| Age | Commit message (Collapse) | Author |
|---|
|
The generic CPU cache maintenance framework provides a way to register
drivers for devices implementing the underlying support for
cpu_cache_has_invalidate_memregion(). Enable it for arm64 by selecting
GENERIC_CPU_CACHE_MAINTENANCE which provides the implementation for,
and in turn selects, ARCH_HAS_CPU_CACHE_INVALIDATE_MEMREGION.
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
|
|
ARCH_HAS_CPU_CACHE_INVALIDATE_MEMREGION provides the mechanism for
invalidating certain memory regions in a cache-incoherent manner. Currently
this is used by NVDIMM and CXL memory drivers in cases where it is
necessary to flush all data from caches by physical address range.
The operations in question are effectively memory hotplug, where stale
data might otherwise remain in the caches.
This is separate from the invalidates done to enable use of non-coherent
DMA masters, primarily in terms of when it is needed (not related to DMA
mappings) and how deep the flush must push data. The flushes done for
non-coherent DMA only need to reach the Point of Coherence of a single host
(which is often nearer CPUs and DMA masters than the physical storage).
This operation must push the data out of non architectural caches
(memory-side caches, write buffers etc) and typically all the way to the
memory device.
In some architectures these operations are supported by system components
that may become available only later in boot as they are either present
on a discoverable bus, or via a firmware description of an MMIO interface
(e.g. ACPI DSDT). Provide a framework to handle this case.
Architectures can opt in for this support via
CONFIG_GENERIC_CPU_CACHE_MAINTENANCE
Add a registration framework. Each driver provides an ops structure and
the first op is Write Back and Invalidate by PA Range. The driver may
over invalidate.
For systems that can perform this operation asynchronously an optional
completion check operation is also provided. If present that must be called
to ensure that the action has finished. This provides a considerable
performance advantage if multiple agents are involved in the maintenance
operation.
When multiple agents are present in the system each should register with
this framework and the core code will issue the invalidate to all of them
before checking for completion on each. This is done to avoid need for
filtering in the core code which can become complex when interleave,
potentially across different cache coherency hardware is going on, so it
is easier to tell everyone and let those who don't care do nothing.
Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
Co-developed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
|
|
Fully initialize *ctx, including the buf field which sha256_init()
doesn't initialize, to avoid a KMSAN warning when comparing *ctx to
orig_ctx. This KMSAN warning slipped in while KMSAN was not working
reliably due to a stackdepot bug, which has now been fixed.
Fixes: 6733968be7cb ("lib/crypto: tests: Add tests and benchmark for sha256_finup_2x()")
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20251121033431.34406-1-ebiggers@kernel.org
Signed-off-by: Eric Biggers <ebiggers@kernel.org>
|
|
The igen6_edac driver calls device_initialize() for all memory
controllers in igen6_register_mci(), but misses corresponding
put_device() calls in error paths and during normal shutdown in
igen6_unregister_mcis().
Adding the missing put_device() calls improves code readability and
ensures proper reference counting for the device structure.
Found by code review.
Signed-off-by: Ma Ke <make24@iscas.ac.cn>
Reviewed-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://patch.msgid.link/20251105090244.23327-1-make24@iscas.ac.cn
|
|
Setup the following debugfs testing node to enable fake memory error
address decoding tests for the imh_edac driver.
/sys/kernel/debug/edac/imh_test/addr
Tested-by: Yi Lai <yi1.lai@intel.com>
Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://patch.msgid.link/20251119134132.2389472-8-qiuxu.zhuo@intel.com
|
|
Detect 2-level memory configurations and notify the 'skx_common' library
to enable ADXL 2-level memory error decoding.
Tested-by: Yi Lai <yi1.lai@intel.com>
Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://patch.msgid.link/20251119134132.2389472-7-qiuxu.zhuo@intel.com
|
|
The allowed maximum number of row bits for DRAM chips in the Diamond
Rapids server processor is up to 19. Extend the current maximum row
bits from 18 to 19.
Tested-by: Yi Lai <yi1.lai@intel.com>
Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://patch.msgid.link/20251119134132.2389472-6-qiuxu.zhuo@intel.com
|
|
Intel Diamond Rapids CPUs include Integrated Memory and I/O Hubs (IMH).
The memory controllers within the IMHs provide memory stacks to the
processor. Create a new driver for this IMH-based memory controllers
rather than applying additional patches to the existing i10nm_edac.c
for the following reasons:
1) The memory controllers are not presented as PCI devices; instead,
the detection and all their registers have been transitioned to
MMIO-based memory spaces.
2) Validation processes are costly. Modifications to i10nm_edac would
require extensive validation checks against multiple platforms,
including Ice Lake, Sapphire Rapids, Emerald Rapids, Granite Rapids,
Sierra Forest, and Grand Ridge.
