When checking a memory buffer to be consecutive in machine memory,
the alignment needs to be checked, too. Failing to do so might result
in DMA memory not being aligned according to its requested size,
leading to error messages like:
4xxx 0000:2b:00.0: enabling device (0140 -> 0142)
4xxx 0000:2b:00.0: Ring address not aligned
4xxx 0000:2b:00.0: Failed to initialise service qat_crypto
4xxx 0000:2b:00.0: Resetting device qat_dev0
4xxx: probe of 0000:2b:00.0 failed with error -14
In order to minimize required special handling for running as Xen PV
dom0, the memory layout is modified to match that of the host. This
requires to have only RAM at the locations where Xen allocated memory
is living. Unfortunately there seem to be some machines, where ACPI
NVS is located at 64 MB, resulting in a conflict with the loaded
kernel or the initial page tables built by Xen.
Avoid this conflict by swapping the ACPI NVS area in the memory map
with unused RAM. This is possible via modification of the dom0 P2M map.
Accesses to the ACPI NVS area are done either for saving and restoring
it across suspend operations (this will work the same way as before),
or by ACPI code when NVS memory is referenced from other ACPI tables.
The latter case is handled by a Xen specific indirection of
acpi_os_ioremap().
While the E820 map can (and should) be modified right away, the P2M
map can be updated only after memory allocation is working, as the P2M
map might need to be extended.
Fixes: 808fdb71936c ("xen: check for kernel memory conflicting with memory layout") Signed-off-by: Juergen Gross <jgross@suse.com> Tested-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Juergen Gross <jgross@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When running as a Xen PV dom0 it can happen that the kernel is being
loaded to a guest physical address conflicting with the host memory
map.
In order to be able to resolve this conflict, add the capability to
remap non-RAM areas to different guest PFNs. A function to use this
remapping information for other purposes than doing the remap will be
added when needed.
As the number of conflicts should be rather low (currently only
machines with max. 1 conflict are known), save the remap data in a
small statically allocated array.
Instead of having max_pfn as a local variable of xen_memory_setup(),
make it a static variable in setup.c instead. This avoids having to
pass it to subfunctions, which will be needed in more cases in future.
Rename it to ini_nr_pages, as the value denotes the currently usable
number of memory pages as passed from the hypervisor at boot time.
When booting as a Xen PV dom0 the memory layout of the dom0 is
modified to match that of the host, as this requires less changes in
the kernel for supporting Xen.
There are some cases, though, which are problematic, as it is the Xen
hypervisor selecting the kernel's load address plus some other data,
which might conflict with the host's memory map.
These conflicts are detected at boot time and result in a boot error.
In order to support handling at least some of these conflicts in
future, introduce a generic helper function which will later gain the
ability to adapt the memory layout when possible.
Add the missing check for the xen_start_info area.
Note that possible p2m map and initrd memory conflicts are handled
already by copying the data to memory areas not conflicting with the
memory map. The initial stack allocated by Xen doesn't need to be
checked, as early boot code is switching to the statically allocated
initial kernel stack. Initial page tables and the kernel itself will
be handled later.
We have some very fancy min/max macros that have tons of sanity checking
to warn about mixed signedness etc.
This is all things that a sane compiler should warn about, but there are
no sane compiler interfaces for this, and '-Wsign-compare' is broken [1]
and not useful.
So then we compensate (some would say over-compensate) by doing the
checks manually with some truly horrid macro games.
And no, we can't just use __builtin_types_compatible_p(), because the
whole question of "does it make sense to compare these two values" is a
lot more complicated than that.
For example, it makes a ton of sense to compare unsigned values with
simple constants like "5", even if that is indeed a signed type. So we
have these very strange macros to try to make sensible type checking
decisions on the arguments to 'min()' and 'max()'.
But that can cause enormous code expansion if the min()/max() macros are
used with complicated expressions, and particularly if you nest these
things so that you get the first big expansion then expanded again.
The xen setup.c file ended up ballooning to over 50MB of preprocessed
noise that takes 15s to compile (obviously depending on the build host),
largely due to one single line.
So let's split that one single line to just be simpler. I think it ends
up being more legible to humans too at the same time. Now that single
file compiles in under a second.
When ata_qc_complete() schedules a command for EH using
ata_qc_schedule_eh(), blk_abort_request() will be called, which leads to
req->q->mq_ops->timeout() / scsi_timeout() being called.
scsi_timeout(), if the LLDD has no abort handler (libata has no abort
handler), will set host byte to DID_TIME_OUT, and then call
scsi_eh_scmd_add() to add the command to EH.
Thus, when commands first enter libata's EH strategy_handler, all the
commands that have been added to EH will have DID_TIME_OUT set.
libata has its own flag (AC_ERR_TIMEOUT), that it sets for commands that
have not received a completion at the time of entering EH.
Thus, libata doesn't really care about DID_TIME_OUT at all, and currently
clears the host byte at the end of EH, in ata_scsi_qc_complete(), before
scsi_eh_finish_cmd() is called.
However, this clearing in ata_scsi_qc_complete() is currently only done
for commands that are not ATA passthrough commands.
Since the host byte is visible in the completion that we return to user
space for ATA passthrough commands, for ATA passthrough commands that got
completed via EH (commands with sense data), the user will incorrectly see:
ATA pass-through(16): transport error: Host_status=0x03 [DID_TIME_OUT]
Fix this by moving the clearing of the host byte (which is currently only
done for commands that are not ATA passthrough commands) from
ata_scsi_qc_complete() to the start of EH (regardless if the command is
ATA passthrough or not).
While at it, use the proper helper function to clear the host byte, rather
than open coding the clearing.
This will make sure that we:
-Correctly clear DID_TIME_OUT for both ATA passthrough commands and
commands that are not ATA passthrough commands.
-Do not needlessly clear the host byte for commands that did not go via EH.
ata_scsi_qc_complete() is called both for commands that are completed
normally (without going via EH), and for commands that went via EH,
however, only commands that went via EH will have DID_TIME_OUT set.
