In komeda_plane_add(), komeda_get_layer_fourcc_list() is assigned to
formats and used in drm_universal_plane_init().
drm_universal_plane_init() passes formats to
__drm_universal_plane_init(). __drm_universal_plane_init() further
passes formats to memcpy() as src parameter, which could lead to an
undefined behavior bug on failure of komeda_get_layer_fourcc_list().
Fix this bug by adding a check of formats.
This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.
Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.
Builds with CONFIG_DRM_KOMEDA=m show no new warnings,
and our static analyzer no longer warns about this code.
There's no real reason not to send the SSID to userspace
when it requests information about P2P_GO, it is, in that
respect, exactly the same as AP interfaces. Fix that.
Similar to the previous patch, for priority changes
requested by the local PM.
Reported-and-suggested-by: Davide Caratti <dcaratti@redhat.com> Fixes: 641df33b6ffc ("mptcp: add the outgoing MP_PRIO support") Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fix an issue with commit 66255dfe9d6c ("x86/PCI: Add support for the ALi
M1487 (IBC) PIRQ router") and correct ALi M1487 (IBC) PIRQ router link
value (`pirq' cookie) interpretation according to findings in the BIOS.
Credit to Nikolai Zhubr for the detective work as to the bit layout.
Fixes: 66255dfe9d6c ("x86/PCI: Add support for the ALi M1487 (IBC) PIRQ router") Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Link: https://lore.kernel.org/r/alpine.DEB.2.21.2203310013270.44113@angie.orcam.me.uk Signed-off-by: Sasha Levin <sashal@kernel.org>
During BPF static linking, all the ELF relocations and .BTF.ext
information (including CO-RE relocations) are preserved for __weak
subprograms that were logically overriden by either previous weak
subprogram instance or by corresponding "strong" (non-weak) subprogram.
This is just how native user-space linkers work, nothing new.
But libbpf is over-zealous when processing CO-RE relocation to error out
when CO-RE relocation belonging to such eliminated weak subprogram is
encountered. Instead of erroring out on this expected situation, log
debug-level message and skip the relocation.
Fixes: 1dbf03242c2c ("libbpf: Support CO-RE relocations for multi-prog sections") Signed-off-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://lore.kernel.org/bpf/20220408181425.2287230-2-andrii@kernel.org Signed-off-by: Sasha Levin <sashal@kernel.org>
If we use a format that has padding instead of the alpha component (such
as XRGB8888), it appears that the Transposer will fill the padding to 0,
disregarding what was stored in the input buffer padding.
This leads to issues with IGT, since it will set the padding to 0xff,
but will then compare the CRC of the two frames which will thus fail.
Another nice side effect is that it is now possible to just use the
buffer as ARGB.
Fixes: 2040cc595849 ("drm/vc4: Add support for the transposer block") Signed-off-by: Maxime Ripard <maxime@cerno.tech> Acked-by: Thomas Zimmermann <tzimmermann@suse.de> Link: https://lore.kernel.org/r/20220328153659.2382206-4-maxime@cerno.tech Signed-off-by: Sasha Levin <sashal@kernel.org>
The TXP_VSTART_AT_EOF will generate a second VSTART signal to the HVS.
However, the HVS waits for VSTART to enable the FIFO and will thus start
filling the FIFO before the start of the frame.
This leads to corruption at the beginning of the first frame, and
content from the previous frame at the beginning of the next frames.
Since one VSTART is enough, let's get rid of it.
Fixes: 2040cc595849 ("drm/vc4: Add support for the transposer block") Signed-off-by: Maxime Ripard <maxime@cerno.tech> Acked-by: Thomas Zimmermann <tzimmermann@suse.de> Link: https://lore.kernel.org/r/20220328153659.2382206-3-maxime@cerno.tech Signed-off-by: Sasha Levin <sashal@kernel.org>
By default, the HVS driver will force the HVS output 3 to be muxed to
the HVS channel 2. However, the Transposer can only be assigned to the
HVS channel 2, so whenever we try to use the writeback connector, we'll
mux its associated output (Output 2) to the channel 2.
This leads to both the output 2 and 3 feeding from the same channel,
which is explicitly discouraged in the documentation.
In order to avoid this, let's reset all the output muxes to their reset
value.
Fixes: 51c4f5a96dda ("drm/vc4: crtc: Assign output to channel automatically") Signed-off-by: Maxime Ripard <maxime@cerno.tech> Acked-by: Thomas Zimmermann <tzimmermann@suse.de> Link: https://lore.kernel.org/r/20220328153659.2382206-2-maxime@cerno.tech Signed-off-by: Sasha Levin <sashal@kernel.org>
In current implementation, mtk_cec_mask() writes val into target register
and ignores the mask. After talking to our hdmi experts, mtk_cec_mask()
should read a register, clean only mask bits, and update (val | mask) bits
to the register.