3) Future Intel CPUs will likely only need patches on top of this new
EDAC driver. Validation can be limited to Diamond Rapids servers
and future Intel CPU generations.
[Tony: Fix kerneldoc for struct local_reg]
[randconfig: Added dependencies on NFIT and DMI]
Tested-by: Yi Lai <yi1.lai@intel.com>
Signed-off-by: Qiuxu Zhuo <qiuxu.zhuo@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
Link: https://patch.msgid.link/20251119134132.2389472-5-qiuxu.zhuo@intel.com
|
|
Call paths leading to __virt_pg_map() are currently:
(a) virt_pg_map() -> virt_arch_pg_map() -> __virt_pg_map()
(b) virt_map_level() -> __virt_pg_map()
For (a), calls to virt_pg_map() from kvm_util.c make sure they update
vm->vpages_mapped, but other callers do not. Move the sparsebit_set()
call into virt_pg_map() to make sure all callers are captured.
For (b), call sparsebit_set_num() from virt_map_level().
It's tempting to have a single the call inside __virt_pg_map(), however:
- The call path in (a) is not x86-specific, while (b) is. Moving the
call into __virt_pg_map() would require doing something similar for
other archs implementing virt_pg_map().
- Future changes will reusue __virt_pg_map() for nested PTEs, which should
not update vm->vpages_mapped, i.e. a triple underscore version that does
not update vm->vpages_mapped would need to be provided.
Signed-off-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Link: https://patch.msgid.link/20251021074736.1324328-12-yosry.ahmed@linux.dev
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Replace __virt_pg_map() calls in tests by high-level equivalent
functions, removing some loops in the process.
No functional change intended.
Signed-off-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Link: https://patch.msgid.link/20251021074736.1324328-11-yosry.ahmed@linux.dev
Signed-off-by: Sean Christopherson <seanjc@google.com>
|
|
Pull drm fixes from Dave Airlie:
"A range of small fixes across the board, the i915 display
disambiguation is probably the biggest otherwise amdgpu and xe as
usual with tegra, nouveau, radeon and a core atomic fix.
Looks mostly normal.
atomic:
- Return error codes on failed blob creation for planes
nouveau:
- Fix memory leak
tegra:
- Fix device ref counting
- Fix pid ref counting
- Revert booting on Pixel C
xe:
- Fix out-of-bounds access with BIT()
- Fix kunit test checking wrong condition
- Drop duplicate kconfig select
- Fix guc2host irq handler with MSI-X
i915:
- Wildcat Lake and Panther Lake detangled for display fixes
amdgpu:
- DTBCLK gating fix
- EDID fetching retry improvements
- HDMI HPD debounce filtering
- DCN 2.0 cursor fix
- DP MST PBN fix
- VPE fix
- GC 11 fix
- PRT fix
- MMIO remap page fix
- SR-IOV fix
radeon:
- Fence deadlock fix"
* tag 'drm-fixes-2025-11-21' of https://gitlab.freedesktop.org/drm/kernel: (25 commits)
drm/amdgpu: Add sriov vf check for VCN per queue reset support.
drm/amdgpu/ttm: Fix crash when handling MMIO_REMAP in PDE flags
drm/amdgpu/vm: Check PRT uAPI flag instead of PTE flag
drm/amdgpu: Skip emit de meta data on gfx11 with rs64 enabled
drm/amd: Skip power ungate during suspend for VPE
drm/plane: Fix create_in_format_blob() return value
drm/xe/irq: Handle msix vector0 interrupt
drm/xe: Remove duplicate DRM_EXEC selection from Kconfig
drm/xe/kunit: Fix forcewake assertion in mocs test
drm/xe: Prevent BIT() overflow when handling invalid prefetch region
drm/radeon: delete radeon_fence_process in is_signaled, no deadlock
drm/amd/display: Fix pbn to kbps Conversion
drm/amd/display: Clear the CUR_ENABLE register on DCN20 on DPP5
drm/amd/display: Add an HPD filter for HDMI
drm/amd/display: Increase DPCD read retries
drm/amd/display: Move sleep into each retry for retrieve_link_cap()
drm/amd/display: Prevent Gating DTBCLK before It Is Properly Latched
drm/i915/xe3: Restrict PTL intel_encoder_is_c10phy() to only PHY A
drm/i915/display: Add definition for wcl as subplatform
drm/pcids: Split PTL pciids group to make wcl subplatform
...
|
|
Mauro says:
That's the final series to complete the migration of documentation
build: it converts get_feat from Perl to Python.
V2 is technically identical to v1: the only difference is that it
now uses tools/lib/python/feat to store the library logic.
With that, no Sphinx in-kernel extensions use fork anymore to call
ancillary scripts: everything is now importing Python methods
directly from the libraries.
