In eb_lookup_vma(), the code checks that the context vm matches before
incrementing the i915 vma usage count, but for the non-matching case it
didn't clear the non-matching vma pointer, so it would then mistakenly
be returned, causing potential UaF and refcount issues.
Reported-by: Yassine Mounir <sosohero200@gmail.com>
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In Hamoa, SMMU invalidation requires the GCC_AGGRE_USB_NOC_AXI_CLK
to be on for the USB QTB to be functional. This is currently
explicitly enabled by the DWC3 glue driver, so an invalidation
happening while the USB controller is suspended will fault.
Solve this by voting for the GCC MMU USB QTB clock.
Fixes: 161b7c401f ("clk: qcom: Add Global Clock controller (GCC) driver for X1E80100")
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Jagadeesh Kona <jagadeesh.kona@oss.qualcomm.com>
Reviewed-by: Taniya Das <taniya.das@oss.qualcomm.com>
Reviewed-by: Abel Vesa <abel.vesa@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260327-hamoa-usb-qtb-clk-always-on-v2-1-7d8a406e650f@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
The static array 'xxx_critical_cbcrs' contains probe match-like data and
is not modified: neither by the driver defining it nor by common.c code
using it.
Make it const for code safety and code readability.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260331091721.61613-4-krzysztof.kozlowski@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
The static 'struct qcom_cc_driver_data' contains probe match-like data
and is not modified: neither by the driver defining it nor by common.c
code using it.
Make it const for code safety and code readability.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260331091721.61613-3-krzysztof.kozlowski@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Static 'struct qcom_cc_desc' is not modified by drivers and can be made
const for code safety.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Taniya Das <taniya.das@oss.qualcomm.com>
Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260331085521.37337-2-krzysztof.kozlowski@oss.qualcomm.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
This version includes the following changes:
- Setting current base frequency as maximum for SST-BF with
kernel QOS changes
- Harmonize extended family decoded with the rest of the kernel
- Minor changes for error codes and messages
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
When running intel-speed-select on unsupported CLX platforms, it prints
intel-speed-select: Invalid CPU model (85)
: Success
Because this is not a system error and errno is not set.
Replace err() with exit().
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
When running an old version intel-speed-select tool on newer platforms,
even with "intel-speed-select -v", the tool only complains about
"Incompatible API version", without giving the current version info.
Print Version info whenever Incompatible API version is detected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
When running the "intel-speed-select -h" command, it returns
1. 0 when using a version that is API incompatible.
2. 1 when using a version that is API compatible.
And this is confusing.
Fix the program to return 0 for "-h" parameter, and return 1 whenever
"Incompatible API versions" is detected.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
When decode and use CPU extended family ID in intel-speed-select, there
are several potential issues,
1. Mask with 0x0f to get CPU extended family ID is bogus because
CPU extended family ID takes 8 bits (bit 27:20).
2. Use CPU extended family ID fields without checking CPU family ID is
risky. Because Intel SDM says, "The Extended Family ID needs to be
examined only when the Family ID is 0FH."
3. Saving cpu family ID and cpu extended family ID separately doesn't
align with Linux kernel. And it may bring extra complexity when
making family specific changes in the future.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
SST-PP level change results in online/offline of CPUs with -o option.
The Linux intel-pstate driver internally stores the current HWP_REQ MSR
value during offline and restores them during online.
It is possible that during SST-PP level change, the new HWP_CAP limits
can be updated. So, when a CPU is online, the HWP_REQ MSR should be
updated to new values based on HWP_CAP values.
This is particularly problematic when either turbo is disabled or the
current HWP_REQ value (stored before online) is less than the base
frequency from the updated HWP_CAP MSR guaranteed value. If the HWP_REQ
MSR is not updated, then the performance will be limited to the value
before perf level change.
Hence the tool updates cpufreq scaling_max_freq to the newer
base_frequency value in this case. This step is not required when HWP
interrupts are enabled, as the perf level change should result in a new
interrupt with HWP_GUARANTEED_PERF_CHANGE_STATUS and the intel_pstate
driver will update to new limits.
But the tool needs to handle the case when HWP interrupts are not
enabled but there is no way for the tool to know that HWP interrupts are
enabled or not. So, it has to still update the scaling_max_freq.
With the QOS changes in the kernel, user space writes to scaling_max_freq
are treated as hard limits. So, when base frequency is increased with
SST-BF enabled, the cpufreq subsystem will still not allow setting to the
SST-BF high priority core frequency. So, the HWP_REQ MSR will still be
capped to the user-set scaling_max_freq after SST-PP level change.