Link: https://patchwork.kernel.org/project/linux-mediatek/patch/20220315232301.2434-1-miles.chen@mediatek.com/ Fixes: 98b4e84d5174 ("drm/mediatek: Add HDMI support") Signed-off-by: Miles Chen <miles.chen@mediatek.com> Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com> Cc: Zhiqiang Lin <zhiqiang.lin@mediatek.com> Cc: CK Hu <ck.hu@mediatek.com> Cc: Matthias Brugger <matthias.bgg@gmail.com> Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com> Signed-off-by: Chun-Kuang Hu <chunkuang.hu@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In order to get the field currently being output, the driver has been
using the display FIFO frame count in the HVS, reading a 6-bit field at
the offset 12 in the DISPSTATx register.
While that field is indeed at that location for the FIFO 1 and 2, the
one for the FIFO0 is actually in the DISPSTAT1 register, at the offset
18.
Fixes: 4ec28cb9da73 ("drm/vc4: Enable precise vblank timestamping for interlaced modes.") Signed-off-by: Maxime Ripard <maxime@cerno.tech> Acked-by: Thomas Zimmermann <tzimmermann@suse.de> Link: https://lore.kernel.org/r/20220331143744.777652-3-maxime@cerno.tech Signed-off-by: Sasha Levin <sashal@kernel.org>
The asm constraint does not reflect the fact that the asm statement can
modify the value of the local variable loops. Which it does.
Specifying the wrong constraint may lead to undefined behavior, it may
clobber random stuff (e.g. local variable, important temporary value in
regs, etc.). This is especially dangerous when the compiler decides to
inline the function and since it doesn't know that the value gets
modified, it might decide to use it from a register directly without
reloading it.
Change the constraint to "+a" to denote that the first argument is an
input and an output argument.
The HFP_HSW_HBP_HI register must be programmed with 2 LSbits of each
Horizontal Front Porch/Sync/Back Porch. Currently the driver programs
this register to 0, which breaks displays with either value above 255.
The HFP_MIN register must be set to the same value as HFP_LI, otherwise
there is visible image distortion, usually in the form of missing lines
at the bottom of the panel.
Fix this by correctly programming the HFP_HSW_HBP_HI and HFP_MIN registers.
Acked-by: Maxime Ripard <maxime@cerno.tech> Fixes: 67fcaad720f44 ("drm: bridge: Add Chipone ICN6211 MIPI-DSI to RGB bridge") Signed-off-by: Marek Vasut <marex@denx.de> Cc: Jagan Teki <jagan@amarulasolutions.com> Cc: Maxime Ripard <maxime@cerno.tech> Cc: Robert Foss <robert.foss@linaro.org> Cc: Sam Ravnborg <sam@ravnborg.org> Cc: Thomas Zimmermann <tzimmermann@suse.de>
To: dri-devel@lists.freedesktop.org Signed-off-by: Robert Foss <robert.foss@linaro.org> Link: https://patchwork.freedesktop.org/patch/msgid/20220331150509.9838-3-marex@denx.de Signed-off-by: Sasha Levin <sashal@kernel.org>
The chip register layout has nothing to do with MIPI DCS, the registers
incorrectly marked as MIPI DCS in the driver are regular chip registers
often with completely different function.
Fill in the actual register names and bits from [1] and [2] and add the
entire register layout, since the documentation for this chip is hard to
come by.
The invalid EDID block filtering uses the number of valid EDID
extensions instead of all EDID extensions for looping the extensions in
the copy. This is fine, by coincidence, if all the invalid blocks are at
the end of the EDID. However, it's completely broken if there are
invalid extensions in the middle; the invalid blocks are included and
valid blocks are excluded.
Fix it by modifying the base block after, not before, the copy.
Fixes: 68b5a7812577 ("drm/edid: Only print the bad edid when aborting") Reported-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Jani Nikula <jani.nikula@intel.com> Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220330170426.349248-1-jani.nikula@intel.com Signed-off-by: Sasha Levin <sashal@kernel.org>
ath11k_peer_find_by_addr states via lockdep that ab->base_lock must be
held when calling that function in order to protect the list. All
callers except ath11k_mac_op_unassign_vif_chanctx have that lock
acquired when calling ath11k_peer_find_by_addr. That lock is also not
transitively held by a path towards ath11k_mac_op_unassign_vif_chanctx.
The solution is to acquire the lock when calling
ath11k_peer_find_by_addr inside ath11k_mac_op_unassign_vif_chanctx.
I am currently working on a static analyser to detect missing locks and
this was a reported case. I manually verified the report by looking at
the code, but I do not have real hardware so this is compile tested
only.
