accident waiting to happen.
Let's make sure that we always take the mmap lock around GUP paths as long
as another process might be able to see the mm.
(Yes, this diff looks suspicious because we drop the lock before doing
anything with `vma`, but that's because we actually don't do anything with
it apart from the NULL check.)
projects / linux.git / log