To address this, instead of setting scaling_max_freq to the current HWP_CAP
highest frequency, set it to the maximum integer value to set the QOS limit
as unconstrained. In this case, the actual HWP_REQ maximum frequency will
still be capped to HWP_CAP highest performance by the intel-pstate driver.
So, it will not result in invalid HWP_REQ values.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Fix allocation in TLB uniquification
Fix SiByte cache initialisation
Check uart parameters from firmware on Loongson64 systems
Fix clock id mismatch for Ralink SoCs
Fix GCC version check for __mutli3 workaround
-----BEGIN PGP SIGNATURE-----
iQJOBAABCAA4FiEEbt46xwy6kEcDOXoUeZbBVTGwZHAFAmnSLdQaHHRzYm9nZW5k
QGFscGhhLmZyYW5rZW4uZGUACgkQeZbBVTGwZHC6fA/7B13xK55Q70sZsiWK+02m
XhE2cEOctWpzpwoMGYov4lY4alW4IzL7ZVWqeD96QeEILopHf448w/3+KCbdjmKr
0+6AY9n1k91i0R2hhEWBor/B+4EXjUgKGrGTzQxyCKkRnzQcy0HBuoZuY7xoW2j/
432DmEs4ogTFHMpkwrz0jB/dv7286KY5VUOr4y2G3SXMgF9mrPqQsFTypR2IC7IS
TZ8rr+tjtsM2WYcayW6pbo5jBX+TfJbRxfdQELjX1YStGkThtIg1ZNUD+DK3CTUO
M9Id1eknSUYcyuxlJruVwEjRZcv49uPEE4JTPjs50dZQxSsWPp1B70FZN/KsP3D3
miJS2Shx0ptPVta/DiNB0sdLFdMvo2FeAQBzl3ZrAYXVuIT9/KrXY/ZiInLevMFa
+m7J85XvfW7vcLtHM0ptz3+Zk836nnTaqUgLkTTtCEBLH/YAjkY8/orZOHaFfOHJ
zXzLZhnhyosJ3Hd4Ie49YekIPjcKf9bMastGzD887syLOGtGqU6FpV0FSgJtER71
SXYmUfZfS7Aj4nfwQuijAJSs9aZnSXUEvxYOQvAM9Go1zLBIZXEVJNmYd8smfpF0
OduLVKDliirEmHPDMOLKmX26E6EZxeres4aHVAJ7xOsyoTwxFY/8PzInQFP4wV7X
/aVMB91RS16sJIphpHcJ6E8=
=1q4p
-----END PGP SIGNATURE-----
Merge tag 'mips-fixes_7.0_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux
Pull MIPS fixes from Thomas Bogendoerfer:
- Fix TLB uniquification for systems with TLB not initialised by
firmware
- Fix allocation in TLB uniquification
- Fix SiByte cache initialisation
- Check uart parameters from firmware on Loongson64 systems
- Fix clock id mismatch for Ralink SoCs
- Fix GCC version check for __mutli3 workaround
* tag 'mips-fixes_7.0_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
mips: mm: Allocate tlb_vpn array atomically
MIPS: mm: Rewrite TLB uniquification for the hidden bit feature
MIPS: mm: Suppress TLB uniquification on EHINV hardware
MIPS: Always record SEGBITS in cpu_data.vmbits
MIPS: Fix the GCC version check for `__multi3' workaround
MIPS: SiByte: Bring back cache initialisation
mips: ralink: update CPU clock index
MIPS: Loongson64: env: Check UARTs passed by LEFI cautiously
Here are a relativly large number of small char/misc/iio and other
driver fixes for 7.0-rc7. There's a bunch, but overall they are all
small fixes for issues that people have been having that I finally
caught up with getting merged due to delays on my end.
The "largest" change overall is just some documentation updates to the
security-bugs.rst file to hopefully tell the AI tools (and any users
that actually read the documentation), how to send us better security
bug reports as the quantity of reports these past few weeks has
increased dramatically due to tools getting better at "finding" things.