There's nothing special on this conversion: it is a direct translation,
almost bug-compatible with the original version (*).
(*) I did solve two or three caveats on patch 1.
Most of the complexity of the script relies at the logic to produce
ReST tables. I do have here on my internal scripts a (somewhat) generic
formatter for ReST tables in Python. I was tempted to convert the logic
to use it, but, as this could cause regressions, I opted to not do it
right now, mainly because the matrix table logic is complex. Also,
I'm tempted to modify a little bit the output there, but extra tests
are required to see if PDF output would work with complex tables (I
remember I had a problem with that in the past). So, I'm postponing
such extra cleanup.
|
|
Now that this was rewritten in Python, we can remove the old
tool.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <1f53e0fa48616af189ce98b45a65cc0c245e7aaf.1763492868.git.mchehab+huawei@kernel.org>
|
|
Now that get_feat is in Python, we don't need to use subprocess
to fork an executable file: we can use the feature classes
directly.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <c59d2542d7cc914fd5f8c84df966e63adc924cdc.1763492868.git.mchehab+huawei@kernel.org>
|
|
As we want to call Python code directly at the Sphinx extension,
convert get_feat.pl to Python.
The code was made to be (almost) bug-compatible with the Perl
version, with two exceptions:
1. Currently, Perl script outputs a wrong table if arch is set
to a non-existing value;
2. the ReST table output when --feat is used without --arch
has an invalid format, as the number of characters for the
table delimiters are wrong.
Those two bugs were fixed while testing the conversion.
Additionally, another caveat was solved:
the output when --feat is used without arch and the feature
doesn't exist doesn't contain an empty table anymore.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <03c26cee1ec567804735a33047e625ef5ab7bfa8.1763492868.git.mchehab+huawei@kernel.org>
|
|
IRQF_SHARED is the only flag handled in __serial8250_isa_init_ports()
and serial8250_probe_platform(). There is no need to precompute the
flags. Instead, initialize port->irqflags directly in the for loop.
Note the "if (bool)" is cheap and these are not hot paths anyway.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119092457.826789-7-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
share_irqs is used solely in 8250_platform. Make it local to that file.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119092457.826789-6-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
8250_core is the only place where skip_txen_test is used. And platform
and core end up in 8250.ko, so there is no change in module name (param
prefix). Therefore, move skip_txen_test there and make it local.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119092457.826789-5-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
In 3.7, 8250 was unintentionally renamed to 8250_core. This happened in
the commit 835d844d1a28 (8250_pnp: do pnp probe before legacy probe).
This made 8250.<xxxx> module options effectively defunct. Instead,
8250_core.<xxxx> worked.
In 3.9, the commit f2b8dfd9e480 (serial: 8250: Keep 8250.<xxxx> module
options functional after driver rename) made the original options work
again by introducing a hack.
Later in 3.9, the commit 9196d8acd7f9 (TTY: 8250, revert module name
change) changed the module name back to 8250 (from 8250_core). Since
then, the hack was there to support the transient 8250_core.<xxxx>
options. Those were present only in the 3.7..3.9 range.
These transient options were deprecated by 9326b047e4fd (TTY: 8250,
deprecated 8250_core.* options) in v3.9 too.
Now, after those 12 years, it is time to get rid of this hack
completely.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119092457.826789-4-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Provided the uart driver is available globally, there is no need to
store a pointer to it in struct cdns_uart. Instead, use the global
cdns_uart_uart_driver in the code directly.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Acked-by: Michal Simek <michal.simek@amd.com>
Link: https://patch.msgid.link/20251119092457.826789-3-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use the designated DIV_ROUND_UP() macro instead of explicit addition
with division.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119092457.826789-2-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use guards in the serial_core code. This improves readability, makes
error handling easier, and marks locked portions of code explicit. All
that while being sure the lock is unlocked.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-11-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Neither uart_do_autoconfig(), nor uart_wait_modem_status() can return
-ENOIOCTLCMD. The ENOIOCTLCMD checks are there to check if 'cmd' matched
against TIOCSERCONFIG, and TIOCMIWAIT respectively. (With 0 or error in
'ret', it does not matter.)
Therefore, the code can simply return from the TIOCSERCONFIG and
TIOCMIWAIT spots immediately.
To be more explicit, use 'if' instead of switch-case for those single
values.
And return without jumping to the 'out' label -- it can be removed too.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-10-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use guards in the vt/keyboard code. This improves readability, makes
error handling easier, and marks locked portions of code explicit. All
that while being sure the lock is unlocked.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-9-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Return immediately when something goes wrong in vt_do_kbkeycode_ioctl().
This makes the code flow more obvious.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-8-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The vt/keyboard code can use __free to ensure the temporary buffers are
freed. Perform the switch.