Fixes: cf1ce93ed628 ("ath11k: add packet log support for QCA6390") Signed-off-by: Niels Dossche <dossche.niels@gmail.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://lore.kernel.org/r/20220314215253.92658-1-dossche.niels@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
vmw_move assumed that buffers to be moved would always be
vmw_buffer_object's but after introduction of new placement for mob
pages that's no longer the case.
The resulting invalid read didn't have any practical consequences
because the memory isn't used unless the object actually is a
vmw_buffer_object.
Fix it by moving the cast to the spot where the results are used.
Signed-off-by: Zack Rusin <zackr@vmware.com> Fixes: 816d1aa08b0b ("drm/vmwgfx: Introduce a new placement for MOB page tables") Reported-by: Chuck Lever III <chuck.lever@oracle.com> Reviewed-by: Martin Krastev <krastevm@vmware.com> Tested-by: Chuck Lever <chuck.lever@oracle.com> Link: https://patchwork.freedesktop.org/patch/msgid/20220318174332.440068-2-zack@kde.org Signed-off-by: Sasha Levin <sashal@kernel.org>
The backlight property was lost during conversion to yaml in commit 66de6f372847 ("dt-bindings: display: sitronix,st7735r: Convert to DT schema").
Put it back.
Hotplug events reported by bridge drivers over drm_bridge_hpd_notify()
get ignored unless somebody calls drm_bridge_hpd_enable(). When the
connector for the bridge is bridge_connector, such a call is done from
drm_bridge_connector_enable_hpd().
However drm_bridge_connector_enable_hpd() is never called on init paths,
documentation suggests that it is intended for suspend/resume paths.
In result, once encoders are switched to bridge_connector,
bridge-detected HPD stops working.
This patch adds a call to that API on init path.
This fixes HDMI HPD with rcar-du + adv7513 case when adv7513 reports HPD
events via interrupts.
When building the kernel for arm with the "-mabi=apcs-gnu" option, gcc
will force alignment of all structures and unions to a word boundary
(see also STRUCTURE_SIZE_BOUNDARY and the "-mstructure-size-boundary=XX"
option if you're a gcc person), even when the members of said structures
do not want or need said alignment.
This completely messes up the structure alignment of 'struct edid' on
those targets, because even though all the embedded structures are
marked with "__attribute__((packed))", the unions that contain them are
not.
This was exposed by commit 0a0798bc5009 ("drm/edid: add EDID block count
and size helpers"), but the bug is pre-existing. That commit just made
the structure layout problem cause a build failure due to the addition
of the
BUILD_BUG_ON(sizeof(*edid) != EDID_LENGTH);
sanity check in drivers/gpu/drm/drm_edid.c:edid_block_data().
This legacy union alignment should probably not be used in the first
place, but we can fix the layout by adding the packed attribute to the
union entries even when each member is already packed and it shouldn't
matter in a sane build environment.
You can see this issue with a trivial test program:
gpio_keys module can either accept gpios or interrupts. The module
initializes delayed work in case of gpios only and is only used if
debounce timer is not used, so make sure cancel_delayed_work_sync()
is called only when its gpio-backed and debounce_use_hrtimer is false.
This fixes the issue seen below when the gpio_keys module is unloaded and
an interrupt pin is used instead of GPIO:
drivers/macintosh/via-pmu-event.o: In function `via_pmu_event':
via-pmu-event.c:(.text+0x44): undefined reference to `input_event'
via-pmu-event.c:(.text+0x68): undefined reference to `input_event'
via-pmu-event.c:(.text+0x94): undefined reference to `input_event'
via-pmu-event.c:(.text+0xb8): undefined reference to `input_event'
drivers/macintosh/via-pmu-event.o: In function `via_pmu_event_init':
via-pmu-event.c:(.init.text+0x20): undefined reference to `input_allocate_device'
via-pmu-event.c:(.init.text+0xc4): undefined reference to `input_register_device'
via-pmu-event.c:(.init.text+0xd4): undefined reference to `input_free_device'
make[1]: *** [Makefile:1155: vmlinux] Error 1
make: *** [Makefile:350: __build_one_by_one] Error 2
Don't call into the input subsystem unless CONFIG_INPUT is built-in.
Reported-by: kernel test robot <lkp@intel.com> Signed-off-by: Finn Thain <fthain@linux-m68k.org> Tested-by: Randy Dunlap <rdunlap@infradead.org> Reviewed-by: Christophe Leroy <christophe.leroy@csgroup.eu> Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/5edbe76ce68227f71e09af4614cc4c1bd61c7ec8.1649326292.git.fthain@linux-m68k.org Signed-off-by: Sasha Levin <sashal@kernel.org>
In init_winctx_regs(), __pa() is called on winctx->rx_fifo and this
function is called to initialize registers for receive and fault
windows. But the real address is passed in winctx->rx_fifo for
receive windows and the virtual address for fault windows which
causes errors with DEBUG_VIRTUAL enabled. Fixes this issue by
assigning only real address to rx_fifo in vas_rx_win_attr struct
for both receive and fault windows.