Included in here are:
- lots of small IIO driver fixes for issues reported in 7.0-rc
- gpib driver fixes
- comedi driver fixes
- interconnect driver fix
- nvmem driver fixes
- mei driver fix
- counter driver fix
- binder rust driver fixes
- some other small misc driver fixes
All of these have been in linux-next this week with no reported issues.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
iG0EABECAC0WIQT0tgzFv3jCIUoxPcsxR9QN2y37KQUCadIaEg8cZ3JlZ0Brcm9h
aC5jb20ACgkQMUfUDdst+ym34ACfTtx3DzFMxNmE2eQrvptvqmSeVEoAoIt9h9kF
pVDpO6Njqetq0zzWfBL3
=uv/h
-----END PGP SIGNATURE-----
Merge tag 'char-misc-7.0-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc
Pull char/misc/iio driver fixes from Greg KH:
"Here are a relativly large number of small char/misc/iio and other
driver fixes for 7.0-rc7. There's a bunch, but overall they are all
small fixes for issues that people have been having that I finally
caught up with getting merged due to delays on my end.
The "largest" change overall is just some documentation updates to the
security-bugs.rst file to hopefully tell the AI tools (and any users
that actually read the documentation), how to send us better security
bug reports as the quantity of reports these past few weeks has
increased dramatically due to tools getting better at "finding"
things.
Included in here are:
- lots of small IIO driver fixes for issues reported in 7.0-rc
- gpib driver fixes
- comedi driver fixes
- interconnect driver fix
- nvmem driver fixes
- mei driver fix
- counter driver fix
- binder rust driver fixes
- some other small misc driver fixes
All of these have been in linux-next this week with no reported issues"
* tag 'char-misc-7.0-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (63 commits)
Documentation: fix two typos in latest update to the security report howto
Documentation: clarify the mandatory and desirable info for security reports
Documentation: explain how to find maintainers addresses for security reports
Documentation: minor updates to the security contacts
.get_maintainer.ignore: add myself
nvmem: zynqmp_nvmem: Fix buffer size in DMA and memcpy
nvmem: imx: assign nvmem_cell_info::raw_len
misc: fastrpc: check qcom_scm_assign_mem() return in rpmsg_probe
misc: fastrpc: possible double-free of cctx->remote_heap
comedi: dt2815: add hardware detection to prevent crash
comedi: runflags cannot determine whether to reclaim chanlist
comedi: Reinit dev->spinlock between attachments to low-level drivers
comedi: me_daq: Fix potential overrun of firmware buffer
comedi: me4000: Fix potential overrun of firmware buffer
comedi: ni_atmio16d: Fix invalid clean-up after failed attach
gpib: fix use-after-free in IO ioctl handlers
gpib: lpvo_usb: fix memory leak on disconnect
gpib: Fix fluke driver s390 compile issue
lis3lv02d: Omit IRQF_ONESHOT if no threaded handler is provided
lis3lv02d: fix kernel-doc warnings
...
Here are 2 small tty vt fixes for 7.0-rc7 to resolve some reported
issues with the resize ability of the alt screen buffer. Both of these
have been in linux-next all week with no reported issues.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
iG0EABECAC0WIQT0tgzFv3jCIUoxPcsxR9QN2y37KQUCadIatA8cZ3JlZ0Brcm9h
aC5jb20ACgkQMUfUDdst+ylqVQCfecM0mxjbI8EmKsfL6FtpAstcW1IAn2XwT7S0
ITAhHAylAuh3HHpXqRfi
=eulV
-----END PGP SIGNATURE-----
Merge tag 'tty-7.0-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Pull tty fixes from Greg KH:
"Here are two small tty vt fixes for 7.0-rc7 to resolve some reported
issues with the resize ability of the alt screen buffer. Both of these
have been in linux-next all week with no reported issues"
* tag 'tty-7.0-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
vt: resize saved unicode buffer on alt screen exit after resize
vt: discard stale unicode buffer on alt screen exit after resize
Here are a bunch of USB and Thunderbolt fixes (most all are USB) for
7.0-rc7. More than I normally like this late in the release cycle,
partly due to my recent travels, and partly due to people banging away
on the USB gadget interfaces and apis more than normal (big shoutout to
Android for getting the vendors to actually work upstream on this,
that's a huge win overall for everyone here.)
Included in here are:
- Small thunderbolt fix
- new USB serial driver ids added
- typec driver fixes
- gadget driver fixes for some disconnect issues
- other usb gadget driver fixes for reported problems with binding and
unbinding devices as happens when a gadget device connects /
disconnects from a system it is plugged into (or it switches device
mode at a user's request, these things are complex little beasts...)
- usb offload fixes (where USB audio tunnels through the controller
while the main CPU is asleep) for when EMP spikes hit the system
causing disconnects to happen (as often happens with static
electricity in the winter months). This has been much reported by
at least one vendor, and resolves the issues they have been seeing
with this codepath. Can't wait for the "formal methods are the
answer!" people to try to model that one properly...