Fixes: 24aeebbf8ea9 ("scsi: ata: libata: Change ata_eh_request_sense() to not set CHECK_CONDITION") Reported-by: Igor Pylypiv <ipylypiv@google.com> Closes: https://lore.kernel.org/linux-ide/ZttIN8He8TOZ7Lct@google.com/ Signed-off-by: Niklas Cassel <cassel@kernel.org> Tested-by: Igor Pylypiv <ipylypiv@google.com> Reviewed-by: Hannes Reinecke <hare@suse.de> Signed-off-by: Damien Le Moal <dlemoal@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The driver currently assumes that the first sequence number it will see
is going to be 0. This is not a realiable assumption and can break if,
for example, the tablet has already been running for some time prior to
the kernel driver connecting to the device. This commit initializes the
expected sequence number to -1 and will only print the "Dropped" warning
the it has been updated to a non-negative value.
Signed-off-by: Jason Gerecke <jason.gerecke@wacom.com> Tested-by: Joshua Dickens <joshua.dickens@wacom.com> Fixes: 6d09085b38e5 ("HID: wacom: Adding Support for new usages") Signed-off-by: Jiri Kosina <jkosina@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The current dropped packet reporting assumes that all sequence numbers
are 16 bits in length. This results in misleading "Dropped" messages if
the hardware uses fewer bits. For example, if a tablet uses only 8 bits
to store its sequence number, once it rolls over from 255 -> 0, the
driver will still be expecting a packet "256". This patch adjusts the
logic to reset the next expected packet to logical_minimum whenever
it overflows beyond logical_maximum.
Signed-off-by: Jason Gerecke <jason.gerecke@wacom.com> Tested-by: Joshua Dickens <joshua.dickens@wacom.com> Fixes: 6d09085b38e5 ("HID: wacom: Adding Support for new usages") Signed-off-by: Jiri Kosina <jkosina@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When running as a Xen PV dom0 the kernel is loaded by the hypervisor
using a different memory map than that of the host. In order to
minimize the required changes in the kernel, the kernel adapts its
memory map to that of the host. In order to do that it is checking
for conflicts of its load address with the host memory map.
Unfortunately the tested memory range does not include the .brk
area, which might result in crashes or memory corruption when this
area does conflict with the memory map of the host.
Fix the test by using the _end label instead of __bss_stop.
Fixes: 808fdb71936c ("xen: check for kernel memory conflicting with memory layout") Signed-off-by: Juergen Gross <jgross@suse.com> Tested-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> Reviewed-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Juergen Gross <jgross@suse.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
According to the display-drivers, 5nm DSI PLL (v4.2, v4.3) have
different boundaries for pll_clock_inverters programming. Follow the
vendor code and use correct values.
Fixes: 2f9ae4e395ed ("drm/msm/dsi: add support for DSI-PHY on SM8350 and SM8450") Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Reviewed-by: Abhinav Kumar <quic_abhinavk@quicinc.com>
Patchwork: https://patchwork.freedesktop.org/patch/606947/ Link: https://lore.kernel.org/r/20240804-sm8350-fixes-v1-3-1149dd8399fe@linaro.org Signed-off-by: Sasha Levin <sashal@kernel.org>
There is another cause for soft lock-up of GPU in empty ring-buffer:
race between GPU executing last commands and CPU checking ring for
emptiness. On GPU side IRQ for retire is triggered by CACHE_FLUSH_TS
event and RPTR shadow (which is used to check ring emptiness) is updated
a bit later from CP_CONTEXT_SWITCH_YIELD. Thus if GPU is executing its
last commands slow enough or we check that ring too fast we will miss a
chance to trigger switch to lower priority ring because current ring isn't
empty just yet. This can escalate to lock-up situation described in
previous patch.
To work-around this issue we keep track of last submit sequence number
for each ring and compare it with one written to memptrs from GPU during
execution of CACHE_FLUSH_TS event.
Fixes: b1fc2839d2f9 ("drm/msm: Implement preemption for A5XX targets") Signed-off-by: Vladimir Lypak <vladimir.lypak@gmail.com>
Patchwork: https://patchwork.freedesktop.org/patch/612047/ Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
On A5XX GPUs when preemption is used it's invietable to enter a soft
lock-up state in which GPU is stuck at empty ring-buffer doing nothing.
This appears as full UI lockup and not detected as GPU hang (because
it's not). This happens due to not triggering preemption when it was
needed. Sometimes this state can be recovered by some new submit but
generally it won't happen because applications are waiting for old
submits to retire.
One of the reasons why this happens is a race between a5xx_submit and
a5xx_preempt_trigger called from IRQ during submit retire. Former thread
updates ring->cur of previously empty and not current ring right after
latter checks it for emptiness. Then both threads can just exit because
for first one preempt_state wasn't NONE yet and for second one all rings
appeared to be empty.
To prevent such situations from happening we need to establish guarantee
for preempt_trigger to make decision after each submit or retire. To
implement this we serialize preemption initiation using spinlock. If
switch is already in progress we need to re-trigger preemption when it
finishes.
Fixes: b1fc2839d2f9 ("drm/msm: Implement preemption for A5XX targets") Signed-off-by: Vladimir Lypak <vladimir.lypak@gmail.com>
Patchwork: https://patchwork.freedesktop.org/patch/612045/ Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Two fields of preempt_record which are used by CP aren't reset on
resume: "data" and "info". This is the reason behind faults which happen
when we try to switch to the ring that was active last before suspend.
In addition those faults can't be recovered from because we use suspend
and resume to do so (keeping values of those fields again).
Fixes: b1fc2839d2f9 ("drm/msm: Implement preemption for A5XX targets") Signed-off-by: Vladimir Lypak <vladimir.lypak@gmail.com> Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/612043/ Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fine grain preemption (switching from/to points within submits)
requires extra handling in command stream of those submits, especially
when rendering with tiling (using GMEM). However this handling is
missing at this point in mesa (and always was). For this reason we get
random GPU faults and hangs if more than one priority level is used
because local preemption is enabled prior to executing command stream
from submit.
With that said it was ahead of time to enable local preemption by
default considering the fact that even on downstream kernel it is only
enabled if requested via UAPI.
Fixes: a7a4c19c36de ("drm/msm/a5xx: fix setting of the CP_PREEMPT_ENABLE_LOCAL register") Signed-off-by: Vladimir Lypak <vladimir.lypak@gmail.com>
Patchwork: https://patchwork.freedesktop.org/patch/612041/ Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In adreno_request_fw() when debugging information is printed to the log
after firmware load, an incorrect filename is printed. 'newname' is used
instead of 'fwname', so prefix "qcom/" is being added to filename.
Looks like "copy-paste" mistake.