Due to a typo, the final argument to alloc_page_vma() didn't refer to a
real variable. This only affected CONFIG_NUMA, which was marked BROKEN in
2006 and removed from alpha in 2021. Found due to a refactoring patch.
Link: https://lkml.kernel.org/r/20220504182857.4013401-4-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> Reported-by: kernel test robot <lkp@intel.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The L1 should not be able to adjust LPES mode for the L2. Setting LPES
if the L0 needs it clear would cause external interrupts to be sent to
L2 and missed by the L0.
Clearing LPES when it may be set, as typically happens with XIVE enabled
could cause a performance issue despite having no native XIVE support in
the guest, because it will cause mediated interrupts for the L2 to be
taken in HV mode, which then have to be injected.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Reviewed-by: Fabiano Rosas <farosas@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220303053315.1056880-7-npiggin@gmail.com Signed-off-by: Sasha Levin <sashal@kernel.org>
The sysfs sriov_numvfs_store() path acquires the device lock before the
config space access lock:
sriov_numvfs_store
device_lock # A (1) acquire device lock
sriov_configure
vfio_pci_sriov_configure # (for example)
vfio_pci_core_sriov_configure
pci_disable_sriov
sriov_disable
pci_cfg_access_lock
pci_wait_cfg # B (4) wait for dev->block_cfg_access == 0
Previously, pci_dev_lock() acquired the config space access lock before the
device lock:
pci_dev_lock
pci_cfg_access_lock
dev->block_cfg_access = 1 # B (2) set dev->block_cfg_access = 1
device_lock # A (3) wait for device lock
Any path that uses pci_dev_lock(), e.g., pci_reset_function(), may
deadlock with sriov_numvfs_store() if the operations occur in the sequence
(1) (2) (3) (4).
Avoid the deadlock by reversing the order in pci_dev_lock() so it acquires
the device lock before the config space access lock, the same as the
sriov_numvfs_store() path.
[bhelgaas: combined and adapted commit log from Jay Zhou's independent
subsequent posting:
https://lore.kernel.org/r/20220404062539.1710-1-jianjay.zhou@huawei.com] Link: https://lore.kernel.org/linux-pci/1583489997-17156-1-git-send-email-yangyicong@hisilicon.com/ Also-posted-by: Jay Zhou <jianjay.zhou@huawei.com> Signed-off-by: Yicong Yang <yangyicong@hisilicon.com> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
This happens because MSR[RI] is unset when entering RTAS but there is no
valid reason to not set it here.
RTAS is expected to be called with MSR[RI] as specified in PAPR+ section
"7.2.1 Machine State":
R1–7.2.1–9. If called with MSR[RI] equal to 1, then RTAS must protect
its own critical regions from recursion by setting the MSR[RI] bit to
0 when in the critical regions.
Fixing this by reviewing the way MSR is compute before calling RTAS. Now a
hardcoded value meaning real mode, 32 bits big endian mode and Recoverable
Interrupt is loaded. In the case MSR[S] is set, it will remain set while
entering RTAS as only urfid can unset it (thanks Fabiano).
In addition a check is added in do_enter_rtas() to detect calls made with
MSR[RI] unset, as we are forcing it on later.
This patch has been tested on the following machines:
Power KVM Guest
P8 S822L (host Ubuntu kernel 5.11.0-49-generic)
PowerVM LPAR
P8 9119-MME (FW860.A1)
p9 9008-22L (FW950.00)
P10 9080-HEX (FW1010.00)
Suggested-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Laurent Dufour <ldufour@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20220504101244.12107-1-ldufour@linux.ibm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
For some platforms, the frequency returned by hardware may be slightly
different from what is provided in the frequency table. For example,
hardware may return 499 MHz instead of 500 MHz. In such cases it is
better to avoid getting into unnecessary frequency updates, as we may
end up switching policy->cur between the two and sending unnecessary
pre/post update notifications, etc.
This patch has chosen allows the hardware frequency and table frequency
to deviate by 1 MHz for now, we may want to increase it a bit later on
if someone still complains.
It will cause null-ptr-deref when using 'res', if platform_get_resource()
returns NULL, so move using 'res' after devm_ioremap_resource() that
will check it to avoid null-ptr-deref.
And use devm_platform_get_and_ioremap_resource() to simplify code.
Dump capture would fail if capture kernel is not of the endianess as the
production kernel, because the in-memory data structure (struct
opal_fadump_mem_struct) shared across production kernel and capture
kernel assumes the same endianess for both the kernels, which doesn't
have to be true always. Fix it by having a well-defined endianess for
struct opal_fadump_mem_struct.
Since its introduction to the mainline kernel, omap1_uart_recalc() helper
makes incorrect use of clk->enable_bit as a ready to use bitmap mask while
it only provides the bit number. Fix it.