- Other small usb driver fixes for issues reported.
All of these have been in linux-next this week, and before, with no
reported issues, and I've personally been stressing these harder than
normal on my systems here with no problems.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
iG0EABECAC0WIQT0tgzFv3jCIUoxPcsxR9QN2y37KQUCadIdow8cZ3JlZ0Brcm9h
aC5jb20ACgkQMUfUDdst+ylTVQCfYg21ERvg753HnW1zbnzjMFlJZTQAoL+/0zeH
VcNijzt/8uiEKvQWAhcB
=BcZf
-----END PGP SIGNATURE-----
Merge tag 'usb-7.0-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Pull USB/Thunderbolt fixes from Greg KH:
"Here are a bunch of USB and Thunderbolt fixes (most all are USB) for
7.0-rc7. More than I normally like this late in the release cycle,
partly due to my recent travels, and partly due to people banging away
on the USB gadget interfaces and apis more than normal (big shoutout
to Android for getting the vendors to actually work upstream on this,
that's a huge win overall for everyone here)
Included in here are:
- Small thunderbolt fix
- new USB serial driver ids added
- typec driver fixes
- gadget driver fixes for some disconnect issues
- other usb gadget driver fixes for reported problems with binding
and unbinding devices as happens when a gadget device connects /
disconnects from a system it is plugged into (or it switches device
mode at a user's request, these things are complex little
beasts...)
- usb offload fixes (where USB audio tunnels through the controller
while the main CPU is asleep) for when EMP spikes hit the system
causing disconnects to happen (as often happens with static
electricity in the winter months). This has been much reported by
at least one vendor, and resolves the issues they have been seeing
with this codepath. Can't wait for the "formal methods are the
answer!" people to try to model that one properly...
- Other small usb driver fixes for issues reported.
All of these have been in linux-next this week, and before, with no
reported issues, and I've personally been stressing these harder than
normal on my systems here with no problems"
* tag 'usb-7.0-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (39 commits)
usb: gadget: f_hid: move list and spinlock inits from bind to alloc
usb: host: xhci-sideband: delegate offload_usage tracking to class drivers
usb: core: use dedicated spinlock for offload state
usb: cdns3: gadget: fix state inconsistency on gadget init failure
usb: dwc3: imx8mp: fix memory leak on probe failure path
usb: gadget: f_uac1_legacy: validate control request size
usb: ulpi: fix double free in ulpi_register_interface() error path
usb: misc: usbio: Fix URB memory leak on submit failure
USB: core: add NO_LPM quirk for Razer Kiyo Pro webcam
usb: cdns3: gadget: fix NULL pointer dereference in ep_queue
usb: core: phy: avoid double use of 'usb3-phy'
USB: serial: option: add MeiG Smart SRM825WN
usb: gadget: f_rndis: Fix net_device lifecycle with device_move
usb: gadget: f_subset: Fix net_device lifecycle with device_move
usb: gadget: f_eem: Fix net_device lifecycle with device_move
usb: gadget: f_ecm: Fix net_device lifecycle with device_move
usb: gadget: u_ncm: Add kernel-doc comments for struct f_ncm_opts
usb: gadget: f_rndis: Protect RNDIS options with mutex
usb: gadget: f_subset: Fix unbalanced refcnt in geth_free
dt-bindings: connector: add pd-disable dependency
...
When the mci->pvt_info allocation in edac_mc_alloc() fails, the error path
will call put_device() which will end up calling the device's release
function.
However, the init ordering is wrong such that device_initialize() happens
*after* the failed allocation and thus the device itself and the release
function pointer are not initialized yet when they're called:
MCE: In-kernel MCE decoding enabled.
------------[ cut here ]------------
kobject: '(null)': is not initialized, yet kobject_put() is being called.
WARNING: lib/kobject.c:734 at kobject_put, CPU#22: systemd-udevd
CPU: 22 UID: 0 PID: 538 Comm: systemd-udevd Not tainted 7.0.0-rc1+ #2 PREEMPT(full)
RIP: 0010:kobject_put
Call Trace:
<TASK>
edac_mc_alloc+0xbe/0xe0 [edac_core]
amd64_edac_init+0x7a4/0xff0 [amd64_edac]
? __pfx_amd64_edac_init+0x10/0x10 [amd64_edac]
do_one_initcall
...
Reorder the calling sequence so that the device is initialized and thus the
release function pointer is properly set before it can be used.