And even one non-temporary in kbd_connect(). There are fail paths, so
ensure the buffer is freed in them and not when returning 0 -- by
retain_and_null_ptr().
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-7-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use guards in the moxa code. This improves readability, makes error
handling easier, and marks locked portions of code explicit. All that
while being sure the lock is unlocked.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-6-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use guards in the n_hdlc code. This improves readability, makes error
handling easier, and marks locked portions of code explicit. All that
while being sure the lock is unlocked.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-5-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
The cases in the switch() of n_hdlc_tty_ioctl() can return immediately
-- no need to store into error and return later.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-4-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use guards in the n_tty code. This improves readability, makes error
handling easier, and marks locked portions of code explicit. All that
while being sure the lock is unlocked.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-3-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use guards in the pty code. This improves readability, makes error
handling easier, and marks locked portions of code explicit. All that
while being sure the lock is unlocked.
pty_set_pktmode() is handled specially -- the conditions are inverted
and return called if conditions unmet. This avoid double nested 'if's.
The variable is renamed to want_pktmode so it is not confused with the
current state of pktmode.
Signed-off-by: Jiri Slaby (SUSE) <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119100140.830761-2-jirislaby@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
This patch updates the Linux documentation for cscope, fixing two issues:
1. Corrects the typo in the command line:
c"scope -d -p10 -> cscope -d -p10
2. Fixes the related documentation comment for clarity and correctness:
cscope by default cscope.out database.
->
cscope by default uses the cscope.out database.
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <20251119065727.3500015-1-jiakaiPeanut@gmail.com>
|
|
Apparently as of version 2.42, glibc headers define AT_RENAME_NOREPLACE
and some of the other flags for renameat2() and friends in <stdio.h>.
Which would all be fine, except for inexplicable reasons glibc decided
to define them _differently_ from the kernel definitions, which then
makes some of our sample code that includes both kernel headers and user
space headers unhappy, because the compiler will (correctly) complain
about redefining things.
Now, mixing kernel headers and user space headers is always a somewhat
iffy proposition due to namespacing issues, but it's kind of inevitable
in our sample and selftest code. And this is just glibc being stupid.
Those defines come from the kernel, glibc is exposing the kernel
interfaces, and glibc shouldn't make up some random new expressions for
these values.
It's not like glibc headers changed the actual result values, but they
arbitrarily just decided to use a different expression to describe those
values. The kernel just does
#define AT_RENAME_NOREPLACE 0x0001
while glibc does
# define RENAME_NOREPLACE (1 << 0)
# define AT_RENAME_NOREPLACE RENAME_NOREPLACE
instead. Same value in the end, but very different macro definition.
For absolutely no reason.
This has since been fixed in the glibc development tree, so eventually
we'll end up with the canonical expressions and no clashes. But in the
meantime the broken headers are in the glibc-2.42 release and have made
it out into distributions.
Do a minimal work-around to make the samples build cleanly by just
undefining the affected macros in between the user space header include
and the kernel header includes.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Better split CONFIG_DRM_XE_DEBUG and CONFIG_DRM_XE_DEBUG_GUC optional
parts from the main code, creating smaller ct_dead_* and fast_req_*
interfaces.
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Link: https://patch.msgid.link/20251119152157.1675188-2-lucas.demarchi@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
|
|
Add missing stack_depot_init() call when CONFIG_DRM_XE_DEBUG_GUC is
enabled to fix the following call stack:
[] BUG: kernel NULL pointer dereference, address: 0000000000000000
[] Workqueue: drm_sched_run_job_work [gpu_sched]
[] RIP: 0010:stack_depot_save_flags+0x172/0x870
[] Call Trace:
[] <TASK>
[] fast_req_track+0x58/0xb0 [xe]
Fixes: 16b7e65d299d ("drm/xe/guc: Track FAST_REQ H2Gs to report where errors came from")
Tested-by: Sagar Ghuge <sagar.ghuge@intel.com>
Cc: stable@vger.kernel.org # v6.17+
Reviewed-by: Stuart Summers <stuart.summers@intel.com>
Link: https://patch.msgid.link/20251118-fix-debug-guc-v1-1-9f780c6bedf8@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
|
|
The row/column bounds (for a screen window box) are changed from
'offset one' to 'offset zero' and bound to the screen size using:
v->xs = min_t(u16, v->xs - 1, vc->vc_cols - 1);
This has the side effect of converting zero to the limit.
A check I'm adding to min_t() reports that (u16)(v->xs - 1) (etc)
discards signiticant bits (because v->xs is promoted to 'int' before
the addition).
If v->xs is zero (it comes from userspace) it converts -1 to 0xffff.
This is then bounded to 'vc->vc_cols - 1' which will be fine.