Signed-off-by: Janusz Krzysztofik <jmkrzyszt@gmail.com> Acked-by: Tony Lindgren <tony@atomide.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
For making easier to test, add the new quirk_flags bits 17 and 18 to
enable and disable the generic implicit feedback mode. The bit 17 is
equivalent with implicit_fb=1 option, applying the generic implicit
feedback sync mode. OTOH, the bit 18 disables the implicit fb mode
forcibly.
In our fault-injection testing, the variable "nblocks" in dbFree() can be
zero when kmalloc_array() fails in dtSearch(). In this case, the variable
"mp" in dbFree() would be NULL and then it is dereferenced in
"write_metapage(mp)".
The allocation funciton devm_kcalloc may fail and return a null pointer,
which would cause a null-pointer dereference later.
It might be better to check it and directly return -ENOMEM just like the
usage of devm_kcalloc in previous code.
Use a fine grained specification of DMA mapping directions
in certain cases, allowing both a more optimized operation
as well as shushing out a harmless, though persky
dma-debug warning.
dtschema expects DMA channels in specific order (tx, rx and tx-sec).
The order actually should not matter because dma-names is used however
let's make it aligned with dtschema to suppress warnings like:
i2s@eee30000: dma-names: ['rx', 'tx', 'tx-sec'] is not valid under any of the given schemas
The documentation of the function rvt_error_qp says both r_lock and
s_lock need to be held when calling that function.
It also asserts using lockdep that both of those locks are held.
rvt_error_qp is called form rvt_send_cq, which is called from
rvt_qp_complete_swqe, which is called from rvt_send_complete, which is
called from rvt_ruc_loopback in two places. Both of these places do not
hold r_lock. Fix this by acquiring a spin_lock of r_lock in both of
these places.
The r_lock acquiring cannot be added in rvt_qp_complete_swqe because
some of its other callers already have r_lock acquired.
In tcmu_blocks_release(), lock_page() is called to prevent a race causing
possible data corruption. Since lock_page() might sleep, calling it while
holding XArray lock is a bug.
To fix this, replace the xas_for_each() call with xa_for_each_range().
Since the latter does its own handling of XArray locking, the xas_lock()
and xas_unlock() calls around the original loop are no longer necessary.
The switch to xa_for_each_range() slows down the loop slightly. This is
acceptable since tcmu_blocks_release() is not relevant for performance.
Link: https://lore.kernel.org/r/20220517192913.21405-1-bostroesser@gmail.com Fixes: d4a5f3fe3517 ("scsi: target: tcmu: Fix possible data corruption") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Bodo Stroesser <bostroesser@gmail.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Before this patch, functions gfs2_qa_get and _put used the i_rw_mutex to
prevent simultaneous access to its i_qadata. But i_rw_mutex is now used
for many other things, including iomap_begin and end, which causes a
conflict according to lockdep. We cannot just remove the lock since
simultaneous opens (gfs2_open -> gfs2_open_common -> gfs2_qa_get) can
then stomp on each others values for i_qadata.
This patch solves the conflict by using the i_lock spin_lock in the inode
to prevent simultaneous access.
Signed-off-by: Bob Peterson <rpeterso@redhat.com> Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Latest llvm-project upstream had a change of behavior
related to qualifiers on function return type ([1]).
This caused selftests btf_dump/btf_dump failure.
The following example shows what changed.
$ cat t.c
typedef const char * const (* const (* const fn_ptr_arr2_t[5])())(char * (*)(int));
struct t {
int a;
fn_ptr_arr2_t l;
};
int foo(struct t *arg) {
return arg->a;
}
To adapt the selftest to both old and new llvm, this patch removed
the intermediate const qualifier in const-to-ptr types, to make the
test succeed again.
Returning an error value in an i2c remove callback results in an error
message being emitted by the i2c core, but otherwise it doesn't make a
difference. The device goes away anyhow and the devm cleanups are
called.
As tpm_cr50_i2c_remove() emits an error message already and the
additional error message by the i2c core doesn't add any useful
information, change the return value to zero to suppress this error
message.
Note that if i2c_clientdata is NULL, there is something really fishy.
Assuming no memory corruption happened (then all bets are lost anyhow),
tpm_cr50_i2c_remove() is only called after tpm_cr50_i2c_probe() returned
successfully. So there was a tpm chip registered before and after
tpm_cr50_i2c_remove() its privdata is freed but the associated character
device isn't removed. If after that happened userspace accesses the
character device it's likely that the freed memory is accessed. For that
reason the warning message is made a bit more frightening.
i is guaranteed < tp->irq_max which in turn is either 1 or 5.
There are more loops like this one in the driver, but strangely
GCC 12 dislikes only this single one.
Silence this silliness for now.
Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
If a client's address changes, say if it is NAT'd, this can disrupt an in
progress operation. For most operations, this is not much of a problem,
but StoreData can be different as some servers modify the target file as
the data comes in, so if a store request is disrupted, the file can get
corrupted on the server.
The problem is that the server doesn't recognise packets that come after
the change of address as belonging to the original client and will bounce
them, either by sending an OUT_OF_SEQUENCE ACK to the apparent new call if
the packet number falls within the initial sequence number window of a call
or by sending an EXCEEDS_WINDOW ACK if it falls outside and then aborting
it. In both cases, firstPacket will be 1 and previousPacket will be 0 in
the ACK information.
Fix this by the following means:
(1) If a client call receives an EXCEEDS_WINDOW ACK with firstPacket as 1
and previousPacket as 0, assume this indicates that the server saw the
incoming packets from a different peer and thus as a different call.
Fail the call with error -ENETRESET.
(2) Also fail the call if a similar OUT_OF_SEQUENCE ACK occurs if the
first packet has been hard-ACK'd. If it hasn't been hard-ACK'd, the
ACK packet will cause it to get retransmitted, so the call will just
be repeated.
(3) Make afs_select_fileserver() treat -ENETRESET as a straight fail of
the operation.
(4) Prioritise the error code over things like -ECONNRESET as the server
did actually respond.
(5) Make writeback treat -ENETRESET as a retryable error and make it
redirty all the pages involved in a write so that the VM will retry.
Note that there is still a circumstance that I can't easily deal with: if
the operation is fully received and processed by the server, but the reply
is lost due to address change. There's no way to know if the op happened.
We can examine the server, but a conflicting change could have been made by
a third party - and we can't tell the difference. In such a case, a
message like:
will be logged to dmesg on the next op to touch the file and the client
will reset the inode state, including invalidating clean parts of the
pagecache.
Reported-by: Marc Dionne <marc.dionne@auristor.com> Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org Link: http://lists.infradead.org/pipermail/linux-afs/2021-December/004811.html Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The RX_USER_ABORT code should really only be used to indicate that the user
of the rxrpc service (ie. userspace) implicitly caused a call to be aborted
- for instance if the AF_RXRPC socket is closed whilst the call was in
progress. (The user may also explicitly abort a call and specify the abort
code to use).
Change some of the points of generation to use other abort codes instead:
(1) Abort the call with RXGEN_SS_UNMARSHAL or RXGEN_CC_UNMARSHAL if we see
ENOMEM and EFAULT during received data delivery and abort with
RX_CALL_DEAD in the default case.
(2) Abort with RXGEN_SS_MARSHAL if we get ENOMEM whilst trying to send a
reply.
(3) Abort with RX_CALL_DEAD if we stop hearing from the peer if we had
heard from the peer and abort with RX_CALL_TIMEOUT if we hadn't.
(4) Abort with RX_CALL_DEAD if we try to disconnect a call that's not
completed successfully or been aborted.
Reported-by: Jeffrey Altman <jaltman@auristor.com> Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
If at the end of rxrpc sendmsg() or rxrpc_kernel_send_data() the call that
was being given data was aborted remotely or otherwise failed, return an
error rather than returning the amount of data buffered for transmission.
The call (presumably) did not complete, so there's not much point
continuing with it. AF_RXRPC considers it "complete" and so will be
unwilling to do anything else with it - and won't send a notification for
it, deeming the return from sendmsg sufficient.
Not returning an error causes afs to incorrectly handle a StoreData
operation that gets interrupted by a change of address due to NAT
reconfiguration.
This doesn't normally affect most operations since their request parameters
tend to fit into a single UDP packet and afs_make_call() returns before the
server responds; StoreData is different as it involves transmission of a
lot of data.
This can be triggered on a client by doing something like:
The Auristor fileserver logs code -453 (RXGEN_SS_UNMARSHAL), but the abort
code received by kafs is -5 (RX_PROTOCOL_ERROR) as the rx layer sees the
condition and generates an abort first and the unmarshal error is a
consequence of that at the application layer.
Reported-by: Marc Dionne <marc.dionne@auristor.com> Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org Link: http://lists.infradead.org/pipermail/linux-afs/2021-December/004810.html Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
The macros implementing Atari ROM port I/O writes do not cast away their
output, unlike similar implementations for other I/O buses.
When they are combined using conditional expressions in the definitions of
outb() and friends, this triggers sparse warnings like:
drivers/net/appletalk/cops.c:382:17: error: incompatible types in conditional expression (different base types):
drivers/net/appletalk/cops.c:382:17: unsigned char
drivers/net/appletalk/cops.c:382:17: void
If we program an RX endpoint to have no header (header length is 0),
header-related endpoint configuration values are meaningless and are
ignored.
The only case we support that defines a header is QMAP endpoints.