This was found by Claude while reviewing another EDAC patch.
Fixes: 0bbb265f70 ("EDAC/mc: Get rid of silly one-shot struct allocation in edac_mc_alloc()")
Reported-by: Claude Code:claude-opus-4.5
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Reviewed-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Cc: stable@kernel.org
Link: https://patch.msgid.link/20260331121623.4871-1-bp@kernel.org
Clippy fires two clippy::precedence warnings on the manual
byte-shifting expression:
warning: operator precedence can trip the unwary
--> drivers/gpu/nova-core/vbios.rs:511:17
|
511 | / u32::from(data[29]) << 24
512 | | | u32::from(data[28]) << 16
513 | | | u32::from(data[27]) << 8
| |______________________________________________^
Clear the warnings by replacing manual byte-shifting with
u32::from_le_bytes(). Using from_le_bytes() is also a tiny code
improvement, because it uses less code and is clearer about the intent.
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Link: https://patch.msgid.link/20260404212831.78971-2-jhubbard@nvidia.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
The current implementation does not actually set the default values for
the fields in the bitfield.
Fixes: 3fa145bef5 ("gpu: nova-core: register: generate correct `Default` implementation")
Signed-off-by: Eliot Courtney <ecourtney@nvidia.com>
Link: https://patch.msgid.link/20260401-fix-bitfield-v2-1-2fa68c98114a@nvidia.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
Commit a88831502c ("gpu: nova-core: falcon: use dma::Coherent")
dropped the nova-local `DmaObject` device memory type for the
kernel-global `Coherent` one.
This switch had a side-effect: `DmaObject` always aligned the requested
size to `PAGE_SIZE`, and also reported that adjusted size when queried.
`Coherent`, on the other hand, does page-align allocation sizes but only
allows CPU access on the exact size provided by the caller.
This change runs into a limitation of falcon DMA copies, namely that DMA
accesses are done on blocks of exactly 256 bytes. If the provided data
does not have a length that is a multiple of 256, `dma_wr` returns
an error.
It was expected that all firmwares would present the proper adjusted
size, but this is not the case at least on my GA107:
NovaCore 0000:08:00.0: DMA transfer goes beyond range of DMA object
NovaCore 0000:08:00.0: Failed to load FWSEC firmware: EINVAL
NovaCore 0000:08:00.0: probe with driver NovaCore failed with error -22
Fix this by padding the `Coherent`'s size to `MEM_BLOCK_ALIGNMENT` (i.e.
256) when allocating it and filling it with zeroes, before copying the
firmware on top of it.
Fixes: a88831502c ("gpu: nova-core: falcon: use dma::Coherent")
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Reviewed-by: Gary Guo <gary@garyguo.net>
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Link: https://patch.msgid.link/20260405-falcon-dma-roundup-v2-1-4af5b2ff9c16@nvidia.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
`driver_read_area` and `driver_write_area` are internal methods that
return slices containing the area of the command queue buffer that the
driver has exclusive read or write access, respectively.
While their returned value is correct and safe to use, internally they
temporarily create a reference to the whole command-buffer slice,
including GSP-owned regions. These regions can change without notice,
and thus creating a slice to them, even if never accessed, is undefined
behavior.
Fix this by making these methods create slices to valid regions only.
Fixes: 75f6b1de81 ("gpu: nova-core: gsp: Add GSP command queue bindings and handling")
Reported-by: Danilo Krummrich <dakr@kernel.org>
Closes: https://lore.kernel.org/all/DH47AVPEKN06.3BERUSJIB4M1R@kernel.org/
Signed-off-by: Alexandre Courbot <acourbot@nvidia.com>
Reviewed-by: Gary Guo <gary@garyguo.net>
Link: https://patch.msgid.link/20260404-cmdq-ub-fix-v5-1-53d21f4752f5@nvidia.com
Signed-off-by: Danilo Krummrich <dakr@kernel.org>
Commit 6a859f1a19 ("powerpc: unify two CONFIG_POWERPC64_CPU entries
in the same choice block") removed the only occurrence of this tricky
use case.
Disallow this pattern in choice_check_sanity() and revert commit
4d46b5b623 ("kconfig: fix infinite loop in sym_calc_choice()").
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Link: https://patch.msgid.link/20260330115736.1559962-1-masahiroy@kernel.org
Reviewed-by: Nicolas Schier <nsc@kernel.org>
Signed-off-by: Nicolas Schier <nsc@kernel.org>
The debug information part of the "Absolute filenames" section in the
reproducible builds document only mentions providing
'-fdebug-prefix-map' to KCFLAGS but it needs to be provided to KAFLAGS
as well since debug information has been generated for assembly files
for a long time.