Fix this mistake by replacing 'newname' with 'fwname'.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 2c41ef1b6f7d ("drm/msm/adreno: deal with linux-firmware fw paths") Signed-off-by: Aleksandr Mishin <amishin@t-argos.ru> Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/602382/ Signed-off-by: Rob Clark <robdclark@chromium.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
During merge of commit 4e991e3c16a3 ("powerpc: add CFUNC assembly
label annotation") a fallback version of CFUNC macro was added at
the last minute, so it can be used inconditionally.
Since commit 9132a2e82adc ("powerpc/8xx: Define a MODULE area below
kernel text"), module exec space is below PAGE_OFFSET so not only
space above PAGE_OFFSET, but space above TASK_SIZE need to be seen
as kernel space.
Until now the problem went undetected because by default TASK_SIZE
is 0x8000000 which means address space is determined by just
checking upper address bit. But when TASK_SIZE is over 0x80000000,
PAGE_OFFSET is used for comparison, leading to thinking module
addresses are part of user space.
Fix it by using TASK_SIZE instead of PAGE_OFFSET for address
comparison.
Commit cf209951fa7f ("powerpc/8xx: Map linear memory with huge pages")
introduced an initial mapping of kernel TEXT using PAGE_KERNEL_TEXT,
but the pages that contain kernel TEXT may also contain kernel RODATA,
and depending on selected debug options PAGE_KERNEL_TEXT may be either
RWX or ROX. RODATA must be writable during init because it also
contains ro_after_init data.
So use PAGE_KERNEL_X instead to be sure it is RWX.
In mtk_crtc_ddp_config(), mtk_crtc will use some configuration flags to
generate instructions to cmdq_handle, such as:
state->pending_config
mtk_crtc->pending_planes
plane_state->pending.config
mtk_crtc->pending_async_planes
plane_state->pending.async_config
These configuration flags may be set to false when a GCE IRQ comes calling
ddp_cmdq_cb(). This may result in missing prepare instructions,
especially if mtk_crtc_update_config() with the flase need_vblank (no need
to wait for vblank) cases.
Therefore, the mtk_crtc->config_updating flag is set at the beginning of
mtk_crtc_update_config() to ensure that these configuration flags won't be
changed when the mtk_crtc_ddp_config() is preparing instructions.
But somehow the ddp_cmdq_cb() didn't use the mtk_crtc->config_updating
flag to prevent those pending config flags from being cleared.
To avoid missing the configuration when generating the config instruction,
the config_updating flag should be added into ddp_cmdq_cb() and be
protected with spin_lock.
In dbNextAG() , there is no check for the case where bmp->db_numag is
greater or same than MAXAG due to a polluted image, which causes an
out-of-bounds. Therefore, a bounds check should be added in dbMount().
And in dbNextAG(), a check for the case where agpref is greater than
bmp->db_numag should be added, so an out-of-bounds exception should be
prevented.
Additionally, a check for the case where agno is greater or same than
MAXAG should be added in diAlloc() to prevent out-of-bounds.
Reported-by: Jeongjun Park <aha310510@gmail.com> Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Jeongjun Park <aha310510@gmail.com> Signed-off-by: Dave Kleikamp <dave.kleikamp@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The kref_put() function will call nport->release if the refcount drops to
zero. The nport->release release function is _efc_nport_free() which frees
"nport". But then we dereference "nport" on the next line which is a use
after free. Re-order these lines to avoid the use after free.
Fixes: fcd427303eb9 ("scsi: elx: libefc: SLI and FC PORT state machine interfaces") Signed-off-by: Dan Carpenter <dan.carpenter@linaro.org> Link: https://lore.kernel.org/r/b666ab26-6581-4213-9a3d-32a9147f0399@stanley.mountain Reviewed-by: Daniel Wagner <dwagner@suse.de> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The commit 0f5251339eda ("drm/vc4: hdmi: Make sure the controller is
powered in detect") introduced the necessary power management handling
to avoid register access while controller is powered down.
Unfortunately it just print a warning if pm_runtime_resume_and_get()
fails and proceed anyway.
This could happen during suspend to idle. So we must assume it is unsafe
to access the HDMI register. So bail out properly.
Fixes: 0f5251339eda ("drm/vc4: hdmi: Make sure the controller is powered in detect") Signed-off-by: Stefan Wahren <wahrenst@gmx.net> Reviewed-by: Maíra Canal <mcanal@igalia.com> Acked-by: Maxime Ripard <mripard@kernel.org> Signed-off-by: Maíra Canal <mcanal@igalia.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240821214052.6800-3-wahrenst@gmx.net Signed-off-by: Sasha Levin <sashal@kernel.org>
If the bridge is attached with the DRM_BRIDGE_ATTACH_NO_CONNECTOR flag set,
this driver won't initialize a connector and hence display mode won't be
validated in drm_connector_helper_funcs::mode_valid(). So, move the mode
validation from drm_connector_helper_funcs::mode_valid() to
drm_bridge_funcs::mode_valid(), because the mode validation is always done
for the bridge.
Several cs track offsets (such as 'track->db_s_read_offset')
either are initialized with or plainly take big enough values that,
once shifted 8 bits left, may be hit with integer overflow if the
resulting values end up going over u32 limit.
Same goes for a few instances of 'surf.layer_size * mslice'
multiplications that are added to 'offset' variable - they may
potentially overflow as well and need to be validated properly.
While some debug prints in this code section take possible overflow
issues into account, simply casting to (unsigned long) may be
erroneous in its own way, as depending on CPU architecture one is
liable to get different results.
Fix said problems by:
- casting 'offset' to fixed u64 data type instead of
ambiguous unsigned long.
- casting one of the operands in vulnerable to integer
overflow cases to u64.
- adjust format specifiers in debug prints to properly
represent 'offset' values.
Found by Linux Verification Center (linuxtesting.org) with static
analysis tool SVACE.
Fixes: 285484e2d55e ("drm/radeon: add support for evergreen/ni tiling informations v11") Signed-off-by: Nikita Zhandarovich <n.zhandarovich@fintech.ru> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The struct assertion is failed because sparse cannot parse
`#pragma pack(push, 1)` and `#pragma pack(pop)` correctly.
GCC's output is still 1-byte-aligned. No harm to memory layout.