Replace with:
v->xs = umin(v->xs - 1, vc->vc_cols - 1);
which again converts a -1 to unsigned - this time to 0xffffffff,
with the same overall effect.
Whether zero is meant to mean the 'maximum size' is unknown.
I can't find any documentation for the ioctl and it pre-dates git.
Detected by an extra check added to min_t().
Signed-off-by: David Laight <david.laight.linux@gmail.com>
Reviewed-by: Jiri Slaby <jirislaby@kernel.org>
Link: https://patch.msgid.link/20251119224140.8616-28-david.laight.linux@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Automated testing caught this missing doc comment; add something suitable
(and useful).
Fixes: 330e76d31697 ("drm/imagination: Add power domain control")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Closes: https://lore.kernel.org/r/20251106152448.453b53ad@canb.auug.org.au/
Reviewed-by: Alessio Belle <alessio.belle@imgtec.com>
Link: https://patch.msgid.link/20251121-device-power-doc-fix-v2-1-3417779f36c7@imgtec.com
Signed-off-by: Matt Coster <matt.coster@imgtec.com>
|
|
The ACPI ECRD and ECWR functions have a 10ms sleep at the end. It turns
out, that this is sometimes needed to avoid I2C transmission failures,
especially for functions doing regmap_update_bits (and thus read + write
shortly after each other). This fixes problems like the following
appearing in the kernel log:
leds platform::micmute: Setting an LED's brightness failed (-6)
leds platform::kbd_backlight: Setting an LED's brightness failed (-6)
The ACPI QEVT function used to read the interrupt status register also
has a 10ms sleep at the end. Without that there are problems with
reading multiple events following directly after each other resulting
in the following error message being logged:
thinkpad-t14s-ec 4-0028: Failed to read event
Fixes: 60b7ab6ce030 ("platform: arm64: thinkpad-t14s-ec: new driver")
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Link: https://patch.msgid.link/20251119-thinkpad-t14s-ec-improvements-v2-2-441219857c02@kernel.org
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
|
|
Fix a race condition, that an input key related interrupt might be
triggered before the input handler has been registered, which results
in a NULL pointer dereference. This can happen if the user enables
the keyboard backlight shortly before the driver is being probed.
This fixes the following backtrace visible in dmesg:
Unable to handle kernel NULL pointer dereference at virtual address 00000000000000e0
...
Call trace:
sparse_keymap_report_event+0x2c/0x978 [sparse_keymap] (P)
t14s_ec_irq_handler+0x190/0x3e8 [lenovo_thinkpad_t14s]
irq_thread_fn+0x30/0xb8
irq_thread+0x18c/0x3b0
kthread+0x148/0x228
ret_from_fork+0x10/0x20
Fixes: 60b7ab6ce030 ("platform: arm64: thinkpad-t14s-ec: new driver")
Signed-off-by: Sebastian Reichel <sre@kernel.org>
Reviewed-by: Bryan O'Donoghue <bod@kernel.org>
Link: https://patch.msgid.link/20251119-thinkpad-t14s-ec-improvements-v2-1-441219857c02@kernel.org
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
|
|
Merge series from Andreas Kemnade <andreas@kemnade.info>:
Add a driver for the FP9931/JD9930 regulator which provides the
comparatively high voltages needed for electronic paper displays.
Datasheet for the FP9931 is at
https://www.fitipower.com/dl/file/flXa6hIchVeu0W3K
Although it is in English, it seems to be only downloadable
from the Chinese part of that website.
For the JD9930 there can be a datasheet found at
https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/196/JD9930_2D00_0.7_2D00_JUN_2D00_2019.pdf
To simplify things, include the hwmon part directly which is only
one register read and there are not other functions besides
regulators in this chip.
|
|
ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mani/mhi into char-misc-next
Manivannan writes:
MHI Host
========
- Add support for Telit FE990B40 and Foxconn T99W760 modems
MHI Endpoint
============
- Make 'struct mhi_ep_bus_type' const as the driver core now handles the const
bus_type.
- Add WQ_PERCPU flag to alloc_workqueue() as a part of the workqueue refactoring
* tag 'mhi-for-v6.19' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/mani/mhi:
bus: mhi: host: pci_generic: Add Foxconn T99W760 modem
bus: mhi: ep: add WQ_PERCPU to alloc_workqueue users
bus: mhi: host: pci_generic: Add Telit FE990B40 modem support
bus: mhi: ep: Make mhi_ep_bus_type const
|
|
snprintf() as defined by the C99 standard,returns the
number of characters that *would have been* written if
enough space were available.Use scnprintf() that returns
the actual number of characters written.