Additionally, mention that the build directory may also appear as an
absolute path in the debug information (via DW_AT_comp_dir), so it needs
to be overridden via '-fdebug-prefix-map' as well.
Reported-by: Alexander Coffin <alex@cyberialabs.net>
Closes: https://lore.kernel.org/b8dfe7035d19fd611b9be55ee3145fdb@purelymail.com/
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Link: https://patch.msgid.link/20260313-kbuild-docs-repro-builds-fdebug-prefix-map-updates-v1-1-3aeeef7fa710@kernel.org
Signed-off-by: Nicolas Schier <nsc@kernel.org>
On MIPS the checksyscalls.sh script may be executed multiple times.
Currently these multiple executions are executed on each build as kbuild
see that the commands have changed each time.
Use a dedicated stamp file for each different invocation to avoid the
spurious executions.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nicolas Schier <nsc@kernel.org>
Link: https://patch.msgid.link/20260402-kbuild-missing-syscalls-v3-3-6641be1de2db@weissschuh.net
Signed-off-by: Nicolas Schier <nsc@kernel.org>
Currently checksyscalls.sh is unconditionally executed during each build.
Most of these executions are unnecessary.
Only run checksyscalls.sh if one of its inputs have changed.
This new logic does not work for the multiple invocations done for MIPS.
The effect is that checksyscalls.sh is still executed unconditionally.
However this is not worse than before.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nicolas Schier <nsc@kernel.org>
Link: https://patch.msgid.link/20260402-kbuild-missing-syscalls-v3-2-6641be1de2db@weissschuh.net
Signed-off-by: Nicolas Schier <nsc@kernel.org>
An upcoming patch will need to reuse this path.
Move it into a reusable variable.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Nicolas Schier <nsc@kernel.org>
Link: https://patch.msgid.link/20260402-kbuild-missing-syscalls-v3-1-6641be1de2db@weissschuh.net
Signed-off-by: Nicolas Schier <nsc@kernel.org>
As we are missing ID_AA64PFR2_EL1.GCIE from the kernel feature set,
userspace cannot write ID_AA64PFR2_EL1 with GCIE set, even if we are
on a GICv5 host.
Add the required field description.
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Link: https://patch.msgid.link/20260401170017.369529-1-maz@kernel.org
Signed-off-by: Marc Zyngier <maz@kernel.org>
if de goes negative right under us, there's nothing to prevent inode
getting freed just as we call coda_flag_inode(). We are not holding
->d_lock, so it's not impossible. Not going to be reproducible on
bare hardware unless it's a realtime config, but it could happen on KVM.
Trivial to fix - just hold rcu_read_lock() over that loop.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
d_really_is_negative(dentry) is a check for d_inode(dentry) being NULL;
rechecking that is pointless (and no, it can't race - the caller is holding
->d_lock, so ->d_inode is stable)
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
XIP has a history of being broken for long periods of time. In 2023, it was
broken for 18 months before getting fixed [1]. In 2024 it was 4 months [2].
And now it is broken again since commit a44fb57221 ("riscv: Add runtime
constant support"), 10 months ago.
These are clear signs that XIP feature is not being used.
I occasionally looked after XIP, but mostly because I was bored and had
nothing better to do.
Remove XIP support. Revert is possible if someone shows up complaining.
Link: https://lore.kernel.org/linux-riscv/20231212-customary-hardcover-e19462bf8e75@wendy/ [1]
Link: https://lore.kernel.org/linux-riscv/20240526110104.470429-1-namcao@linutronix.de/ [2]
Signed-off-by: Nam Cao <namcao@linutronix.de>
Cc: Frederik Haxel <haxel@fzi.de>
Cc: Vitaly Wool <vitaly.wool@konsulko.com>
Reviewed-by: Jisheng Zhang <jszhang@kernel.org>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://patch.msgid.link/20260202115403.2119218-1-namcao@linutronix.de
[pjw@kernel.org: updated to apply]
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Scalar misaligned access probe and vector misaligned access probe share
very similar code. Split out this similar part from scalar probe into
compare_unaligned_access(), which will be reused for vector probe in a
follow-up commit.
Signed-off-by: Nam Cao <namcao@linutronix.de>
Link: https://patch.msgid.link/3695f77279d473eead8ed6210d97c941321cd4f1.1770830596.git.namcao@linutronix.de
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Byte cycle measurement and word cycle measurement of scalar misaligned
access are very similar. Split these parts out into a common
measure_cycles() function to avoid duplication.