The error can be filtered out by sparse-diff, but sometimes
multiple lines queezed into one, making the sparse-diff thinks
its a new error. I'm trying to aviod this by fixing errors.
It's not an error for a target to change the bus phase during a transfer.
Unfortunately, the FLAG_DMA_FIXUP workaround does not allow for that -- a
phase change produces a DRQ timeout error and the device borken flag will
be set.
Check the phase match bit during FLAG_DMA_FIXUP processing. Don't forget to
decrement the command residual. While we are here, change shost_printk()
into scmd_printk() for better consistency with other DMA error messages.
Correct a rare multipath failure issue by reverting commit 94a68c814328
("scsi: smartpqi: Quickly propagate path failures to SCSI midlayer") [1].
Reason for revert: The patch propagated the path failure to SML quickly
when one of the path fails during IO and AIO path gets disabled for a
multipath device.
But it created a new issue: when creating a volume on an encryption-enabled
controller, the firmware reports the AIO path is disabled, which cause the
driver to report a path failure to SML for a multipath device.
There will be a new fix to handle "Illegal request" and "Invalid field in
parameter list" on RAID path when the AIO path is disabled on a multipath
device.
The comment in the vbios structure says:
// = 128 means EDID length is 128 bytes, otherwise the EDID length = ucFakeEDIDLength*128
This fake edid struct has not been used in a long time, so I'm
not sure if there were actually any boards out there with a non-128 byte
EDID, but align the code with the comment.
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net> Reported-by: Thomas Weißschuh <linux@weissschuh.net> Link: https://lists.freedesktop.org/archives/amd-gfx/2024-June/109964.html Fixes: c324acd5032f ("drm/radeon/kms: parse the extended LCD info block") Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The comment in the vbios structure says:
// = 128 means EDID length is 128 bytes, otherwise the EDID length = ucFakeEDIDLength*128
This fake edid struct has not been used in a long time, so I'm
not sure if there were actually any boards out there with a non-128 byte
EDID, but align the code with the comment.
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net> Reported-by: Thomas Weißschuh <linux@weissschuh.net> Link: https://lists.freedesktop.org/archives/amd-gfx/2024-June/109964.html Fixes: d38ceaf99ed0 ("drm/amdgpu: add core driver (v4)") Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This commit adds a null check for the set_output_gamma function pointer
in the dcn30_set_output_transfer_func function. Previously,
set_output_gamma was being checked for nullity at line 386, but then it
was being dereferenced without any nullity check at line 401. This
could potentially lead to a null pointer dereference error if
set_output_gamma is indeed null.
To fix this, we now ensure that set_output_gamma is not null before
dereferencing it. We do this by adding a nullity check for
set_output_gamma before the call to set_output_gamma at line 401. If
set_output_gamma is null, we log an error message and do not call the
function.
This fix prevents a potential null pointer dereference error.
drivers/gpu/drm/amd/amdgpu/../display/dc/hwss/dcn30/dcn30_hwseq.c:401 dcn30_set_output_transfer_func()
error: we previously assumed 'mpc->funcs->set_output_gamma' could be null (see line 386)
drivers/gpu/drm/amd/amdgpu/../display/dc/hwss/dcn30/dcn30_hwseq.c
373 bool dcn30_set_output_transfer_func(struct dc *dc,
374 struct pipe_ctx *pipe_ctx,
375 const struct dc_stream_state *stream)
376 {
377 int mpcc_id = pipe_ctx->plane_res.hubp->inst;
378 struct mpc *mpc = pipe_ctx->stream_res.opp->ctx->dc->res_pool->mpc;
379 const struct pwl_params *params = NULL;
380 bool ret = false;
381
382 /* program OGAM or 3DLUT only for the top pipe*/
383 if (pipe_ctx->top_pipe == NULL) {
384 /*program rmu shaper and 3dlut in MPC*/
385 ret = dcn30_set_mpc_shaper_3dlut(pipe_ctx, stream);
386 if (ret == false && mpc->funcs->set_output_gamma) {
^^^^^^^^^^^^^^^^^^^^^^^^^^^^ If this is NULL
387 if (stream->out_transfer_func.type == TF_TYPE_HWPWL)
388 params = &stream->out_transfer_func.pwl;
389 else if (pipe_ctx->stream->out_transfer_func.type ==
390 TF_TYPE_DISTRIBUTED_POINTS &&
391 cm3_helper_translate_curve_to_hw_format(
392 &stream->out_transfer_func,
393 &mpc->blender_params, false))
394 params = &mpc->blender_params;
395 /* there are no ROM LUTs in OUTGAM */
396 if (stream->out_transfer_func.type == TF_TYPE_PREDEFINED)
397 BREAK_TO_DEBUGGER();
398 }
399 }
400
--> 401 mpc->funcs->set_output_gamma(mpc, mpcc_id, params);
^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Then it will crash
402 return ret;
403 }
Fixes: d99f13878d6f ("drm/amd/display: Add DCN3 HWSEQ") Reported-by: Dan Carpenter <dan.carpenter@linaro.org> Cc: Tom Chung <chiahsuan.chung@amd.com> Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Cc: Roman Li <roman.li@amd.com> Cc: Hersen Wu <hersenxs.wu@amd.com> Cc: Alex Hung <alex.hung@amd.com> Cc: Aurabindo Pillai <aurabindo.pillai@amd.com> Cc: Harry Wentland <harry.wentland@amd.com> Cc: Hamza Mahfooz <hamza.mahfooz@amd.com> Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com> Reviewed-by: Tom Chung <chiahsuan.chung@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The inter-column space in the debug summary is two spaces. However, in
one case, the extra space is handled implicitly in a field width
specifier. Make inter-column space explicit to ease future maintenance.
The Qualcomm SDM630 / SDM660 platform requires the same kind of
workaround as MSM8998: some IOMMUs have context banks reserved by
firmware / TZ, touching those banks resets the board.
Apply the num_context_bank workaround to those two SMMU devices in order
to allow them to be used by Linux.
SDM845's Adreno SMMU is unique in that it actually advertizes support
for 16K (and 32M) pages, which doesn't hold for newer SoCs.
This however, seems either broken in the hardware implementation, the
hypervisor middleware that abstracts the SMMU, or there's a bug in the
Linux kernel somewhere down the line that nobody managed to track down.
Booting SDM845 with 16K page sizes and drm/msm results in:
On qcom msm8998, writing to the last context bank of lpass_q6_smmu
(base address 0x05100000) produces a system freeze & reboot.