Link: https://github.com/KSPP/linux/issues/105
Signed-off-by: HariKrishna Sagala <hariconscious@gmail.com>
Link: https://patch.msgid.link/20251113142637.259737-2-hariconscious@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux into soc/dt
Qualcomm Arm64 DeviceTree updates for v6.19
Introduce support for the Redxa Dragon Q6A development board, the Huawei
MateBoot E 2019, the Asus ZenFone 2 Laser/Selfie, the MSM8937 platform
and the Xiaomo Redmi 3S device based on it.
SoC dtsi files for Agatti, Hamoa, Kodiak, Monaco, Purwa, and Talos, are
renamed in order to better facilitate the addition of new boards on the
various SKUs of these.
Cooling maps are introduced for the CPU cores in IPQ5424, and the
network subsystem clock controller is added.
On Lemans, RTC is enabled, the EVK fan controller is described and a
camera mezzanine overlay is introduced.
Touchscreen support is added to the BQ Aquaris M5, and the touchscreen
from Samsung Galaxy Core Prime is moved to the common platform to
benefit the other devices sharing common definitions.
On Agatti two more UARTs are described, as well as APR and the related audio
services, and the LPASS LPI pin controller. The RB1 board gets HDMI
autio playback support.
On Kodiak-based targets, Fairphone FP5 gains definitions of the UW camera
actuator, regulator for the ToF sensor, and haptic module. The SHIFT
SHIFTphone 8 gains RGB and flash LEDs, and Venus support. The Rb3Gen2
development board gets QUP firmware path defined, to support dynamic
loading of the serial engine firmware. Kodiak also gains Coresight
devices for AOSS and QDSS blocks.
Display support is added for the Talos platform, and enabled on the Ride
board. Talos also gains the definitions to scale DDR and L3
interconnects.
On SC8280XP, the camera privacy indicator on Lenovo Thinkpad X13s is
connected to the camera stack. Off-by-one GPI DMA channels are
corrected.
The SDM845-based LG and OnePlus custom defined rmtfs guard pages are
replaced with the inline-support for guard pages.
SDX75 DWC3 node is flattened and marked for USB role switching.
On SM8550, the camera subsystem and the S5K3M5 camera sensor is
introduced for the QRD, and an overlay for the "Rear Camera Card" for
the Hardware Development Kit (HDK) is introduced.
USB support is introduce for the SM8750 platform, and enabled in the MTP
and QRD devices.
On Hamoa, like on other devices the Asus Zenbook A14 definition of the
eDP panel is reworked to support both LCD and OLED configurations. WiFi
and Bluetooth is also enabled on the A14. The CRD gains support for
controlling charge limits.
The refgen regulator supplying DSI is defined and wired up on a variety
of platforms.
* tag 'qcom-arm64-for-6.19' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux: (138 commits)
arm64: dts: qcom: sdx75: Add missing usb-role-switch property
arm64: dts: qcom: sdx75: Flatten usb controller node
arm64: dts: qcom: HAMOA-IOT-SOM: Unreserve GPIOs blocking SPI11 access
arm64: dts: qcom: qrb2210-rb1: Fix UART3 wakeup IRQ storm
Revert "arm64: dts: qcom: sc7280: Increase config size to 256MB for ECAM feature"
arm64: dts: qcom: kodiak: add coresight nodes
arm64: dts: qcom: sdm845-oneplus: Describe TE gpio
arm64: dts: qcom: sdm845-oneplus: Implement panel sleep pinctrl
arm64: dts: qcom: sdm845-oneplus: Group panel pinctrl
arm64: dts: qcom: sdm845-oneplus: Update compatbible and add DDIC supplies
arm64: dts: qcom: qcs6490-rb3gen2: Rename vph-pwr regulator node
arm64: dts: qcom: qcm6490-fairphone-fp5: Add UW cam actuator
arm64: dts: qcom: qcm6490-fairphone-fp5: Enable CCI pull-up
arm64: dts: qcom: sm8750: Add USB support for SM8750 QRD platform
arm64: dts: qcom: sm8750: Add USB support for SM8750 MTP platform
arm64: dts: qcom: sm8750: Add USB support to SM8750 SoCs
arm64: dts: qcom: rename x1p42100 to purwa
arm64: dts: qcom: rename sc7280 to kodiak
arm64: dts: qcom: rename qcm2290 to agatti
arm64: dts: qcom: add gpu_zap_shader label
...
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux into soc/dt
Qualcomm Arm32 DeviceTree updates for v6.19
In addition to a variety of cleanups and reordering of nodes, four GSBIs
are added to the MSM8960 platform.
On the MSM8226-based Samsung Galaxy Grand 2, a simple framebuffer is
defined.