This function will also be reused for vector misaligned access probe in a
follow-up commit.
Signed-off-by: Nam Cao <namcao@linutronix.de>
Link: https://patch.msgid.link/50d0598e45acc56c95176e52fbbe56e1f4becc84.1770830596.git.namcao@linutronix.de
Signed-off-by: Paul Walmsley <pjw@kernel.org>
check_unaligned_access_speed_all_cpus() is more complicated than it should
be:
- It uses on_each_cpu() to probe unaligned memory access on all CPUs but
excludes CPU0 with a check in the callback function. So an IPI to CPU0
is wasted.
- Probing on CPU0 is done with smp_call_on_cpu(), which is not as fast as
on_each_cpu().
The reason for this design is because the probe is timed with jiffies.
Therefore on_each_cpu() excludes CPU0 because that CPU needs to tend to
jiffies.
Instead, replace jiffies usage with ktime_get_mono_fast_ns(). With jiffies
out of the way, on_each_cpu() can be used for all CPUs and
smp_call_on_cpu() can be dropped.
To make ktime_get_mono_fast_ns() usable, move this probe to late_initcall.
Anything after clocksource's fs_initcall works, but avoid depending on
clocksource staying at fs_initcall.
The choice of probe time is now 8000000 ns, which is the same as before (2
jiffies) for riscv defconfig. This is excessive for the CPUs I have, and
probably should be reduced; but that's a different discussion.
Suggested-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Nam Cao <namcao@linutronix.de>
Link: https://patch.msgid.link/9b9a20affe2e4f5c380926ceb885a47e20a59395.1770830596.git.namcao@linutronix.de
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Add an assembly implementation of strrchr() for RISC-V.
This implementation minimizes instruction count and avoids unnecessary
memory access to the stack. The performance benefits are most visible
on small workloads (1-16 bytes) where the architectural savings in
function overhead outweigh the execution time of the scan loop.
Benchmark results (QEMU TCG, rv64):
Length | Original (MB/s) | Optimized (MB/s) | Improvement
-------|-----------------|------------------|------------
1 B | 20 | 21 | +5.0%
7 B | 111 | 120 | +8.1%
16 B | 189 | 199 | +5.3%
512 B | 361 | 382 | +5.8%
4096 B | 388 | 391 | +0.8%
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Tested-by: Joel Stanley <joel@jms.id.au>
Link: https://patch.msgid.link/20260130025018.172925-9-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Add an assembly implementation of strchr() for RISC-V.
By eliminating stack frame management (prologue/epilogue) and optimizing
the function entries, the assembly version provides significant relative
gains for short strings where the fixed overhead of the C function is
most prominent. As string length increases, performance converges with
the generic C implementation.
Benchmark results (QEMU TCG, rv64):
Length | Original (MB/s) | Optimized (MB/s) | Improvement
-------|-----------------|------------------|------------
1 B | 21 | 22 | +4.8%
7 B | 113 | 121 | +7.1%
16 B | 195 | 202 | +3.6%
512 B | 376 | 389 | +3.5%
4096 B | 394 | 393 | -0.3%
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Tested-by: Joel Stanley <joel@jms.id.au>
Link: https://patch.msgid.link/20260130025018.172925-8-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Add an optimized strnlen() implementation for RISC-V. This version
includes a generic optimization and a Zbb-powered optimization using
the 'orc.b' instruction, derived from the strlen() implementation.
Benchmark results (QEMU TCG, rv64):
Length | Original (MB/s) | Optimized (MB/s) | Improvement
-------|-----------------|------------------|------------
16 B | 179 | 309 | +72.6%
512 B | 347 | 1562 | +350.1%
4096 B | 356 | 1878 | +427.5%
Suggested-by: Qingfang Deng <dqfext@gmail.com>
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Link: https://patch.msgid.link/20260130025018.172925-7-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Extend the string benchmarking suite to include strnlen(), strchr(),
and strrchr().
For character search functions strchr() and strrchr(), the benchmark
targets the NUL character. This ensures the entire string is scanned,
providing a consistent measure of full-length processing efficiency
comparable to strlen().
Suggested-by: Andy Shevchenko <andy@kernel.org>
Suggested-by: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Acked-by: Andy Shevchenko <andy@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20260130025018.172925-6-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Introduce a benchmarking framework to the string_kunit test suite to
measure the execution efficiency of string functions.