The hardware/hypervisor reports 13 context banks for the LPASS SMMU
on msm8998, but only the first 12 are accessible...
Override the number of context banks
Reviewing a series converting the for_each_chil_of_node() loops into
their _scoped variants made me realize there was no cleanup of the
already registered NAND devices upon error which may leak memory on
systems with more than a chip when this error occurs. We should call the
_nand_chips_cleanup() function when this happens.
There are some un-freed resources in one of the error path which would
benefit from a helper going through all the registered mtk chips one by
one and perform all the necessary cleanup. This is precisely what the
remove path does, so let's extract the logic in a helper.
After a CPU is marked offline and until it reaches its final trip to
idle, rcuo has several opportunities to be woken up, either because
a callback has been queued in the meantime or because
rcutree_report_cpu_dead() has issued the final deferred NOCB wake up.
If RCU-boosting is enabled, RCU kthreads are set to SCHED_FIFO policy.
And if RT-bandwidth is enabled, the related hrtimer might be armed.
However this then happens after hrtimers have been migrated at the
CPUHP_AP_HRTIMERS_DYING stage, which is broken as reported by the
following warning:
Fix this with waking up rcuo using an IPI if necessary. Since the
existing API to deal with this situation only handles swait queue, rcuo
is only woken up from offline CPUs if it's not already waiting on a
grace period. In the worst case some callbacks will just wait for a
grace period to complete before being assigned to a subsequent one.
Reported-by: "Cheng-Jui Wang (王正睿)" <Cheng-Jui.Wang@mediatek.com> Fixes: 5c0930ccaad5 ("hrtimers: Push pending hrtimers away from outgoing CPU earlier") Signed-off-by: Frederic Weisbecker <frederic@kernel.org> Signed-off-by: Neeraj Upadhyay <neeraj.upadhyay@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
devm_kasprintf() can return a NULL pointer on failure but this
returned value is not checked.
Fixes: acfe63ec1c59 ("mtd: Convert to using %pOFn instead of device_node.name") Signed-off-by: Charles Han <hanchunchao@inspur.com> Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Link: https://lore.kernel.org/linux-mtd/20240828092427.128177-1-hanchunchao@inspur.com Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 223a3b82834f ("power: supply: max17042_battery: use VFSOC for
capacity when no rsns") made it so that capacity on systems without
current sensing would be read from VFSOC instead of RepSOC. However,
the SOC threshold calculation still read RepSOC to get the SOC
regardless of the current sensing option state.
Fix this by applying the same conditional to determine which register
should be read.
This also seems to be the intended behavior as per the datasheet - SOC
alert config value in MiscCFG on setups without current sensing is set
to a value of 0b11, indicating SOC alerts being generated based on
VFSOC, instead of 0b00 which indicates SOC alerts being generated based
on RepSOC.
This fixes an issue on the Galaxy S3/Midas boards, where the alert
interrupt would be constantly retriggered, causing high CPU usage
on idle (around ~12%-15%).
Fixes: e5f3872d2044 ("max17042: Add support for signalling change in SOC") Signed-off-by: Artur Weber <aweber.kernel@gmail.com> Reviewed-by: Henrik Grimler <henrik@grimler.se> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Link: https://lore.kernel.org/r/20240817-max17042-soc-threshold-fix-v1-1-72b45899c3cc@gmail.com Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The POWER_SUPPLY_PROP_VOLTAGE_MIN_DESIGN and
POWER_SUPPLY_PROP_VOLTAGE_MAX_DESIGN values should be immutable
properties of the battery, but for this driver they are writable values
and used as the minimum and maximum values for charging. Remove the
DESIGN designation from these values.
Fixes: 46c202b5f25f ("power: supply: add battery driver for AXP20X and AXP22X PMICs") Suggested-by: Chen-Yu Tsai <wens@kernel.org> Signed-off-by: Chris Morgan <macromorgan@hotmail.com> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Link: https://lore.kernel.org/r/20240821215456.962564-3-macroalpha82@gmail.com Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
GCC 12.3.0 compiler on linux-next next-20240709 tree found the execution
path in which, due to lazy evaluation, devlength isn't initialised with the
parsed string:
Chips reporting overcurrent alarms report it in the second alarm register.
That means the second alarm register has to be read, even if the chip only
supports 8 or fewer ADC channels.
MAX16067 and MAX16068 report undervoltage and overvoltage alarms in
separate registers. Fold register contents together to report both with
the existing alarm attribute. This requires actually storing the chip type
in struct max16065_data. Rename the variable 'chip' to match the variable
name used in the probe function.
Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Fixes: f5bae2642e3d ("hwmon: Driver for MAX16065 System Manager and compatibles") Signed-off-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The function i2c_match_id() is used to fetch the matching ID from
the i2c_device_id table. This is often used to then retrieve the
matching driver_data. This can be done in one step with the helper
i2c_get_match_data().
This helper has a couple other benefits:
* It doesn't need the i2c_device_id passed in so we do not need
to have that forward declared, allowing us to remove those or
move the i2c_device_id table down to its more natural spot
with the other module info.
* It also checks for device match data, which allows for OF and
ACPI based probing. That means we do not have to manually check
those first and can remove those checks.
In function loongson_card_parse_of(), when get device_node
'codec' failed, the function of_node_put(codec) should not
be invoked, thus fix error release.
Stan Johnson recently reported a failure from the 'dump' command:
DUMP: Date of this level 0 dump: Fri Aug 9 23:37:15 2024
DUMP: Dumping /dev/sda (an unlisted file system) to /dev/null
DUMP: Label: none
DUMP: Writing 10 Kilobyte records
DUMP: mapping (Pass I) [regular files]
DUMP: mapping (Pass II) [directories]
DUMP: estimated 3595695 blocks.
DUMP: Context save fork fails in parent 671
The dump program uses the clone syscall with the CLONE_IO flag, that is,
flags == 0x80000000. When that value is promoted from long int to u64 by
m68k_clone(), it undergoes sign-extension. The new value includes
CLONE_INTO_CGROUP so the validation in cgroup_css_set_fork() fails and
the syscall returns -EBADF. Avoid sign-extension by casting to u32.