* tag 'qcom-arm32-for-6.19' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux:
ARM: dts: qcom: msm8226-samsung-ms013g: add simple-framebuffer
ARM: dts: qcom: msm8960: rename msmgpio node to tlmm
ARM: dts: qcom: msm8960: add I2C nodes for gsbi1 and gsbi8
ARM: dts: qcom: msm8960: add I2C nodes for gsbi10 and gsbi12
ARM: dts: qcom: msm8960: inline qcom-msm8960-pins.dtsi
ARM: dts: qcom: msm8960: reorder nodes and properties
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/khilman/linux-omap into soc/dt
Misc. DT updates, mostly for AM3 platforms.
* tag 'omap-for-v6.19/dt-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/khilman/linux-omap:
ARM: dts: ti/omap: fix incorrect compatible string in internal eeprom node
ARM: dts: omap: am335x-mba335x: Fix stray '/*' in comment
ARM: dts: omap: am335x-tqma335x/mba335x: Fix MicIn routing
dt-bindings: mmc: ti,omap2430-sdhci: convert to DT schema
arm: dts: ti: omap: Drop unnecessary properties for SDHCI node
arm: dts: ti: omap: am335x-pepper: Fix vmmc-supply property typo
ARM: dts: omap3: n900: Correct obsolete TWL4030 power compatible
ARM: dts: omap3: beagle-xm: Correct obsolete TWL4030 power compatible
ARM: dts: omap: Add support for TQMa335x/MBa335x
ARM: dts: omap: AM33xx: add cpu label
dt-bindings: omap: add AM335x-based TQMa335x SOM and MBa335x board
ARM: dts: am335x-netcom-plus-2xx: add missing GPIO labels
ARM: dts: ti: omap: am335x-baltos: add a regulator for the mPCIe slot
ARM: dts: omap: rework Baltos LED nodes
ARM: dts: am33xx: Add missing serial console speed
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/at91/linux into soc/dt
Microchip AT91 device tree updates for v6.19
This update includes:
- FIFO size fixes for some Flexcom UARTs and SPIs
* tag 'at91-dt-6.19' of https://git.kernel.org/pub/scm/linux/kernel/git/at91/linux:
ARM: dts: microchip: sama5d2: fix spi flexcom fifo size to 32
ARM: dts: microchip: sama7g5: fix uart fifo size to 32
ARM: dts: microchip: sama7d65: fix uart fifo size to 32
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
The selective fetch code doesn't handle asycn flips correctly.
There is a nonsense check for async flips in
intel_psr2_sel_fetch_config_valid() but that only gets called
for modesets/fastsets and thus does nothing for async flips.
Currently intel_async_flip_check_hw() is very unhappy as the
selective fetch code pulls in planes that are not even async
flips capable.
Reject async flips when selective fetch is enabled, until
someone fixes this properly (ie. disable selective fetch while
async flips are being issued).
Cc: stable@vger.kernel.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20251105171015.22234-1-ville.syrjala@linux.intel.com
Reviewed-by: Jouni Högander <jouni.hogander@intel.com>
|
|
WCD934x provides register read/write interface to soundwire controller via
slimbus. Eventhough there was no direct compile time dependency for
this so far, however after adding common wcd functions this dependency
became explict.
We can either move the soundwire specific bits from wcd-common to wcd-sdw.c or
something on those lines or explicity add this dependency in Kconfig.
Moving this wcd-sdw can fix reported compile issue but it does not make
sense to select this codec without soundwire configs.
Given the fact that WCD934x is very much providing soundwire functionality
it is better to specify the soundwire dependency explicity in Kconfig
which should also fix the below compile time error with combination
of wcd934x a built-in and soundwire a module and also allow this codec
selection only when soundwire is available.