In ipa_endpoint_init_hdr_ext() we set the endianness mask value
unconditionally, but it should not be done if there is no header
(meaning it is not configured for QMAP).
Set the endianness conditionally, and rearrange the logic in that
function slightly to avoid testing the qmap flag twice.
Delete an incorrect comment in ipa_endpoint_init_aggr().
Signed-off-by: Alex Elder <elder@linaro.org> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Starting from LLVM 14, having an unpacked struct nested in a packed
struct triggers a warning. c.f. [1].
This is a false positive because the field is always being accessed
with the relevant put_unaligned_*() function. Adding __packed to the
structure declaration silences the warning.
In current implementation we set the non-mdts limits by calling
nvme_init_non_mdts_limits() from nvme_init_ctrl_finish().
This also tries to set the limits for the discovery controller which
has no I/O queues resulting in the warning message reported by the
nvme_log_error() when running blktest nvme/002: -
[ 2005.155946] run blktests nvme/002 at 2022-04-09 16:57:47
[ 2005.192223] loop: module loaded
[ 2005.196429] nvmet: adding nsid 1 to subsystem blktests-subsystem-0
[ 2005.200334] nvmet: adding nsid 1 to subsystem blktests-subsystem-1
[ 2008.958108] nvmet: adding nsid 1 to subsystem blktests-subsystem-997
[ 2008.962082] nvmet: adding nsid 1 to subsystem blktests-subsystem-998
[ 2008.966102] nvmet: adding nsid 1 to subsystem blktests-subsystem-999
[ 2008.973132] nvmet: creating discovery controller 1 for subsystem nqn.2014-08.org.nvmexpress.discovery for NQN testhostnqn.
*[ 2008.973196] nvme1: Identify(0x6), Invalid Field in Command (sct 0x0 / sc 0x2) MORE DNR*
[ 2008.974595] nvme nvme1: new ctrl: "nqn.2014-08.org.nvmexpress.discovery"
[ 2009.103248] nvme nvme1: Removing ctrl: NQN "nqn.2014-08.org.nvmexpress.discovery"
Move the call of nvme_init_non_mdts_limits() to nvme_scan_work() after
we verify that I/O queues are created since that is a converging point
for each transport where these limits are actually used.
1. FC :
nvme_fc_create_association()
...
nvme_fc_create_io_queues(ctrl);
...
nvme_start_ctrl()
nvme_scan_queue()
nvme_scan_work()
The transition_delay_us (struct cpufreq_policy) is currently defined
as:
Preferred average time interval between consecutive invocations of
the driver to set the frequency for this policy. To be set by the
scaling driver (0, which is the default, means no preference).
The transition_latency represents the amount of time necessary for a
CPU to change its frequency.
A PCCT table advertises mutliple values:
- pcc_nominal: Expected latency to process a command, in microseconds
- pcc_mpar: The maximum number of periodic requests that the subspace
channel can support, reported in commands per minute. 0 indicates no
limitation.
- pcc_mrtt: The minimum amount of time that OSPM must wait after the
completion of a command before issuing the next command,
in microseconds.
cppc_get_transition_latency() allows to get the max of them.
commit bb8ad01090d9 ("cpufreq / CPPC: Set platform specific
transition_delay_us") allows to select transition_delay_us based on
the platform, and fallbacks to cppc_get_transition_latency()
otherwise.
If _CPC objects are not using PCC channels (no PPCT table), the
transition_delay_us is set to CPUFREQ_ETERNAL, leading to really long
periods between frequency updates (~4s).
If the desired_reg, where performance requests are written, is in
SystemMemory or SystemIo ACPI address space, there is no delay
in requests. So return 0 instead of CPUFREQ_ETERNAL, leading to
transition_delay_us being set to LATENCY_MULTIPLIER us (1000 us).
This patch also adds two macros to check the address spaces.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com> Reviewed-by: Sudeep Holla <sudeep.holla@arm.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The hardware expects FrameNumWrap or long_term_frame_idx. Picture
numbers are per field, and are mostly used during the memory
management process, which is done in userland. This fixes two
ITU conformance tests:
- MR6_BT_B
- MR8_BT_B
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com> Reviewed-by: Sebastian Fricke <sebastian.fricke@collabora.com> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
smartshift apu and dgpu power boost are reported as percentage
with respect to their power limits. adjust the units of power before
calculating the percentage of boost.
Signed-off-by: Sathishkumar S <sathishkumar.sundararaju@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
smartshift apu and dgpu power boost are reported as percentage with
respect to their power limits. This value[0-100] reflects the boost
for the respective device.
Signed-off-by: Sathishkumar S <sathishkumar.sundararaju@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When userspace closes the socket before sending a disconnect
request, the following I/O requests will be blocked in
wait_for_reconnect() until dead timeout. This will cause the
following disconnect request also hung on blk_mq_quiesce_queue().