Reported-by: Stan Johnson <userm57@yahoo.com> Closes: https://lists.debian.org/debian-68k/2024/08/msg00000.html Fixes: 6aabc1facdb2 ("m68k: Implement copy_thread_tls()") Signed-off-by: Finn Thain <fthain@linux-m68k.org> Reviewed-by: Geert Uytterhoeven <geert@linux-m68k.org> Link: https://lore.kernel.org/3463f1e5d4e95468dc9f3368f2b78ffa7b72199b.1723335149.git.fthain@linux-m68k.org Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
kprobe_args_{char,string}.tc are using available_filter_functions file
which is provided by function tracer. Thus if function tracer is disabled,
these tests are failed on recent kernels because tracefs_create_dir is
not raised events by adding a dynamic event.
Add available_filter_functions to requires line.
The code is obtaining a GPIO reset using the reset GPIO
name "reset-gpios", but the gpiolib is already adding the
suffix "-gpios" to anything passed to this function and
will be looking for "reset-gpios-gpios" which is most
certainly not what the author desired.
The tas2781-i2c driver gets an IRQ from either ACPI or device tree,
then proceeds to check if the IRQ has a corresponding GPIO and in
case it does enforce the GPIO as input and set a label on it.
This is abuse of the API:
- First we cannot guarantee that the numberspaces of the GPIOs and
the IRQs are the same, i.e that an IRQ number corresponds to
a GPIO number like that.
- Second, GPIO chips and IRQ chips should be treated as orthogonal
APIs, the irqchip needs to ascertain that the backing GPIO line
is set to input etc just using the irqchip.
- Third it is using the legacy <linux/gpio.h> API which should not
be used in new code yet this was added just a year ago.
Delete the offending code.
If this creates problems the GPIO and irqchip maintainers can help
to fix the issues.
It *should* not create any problems, because the irq isn't
used anywhere in the driver, it's just obtained and then
left unused.
Return devm_of_clk_add_hw_provider() in order to transfer the error, if it
fails due to resource allocation failure or device tree clock provider
registration failure.
Fixes: bdd229ab26be ("ASoC: rt5682s: Add driver for ALC5682I-VS codec") Signed-off-by: Ma Ke <make24@iscas.ac.cn> Link: https://patch.msgid.link/20240717115436.3449492-1-make24@iscas.ac.cn Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
LAM can only be enabled when a process is single-threaded. But _kernel_
threads can temporarily use a single-threaded process's mm.
If LAM is enabled by a userspace process while a kthread is using its
mm, the kthread will not observe LAM enablement (i.e. LAM will be
disabled in CR3). This could be fine for the kthread itself, as LAM only
affects userspace addresses. However, if the kthread context switches to
a thread in the same userspace process, CR3 may or may not be updated
because the mm_struct doesn't change (based on pending TLB flushes). If
CR3 is not updated, the userspace thread will run incorrectly with LAM
disabled, which may cause page faults when using tagged addresses.
Example scenario:
CPU 1 CPU 2
/* kthread */
kthread_use_mm()
/* user thread */
prctl_enable_tagged_addr()
/* LAM enabled on CPU 2 */
/* LAM disabled on CPU 1 */
context_switch() /* to CPU 1 */
/* Switching to user thread */
switch_mm_irqs_off()
/* CR3 not updated */
/* LAM is still disabled on CPU 1 */
Synchronize LAM enablement by sending an IPI to all CPUs running with
the mm_struct to enable LAM. This makes sure LAM is enabled on CPU 1
in the above scenario before prctl_enable_tagged_addr() returns and
userspace starts using tagged addresses, and before it's possible to
run the userspace process on CPU 1.
In switch_mm_irqs_off(), move reading the LAM mask until after
mm_cpumask() is updated. This ensures that if an outdated LAM mask is
written to CR3, an IPI is received to update it right after IRQs are
re-enabled.
[ dhansen: Add a LAM enabling helper and comment it ]
Fixes: 82721d8b25d7 ("x86/mm: Handle LAM on context switch") Suggested-by: Andy Lutomirski <luto@kernel.org> Signed-off-by: Yosry Ahmed <yosryahmed@google.com> Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com> Reviewed-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Link: https://lore.kernel.org/all/20240702132139.3332013-2-yosryahmed%40google.com Signed-off-by: Sasha Levin <sashal@kernel.org>
Driver is leaking OF node reference on memory allocation failure.
Acquire the OF node reference after memory allocation to fix this and
keep it simple.
Fixes: 5a2308da9f60 ("riscv: Add Canaan Kendryte K210 reset controller") Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Damien Le Moal <dlemoal@kernel.org> Link: https://lore.kernel.org/r/20240825-reset-cleanup-scoped-v1-2-03f6d834f8c0@linaro.org Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
Driver is leaking OF node reference on memory allocation failure.
Acquire the OF node reference after memory allocation to fix this and
keep it simple.
Fixes: aed6f3cadc86 ("reset: berlin: convert to a platform driver") Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Reviewed-by: Damien Le Moal <dlemoal@kernel.org> Link: https://lore.kernel.org/r/20240825-reset-cleanup-scoped-v1-1-03f6d834f8c0@linaro.org Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
There is no "fsl,phy" property in pin controller pincfg nodes:
imx7d-zii-rmu2.dtb: pinctrl@302c0000: enet1phyinterruptgrp: 'fsl,pins' is a required property
imx7d-zii-rmu2.dtb: pinctrl@302c0000: enet1phyinterruptgrp: 'fsl,phy' does not match any of the regexes: 'pinctrl-[0-9]+'
Fixes: f496e6750083 ("ARM: dts: Add ZII support for ZII i.MX7 RMU2 board") Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by: Shawn Guo <shawnguo@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
According to datasheet, Chapter 34. Clock Generator, section 34.2,
Embedded characteristics, source clock for RTT is the TD_SLCK, registered
with ID 1 by the slow clock controller driver. Fix RTT clock.
The pm_runtime_disable() is missing in remove function, use
devm_pm_runtime_enable() to fix it. So the pm_runtime_disable() in
the probe error path can also be removed.
Fixes: a38a2233f23b ("spi: bcmbca-hsspi: Add driver for newer HSSPI controller") Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com> Reviewed-by: William Zhang <william.zhang@broadcom.com> Link: https://patch.msgid.link/20240826124903.3429235-2-ruanjinjie@huawei.com Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The DMA carveout for the C6x core 0 is at 0xa6000000 and core 1 is at
0xa7000000. These are reversed in DT. While both C6x can access either
region, so this is not normally a problem, but if we start restricting
the memory each core can access (such as with firewalls) the cores
accessing the regions for the wrong core will not work. Fix this here.