Error log:
wcd-common.c:undefiined reference to `sdw_write'
sound/soc/codecs/wcd-common.o: in function `wcd_bus_config':
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202511210405.zd9wig5Z-lkp@intel.com/
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@oss.qualcomm.com>
Link: https://patch.msgid.link/20251121143258.229138-1-srinivas.kandagatla@oss.qualcomm.com
Signed-off-by: Mark Brown <broonie@kernel.org>
|
|
https://git.kernel.org/pub/scm/linux/kernel/git/ti/linux into soc/dt
TI K3 device tree updates for v6.19
Generic fixes and cleanups:
* Multiple SoCs: Disable CPSW in SoC files and
enable them in board files for better board-level control
* Replace rgmii-rxid with rgmii-id for CPSW ports across multiple boards
New Boards/SoM:
* AM62L SoC and basic support for EVM
* Toradex Aquila AM69 board support
* Kontron SMARC-sAM67 module and ADS2 carrier board support
Platform wide:
* Define possible system states amd wakeup-source (AM62/AM62A/AM62P)
SoC/EVM specific changes:
AM62:
* Add RNG node
* Add OLDI support
AM62P:
* Move audio_refclk to common main dtsi (k3-am62p-j722s-common-main)
* Fix memory ranges for GPU
AM62D2:
* Enable PMIC support on EVM
* Misc fixes
AM64:
* Add DMA support for TSCADC on EVM
AM69:
* Add Aquila board support with Clover variant
J722S:
* Fix audio refclk source in main dtsi
* Explicitly use PLL1_HSDIV6 audio refclk for EVM
J784S4/J742S2:
* Add bootph-all tag to support PCIe boot
Variscite VAR-SOM-AM62P:
* Add support for ADS7846 touchscreen
* Add support for WM8904 audio codec
* tag 'ti-k3-dt-for-v6.19' of https://git.kernel.org/pub/scm/linux/kernel/git/ti/linux: (42 commits)
arm64: dts: ti: k3-am62l: add initial reference board file
arm64: dts: ti: k3-am62l: add initial infrastructure
dt-bindings: arm: ti: Add binding for AM62L SoCs
arm64: dts: ti: am69-aquila: Add Clover
arm64: dts: ti: Add Aquila AM69 Support
dt-bindings: arm: ti: add Toradex Aquila AM69
arm64: dts: ti: k3-j721s2: disable "mcu_cpsw" in SoC file and enable in board files
arm64: dts: ti: k3-j721e: disable "mcu_cpsw" in SoC file and enable it in board file
arm64: dts: ti: k3-j7200: disable "mcu_cpsw" in SoC file and enable in board file
arm64: dts: ti: k3-am65: disable "mcu_cpsw" in SoC file and enable in board file
arm64: dts: ti: k3-am62: disable "cpsw3g" in SoC file and enable in board file
arm64: dts: ti: k3-am62p5-sk: Set wakeup-source system-states
arm64: dts: ti: k3-am62a7-sk: Set wakeup-source system-states
arm64: dts: ti: k3-am62-lp-sk: Set wakeup-source system-states
arm64: dts: ti: k3-am62p: Define possible system states
arm64: dts: ti: k3-am62a: Define possible system states
arm64: dts: ti: k3-am62: Define possible system states
arm64: dts: ti: k3-am62p-j722s-common-main: move audio_refclk here
arm64: dts: ti: k3-*: Replace rgmii-rxid with rgmii-id for CPSW ports
arm64: dts: ti: k3-am642-tqma64xxl: add boot phase tags
...
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux into soc/dt
SoCFPGA DTS updates for v6.19
- Add 4-bit SPI bus width(n5x, stratix10, agilex and agilex5)
- Agilex5 updates:
- Add GMAC0 for NAND daughter card
- Add SMMU support
- Add VGIC maintenance interrupt
- Add L2 and L3 cache
- Add support for the 013b board
- Add I3C support
- Add support for the Enclustra Mercury+ SA1 SoM based on Cyclone5
- Add support for Agilex3 board(a variant of the Agilex5 board)
- dt-bindings update:
- Document iommu in cdns,hp-nfc, snps,dw-axi-dmac and Agilex5
- Document Enclustra Mercury SA1 and AA1 boards
- Document Agilex5 013b board
- Document Agilex3 board
- Fix dtbs_check warnings:
- stratix10-swvp
- Agilex(NAND and Clock manager)
- Move sdmmc-ecc to base DTSI file(Stratix10)
* tag 'socfpga_dts_updates_for_v6.19' of git://git.kernel.org/pub/scm/linux/kernel/git/dinguyen/linux: (35 commits)
arm64: dts: socfpga: agilex5: update qspi partitions for 013b board
arm64: dts: socfpga: add Agilex3 board
dt-bindings: intel: Add Agilex3 SoCFPGA board
arm64: dts: intel: agilex5: Add Altera compatible for I3C controllers
arm64: dts: socfpga: Add Agilex5 SVC node with memory region
dt-bindings: firmware: svc: Add IOMMU support for Agilex5
arm64: dts: socfpga: agilex5: Add SMMU-V3-PMCG nodes
arm64: dts: socfpga: agilex5: Add L2 and L3 cache
arm64: dts: socfpga: agilex5: fix CHECK_DTBS warning for NAND
arm64: dts: socfpga: agilex5: add support for 013b board
dt-bindings: intel: Add Agilex5 SoCFPGA 013b board
arm64: dts: socfpga: agilex5: add VGIC maintenance interrupt
arm64: dts: socfpga: agilex: fix dtbs_check warning for NAND
arm64: dts: socfpga: agilex: fix dtbs_check warning for clock manager
arm64: dts: socfpga: stratix10-swvp: fix dtbs_check warnings swvp
arm64: dts: socfpga: move sdmmc-ecc to the base DTSI file
ARM: dts: socfpga: add Enclustra SoM dts files
dt-bindings: altera: removal of generic PE1 dts
ARM: dts: socfpga: removal of generic PE1 dts
dt-bindings: altera: add Mercury AA1 variants
...
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
|