The implementation is inspired by crc_benchmark(), measuring throughput
(MB/s) and latency (ns/call) across a range of string lengths. It
includes a warm-up phase, disables preemption during measurement, and
uses a fixed seed for reproducible results.
This framework allows for comparing different implementations (e.g.,
generic C vs. architecture-optimized assembly) within the KUnit
environment.
Initially, provide a benchmark for strlen().
Suggested-by: Andy Shevchenko <andy@kernel.org>
Suggested-by: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Reviewed-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20260130025018.172925-5-jiangfeng@kylinos.cn
[pjw@kernel.org: fixed a checkpatch issue]
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Add a KUnit test for strrchr() to verify correctness across
different string lengths and memory alignments. Use vmalloc()
to place the NUL character at the page boundary to ensure
over-reads are detected.
Suggested-by: Kees Cook <kees@kernel.org>
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Reviewed-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20260130025018.172925-4-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Add a KUnit test for strnlen() to verify correctness across
different string lengths and memory alignments. Use vmalloc()
to place the NUL character at the page boundary to ensure
over-reads are detected.
Suggested-by: Andy Shevchenko <andy@kernel.org>
Suggested-by: Kees Cook <kees@kernel.org>
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Reviewed-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20260130025018.172925-3-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
Add a KUnit test for strlen() to verify correctness across
different string lengths and memory alignments. Use vmalloc()
to place the NUL character at the page boundary to ensure
over-reads are detected.
Suggested-by: Kees Cook <kees@kernel.org>
Signed-off-by: Feng Jiang <jiangfeng@kylinos.cn>
Reviewed-by: Kees Cook <kees@kernel.org>
Link: https://patch.msgid.link/20260130025018.172925-2-jiangfeng@kylinos.cn
Signed-off-by: Paul Walmsley <pjw@kernel.org>
The vdso_cfi build process copies source files (*.c, *.S) from the main
vdso directory to the build directory. Without a .gitignore file, these
copied files appear as untracked files in git status, cluttering the
working directory.
Add a .gitignore file to exclude:
- Copied source files (*.c, *.S)
- Temporary build files (vdso.lds, *.tmp, vdso-syms.S)
- While preserving vdso-cfi.S which is the original entry point
This follows the same pattern used in the main vdso directory
and keeps the working directory clean.
Signed-off-by: Chen Pei <cp0613@linux.alibaba.com>
Link: https://patch.msgid.link/20260320021850.1877-3-cp0613@linux.alibaba.com
Signed-off-by: Paul Walmsley <pjw@kernel.org>
When building VDSO with CFI support, source files are copied from the main
VDSO directory to the CFI build directory as part of the build process.
However, these copied source files were not removed during 'make clean',
leaving temporary files in the build directory.
Add the clean-files variable to ensure that these copied .c and .S files
are properly cleaned up. The notdir() function is used to strip the path
prefix, as clean-files expects relative file names without directory
components.
This ensures the build directory is left in a clean state after make clean.
Signed-off-by: Chen Pei <cp0613@linux.alibaba.com>
Link: https://patch.msgid.link/20260320021850.1877-2-cp0613@linux.alibaba.com
Signed-off-by: Paul Walmsley <pjw@kernel.org>
RISC-V has implemented pte_pgprot() and selects GENERIC_IOREMAP,
which provides a generic ioremap_prot() implementation. Enable
HAVE_IOREMAP_PROT to activate generic_access_phys() support, which
is useful for debugging (e.g., accessing /dev/mem via gdb).
Also update the architecture support documentation accordingly.
Signed-off-by: Yufeng Wang <wangyufeng@kylinos.cn>
Link: https://patch.msgid.link/20260306112734.108186-1-r4o5m6e8o@163.com
Signed-off-by: Paul Walmsley <pjw@kernel.org>
ACPI 6.6 is required for RISC-V as it introduces RISC-V specific
tables such as RHCT (RISC-V Hart Capabilities Table) and
RIMT (RISC-V I/O Mapping Table).
Update the FADT revision check from 6.5 to 6.6 and remove
the TODO comment since ACPI 6.6 has been officially released.
Signed-off-by: Yufeng Wang <wangyufeng@kylinos.cn>
Reviewed-by: Sunil V L <sunilvl@oss.qualcomm.com>
Acked-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Yao Zi <me@ziyao.cc>
Link: https://patch.msgid.link/20260305091433.83983-1-r4o5m6e8o@163.com
Signed-off-by: Paul Walmsley <pjw@kernel.org>