The DMA carveout for the C6x core 0 is at 0xa6000000 and core 1 is at
0xa7000000. These are reversed in DT. While both C6x can access either
region, so this is not normally a problem, but if we start restricting
the memory each core can access (such as with firewalls) the cores
accessing the regions for the wrong core will not work. Fix this here.
The vendor prefix for Hardkernel ODROID-M1 is incorrectly listed as
rockchip. Use the proper hardkernel vendor prefix for this board, while
at it also drop the redundant soc prefix.
The RTC and RTT peripherals use the timing domain slow clock (TD_SLCK),
sourced from the 32.768 kHz crystal oscillator or slow rc oscillator.
The previously used Monitoring domain slow clock (MD_SLCK) is sourced
from an internal RC oscillator which is most probably not precise enough
for real time clock purposes.
Fixes: 1e5f532c2737 ("ARM: dts: at91: sam9x60: add device tree for soc and board") Fixes: 5f6b33f46346 ("ARM: dts: sam9x60: add rtt") Signed-off-by: Alexander Dahl <ada@thorsis.com> Link: https://lore.kernel.org/r/20240821055136.6858-1-ada@thorsis.com
[claudiu.beznea: removed () around the last commit description paragraph,
removed " in front of "timing domain slow clock", described that
TD_SLCK can also be sourced from slow rc oscillator] Signed-off-by: Claudiu Beznea <claudiu.beznea@tuxon.dev> Signed-off-by: Sasha Levin <sashal@kernel.org>
If regulator_get() in of_regulator_bulk_get_all() returns an error, that
error gets overridden and -EINVAL is always passed out. This masks probe
deferral requests and likely won't work properly in all cases.
Fix this by letting of_regulator_bulk_get_all() return the original
error code.
Fixes: 27b9ecc7a9ba ("regulator: Add of_regulator_bulk_get_all") Signed-off-by: Chen-Yu Tsai <wenst@chromium.org> Link: https://patch.msgid.link/20240822072047.3097740-3-wenst@chromium.org Signed-off-by: Mark Brown <broonie@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The speedbin eFuse reads a value 'x' from 0 to 7 and, in order to
make that compatible with opp-supported-hw, it gets post processed
as BIT(x).
Change all of the 0x30 supported-hw to 0x20 to avoid getting
duplicate OPPs for speedbin 4, and also change all of the 0x8 to
0xcf because speedbins different from 4 and 5 do support 900MHz,
950MHz, 1000MHz with the higher voltage of 850mV, 900mV, 950mV
respectively.
Zero and negative number is not a valid IRQ for in-kernel code and the
irq_of_parse_and_map() function returns zero on error. So this check for
valid IRQs should only accept values > 0.
The blk_add_partition() function initially used a single if-condition
(IS_ERR(part)) to check for errors when adding a partition. This was
modified to handle the specific case of -ENXIO separately, allowing the
function to proceed without logging the error in this case. However,
this change unintentionally left a path where md_autodetect_dev()
could be called without confirming that part is a valid pointer.
This commit separates the error handling logic by splitting the
initial if-condition, improving code readability and handling specific
error scenarios explicitly. The function now distinguishes the general
error case from -ENXIO without altering the existing behavior of
md_autodetect_dev() calls.
Fixes: b72053072c0b (block: allow partitions on host aware zone devices) Signed-off-by: Riyan Dhiman <riyandhiman14@gmail.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20240911132954.5874-1-riyandhiman14@gmail.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
Utilize the %pe print specifier to get the symbolic error name as a
string (i.e "-ENOMEM") in the log message instead of the error code to
increase its readablility.
This change was suggested in
https://lore.kernel.org/all/92972476-0b1f-4d0a-9951-af3fc8bc6e65@suswa.mountain/
The io worker threads are userland threads that just never exit to the
userland. By that, they are also assigned to a cgroup (the group of the
creating task).
When creating a new io worker, this worker should inherit the cpuset
of the cgroup.
The io worker threads are userland threads that just never exit to the
userland. By that, they are also assigned to a cgroup (the group of the
creating task).
When changing the affinity of the io_wq thread via syscall, we must only
allow cpumasks within the limits defined by the cpuset controller of the
cgroup (if enabled).
Process 1 Process 2 Process 3 Process 4
(BIC1) (BIC2) (BIC3) (BIC4)
Λ | | |
\--------------\ \-------------\ \-------------\|
V V V
bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
ref 0 1 2 4
After commit 0e456dba86c7 ("block, bfq: choose the last bfqq from merge
chain in bfq_setup_cooperator()"), if P1 issues a new IO:
Without the patch:
Process 1 Process 2 Process 3 Process 4
(BIC1) (BIC2) (BIC3) (BIC4)
Λ | | |
\------------------------------\ \-------------\|
V V
bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
ref 0 0 2 4
bfqq3 will be used to handle IO from P1, this is not expected, IO
should be redirected to bfqq4;
With the patch:
-------------------------------------------
| |
Process 1 Process 2 Process 3 | Process 4
(BIC1) (BIC2) (BIC3) | (BIC4)
| | | |
\-------------\ \-------------\|
V V
bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
ref 0 0 2 4
IO is redirected to bfqq4, however, procress reference of bfqq3 is still
2, while there is only P2 using it.
Fix the problem by calling bfq_merge_bfqqs() for each bfqq in the merge
chain. Also change bfqq_merge_bfqqs() to return new_bfqq to simplify
code.
Fixes: 0e456dba86c7 ("block, bfq: choose the last bfqq from merge chain in bfq_setup_cooperator()") Signed-off-by: Yu Kuai <yukuai3@huawei.com> Link: https://lore.kernel.org/r/20240909134154.954924-3-yukuai1@huaweicloud.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
After commit 42c306ed7233 ("block, bfq: don't break merge chain in
bfq_split_bfqq()"), if the current procress is the last holder of bfqq,
the bfqq can be freed after bfq_split_bfqq(). Hence recored the bfqq and
then access bfqq->waker_bfqq may trigger UAF. What's more, the waker_bfqq
may in the merge chain of bfqq, hence just recored waker_bfqq is still
not safe.
Fix the problem by adding a helper bfq_waker_bfqq() to check if
bfqq->waker_bfqq is in the merge chain, and current procress is the only
holder.
Fast symlink can be used if the on-disk symlink data is stored
in the same block as the on-disk inode, so we don’t need to trigger
another I/O for symlink data. However, currently fs correction could be
reported _incorrectly_ if inode xattrs are too large.
In fact, these should be valid images although they cannot be handled as
fast symlinks.
Process 1 Process 2 Process 3 Process 4
(BIC1) (BIC2) (BIC3) (BIC4)
Λ | | |
\-------------\ \-------------\ \--------------\|
V V V
bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
ref 0 1 2 4
If Process 1 issue a new IO and bfqq2 is found, and then bfq_init_rq()
decide to spilt bfqq2 by bfq_split_bfqq(). Howerver, procress reference
of bfqq2 is 1 and bfq_split_bfqq() just clear the coop flag, which will
break the merge chain.
Expected result: caller will allocate a new bfqq for BIC1
Process 1 Process 2 Process 3 Process 4
(BIC1) (BIC2) (BIC3) (BIC4)
| | |
\-------------\ \--------------\|
V V
bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
ref 0 0 1 3
Since the condition is only used for the last bfqq4 when the previous
bfqq2 and bfqq3 are already splited. Fix the problem by checking if
bfqq is the last one in the merge chain as well.
Process 1 Process 2 Process 3 Process 4
(BIC1) (BIC2) (BIC3) (BIC4)
Λ | | |
\--------------\ \-------------\ \-------------\|
V V V
bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
IO from Process 1 will get bfqf2 from BIC1 first, then
bfq_setup_cooperator() will found bfqq2 already merged to bfqq3 and then
handle this IO from bfqq3. However, the merge chain can be much deeper
and bfqq3 can be merged to other bfqq as well.
Fix this problem by iterating to the last bfqq in
bfq_setup_cooperator().
Process 1 Process 2 Process 3
(BIC1) (BIC2) (BIC3)
| Λ | Λ | Λ
| | | | | |
V | V | V |
bfqq1 bfqq2 bfqq3
process ref: 1 1 1
2) bfqq1 merged to bfqq2:
Process 1 Process 2 Process 3
(BIC1) (BIC2) (BIC3)
| | | Λ
\--------------\| | |
V V |
bfqq1--------->bfqq2 bfqq3
process ref: 0 2 1
3) bfqq2 merged to bfqq3:
Process 1 Process 2 Process 3
(BIC1) (BIC2) (BIC3)
here -> Λ | |
\--------------\ \-------------\|
V V
bfqq1--------->bfqq2---------->bfqq3
process ref: 0 1 3
In this case, IO from Process 1 will get bfqq2 from BIC1 first, and then
get bfqq3 through merge chain, and finially handle IO by bfqq3.
Howerver, current code will think bfqq2 is owned by BIC1, like initial
state, and set bfqq2->bic to BIC1.
bfq_insert_request
-> by Process 1
bfqq = bfq_init_rq(rq)
bfqq = bfq_get_bfqq_handle_split
bfqq = bic_to_bfqq
-> get bfqq2 from BIC1
bfqq->ref++
rq->elv.priv[0] = bic
rq->elv.priv[1] = bfqq
if (bfqq_process_refs(bfqq) == 1)
bfqq->bic = bic
-> record BIC1 to bfqq2
__bfq_insert_request
new_bfqq = bfq_setup_cooperator
-> get bfqq3 from bfqq2->new_bfqq
bfqq_request_freed(bfqq)
new_bfqq->ref++
rq->elv.priv[1] = new_bfqq
-> handle IO by bfqq3
Fix the problem by checking bfqq is from merge chain fist. And this
might fix a following problem reported by our syzkaller(unreproducible):
==================================================================
BUG: KASAN: slab-use-after-free in bfq_do_early_stable_merge block/bfq-iosched.c:5692 [inline]
BUG: KASAN: slab-use-after-free in bfq_do_or_sched_stable_merge block/bfq-iosched.c:5805 [inline]
BUG: KASAN: slab-use-after-free in bfq_get_queue+0x25b0/0x2610 block/bfq-iosched.c:5889
Write of size 1 at addr ffff888123839eb8 by task kworker/0:1H/18595
Second to last potentially related work creation:
kasan_save_stack+0x20/0x40 mm/kasan/common.c:45
__kasan_record_aux_stack+0xaf/0xc0 mm/kasan/generic.c:492
__call_rcu_common kernel/rcu/tree.c:2712 [inline]
call_rcu+0xce/0x1020 kernel/rcu/tree.c:2826
ioc_destroy_icq+0x54c/0x830 block/blk-ioc.c:105
ioc_release_fn+0xf0/0x360 block/blk-ioc.c:124
process_one_work kernel/workqueue.c:2627 [inline]
process_scheduled_works+0x432/0x13f0 kernel/workqueue.c:2700
worker_thread+0x6f2/0x1160 kernel/workqueue.c:2781
kthread+0x33c/0x440 kernel/kthread.c:388
ret_from_fork+0x4d/0x80 arch/x86/kernel/process.c:147
ret_from_fork_asm+0x1b/0x30 arch/x86/entry/entry_64.S:305
The buggy address belongs to the object at ffff888123839d68
which belongs to the cache bfq_io_cq of size 1360
The buggy address is located 336 bytes inside of
freed 1360-byte region [ffff888123839d68, ffff88812383a2b8)
If request timetout is handled by nbd_requeue_cmd(), normal completion
has to be stopped for avoiding to complete this requeued request, other
use-after-free can be triggered.
Fix the race by clearing NBD_CMD_INFLIGHT in nbd_requeue_cmd(), meantime
make sure that cmd->lock is grabbed for clearing the flag and the
requeue.
Cc: Josef Bacik <josef@toxicpanda.com> Cc: Yu Kuai <yukuai3@huawei.com> Fixes: 2895f1831e91 ("nbd: don't clear 'NBD_CMD_INFLIGHT' flag if request is not completed") Signed-off-by: Ming Lei <ming.lei@redhat.com> Reviewed-by: Yu Kuai <yukuai3@huawei.com> Link: https://lore.kernel.org/r/20240830034145.1827742-1-ming.lei@redhat.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Sasha Levin <sashal@kernel.org>
projects / thirdparty / kernel / stable.git / log