That means we have no way to disconnect a nbd device if there
are some I/O requests waiting for reconnecting until dead timeout.
It's not expected. So let's wake up the thread waiting for
reconnecting directly when a disconnect request is sent.
There is a logic error when removing rt5645 device as the function
rt5645_i2c_remove() first cancel the &rt5645->jack_detect_work and
delete the &rt5645->btn_check_timer latter. However, since the timer
handler rt5645_btn_check_callback() will re-queue the jack_detect_work,
this cleanup order is buggy.
That is, once the del_timer_sync in rt5645_i2c_remove is concurrently
run with the rt5645_btn_check_callback, the canceled jack_detect_work
will be rescheduled again, leading to possible use-after-free.
This patch fix the issue by placing the del_timer_sync function before
the cancel_delayed_work_sync.
In nvme_alloc_admin_tags, the admin_q can be set to an error (typically
-ENOMEM) if the blk_mq_init_queue call fails to set up the queue, which
is checked immediately after the call. However, when we return the error
message up the stack, to nvme_reset_work the error takes us to
nvme_remove_dead_ctrl()
nvme_dev_disable()
nvme_suspend_queue(&dev->queues[0]).
Here, we only check that the admin_q is non-NULL, rather than not
an error or NULL, and begin quiescing a queue that never existed, leading
to bad / NULL pointer dereference.
Signed-off-by: Kyle Smith <kyles@hpe.com> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com> Reviewed-by: Hannes Reinecke <hare@suse.de> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
In order to measure the boot process, the timer should be switched on as
early in boot as possible. As well, the commit defines the get_cycles
macro, like the previous patches in this series, so that generic code is
aware that it's implemented by the platform, as is done on other archs.
Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Arnd Bergmann <arnd@arndb.de> Cc: Jonas Bonn <jonas@southpole.se> Cc: Stefan Kristiansson <stefan.kristiansson@saunalahti.fi> Acked-by: Stafford Horne <shorne@gmail.com> Reported-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The commit made binding and unbinding of USB Ethernet asymmetrical:
Before, usbnet_probe() first invoked the ->bind() callback and then
register_netdev(). usbnet_disconnect() mirrored that by first invoking
unregister_netdev() and then ->unbind().
Since the commit, the order in usbnet_disconnect() is reversed and no
longer mirrors usbnet_probe().
One consequence is that a PHY disconnected (and stopped) in ->unbind()
is afterwards stopped once more by unregister_netdev() as it closes the
netdev before unregistering. That necessitates a contortion in ->stop()
because the PHY may only be stopped if it hasn't already been
disconnected.
Reverting the commit allows making the call to phy_stop() unconditional
in ->stop().
Tested-by: Oleksij Rempel <o.rempel@pengutronix.de> # LAN9514/9512/9500 Tested-by: Ferry Toth <fntoth@gmail.com> # LAN9514 Signed-off-by: Lukas Wunner <lukas@wunner.de> Acked-by: Oliver Neukum <oneukum@suse.com> Cc: Martyn Welch <martyn.welch@collabora.com> Cc: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
If an adapter is trying to claim a free logical address then it is
in the 'is_configuring' state. If during that process the cable is
disconnected (HPD goes low, which in turn invalidates the physical
address), then cec_adap_unconfigure() is called, and that set the
is_configuring boolean to false, even though the thread that's
trying to claim an LA is still running.
Don't touch the is_configuring bool in cec_adap_unconfigure(), it
will eventually be cleared by the thread. By making that change
the cec_config_log_addr() function also had to change: it was
aborting if is_configuring became false (since that is what
cec_adap_unconfigure() did), but that no longer works. Instead
check if the physical address is invalid. That is a much
more appropriate check anyway.
This fixes a bug where the the adapter could be disabled even
though the device was still configuring. This could cause POLL
transmits to time out.
Since usb_register_dev() from imon_init_display() from imon_probe() holds
minor_rwsem while display_open() which holds driver_lock and ictx->lock is
called with minor_rwsem held from usb_open(), holding driver_lock or
ictx->lock when calling usb_register_dev() causes circular locking
dependency problem.
Since usb_deregister_dev() from imon_disconnect() holds minor_rwsem while
display_open() which holds driver_lock is called with minor_rwsem held,
holding driver_lock when calling usb_deregister_dev() also causes circular
locking dependency problem.
Sean Young explained that the problem is there are imon devices which have
two usb interfaces, even though it is one device. The probe and disconnect
function of both usb interfaces can run concurrently.
Alan Stern responded that the driver and USB cores guarantee that when an
interface is probed, both the interface and its USB device are locked.
Ditto for when the disconnect callback gets run. So concurrent probing/
disconnection of multiple interfaces on the same device is not possible.
Therefore, we don't need locks for handling race between imon_probe() and
imon_disconnect(). But we still need to handle race between display_open()
/vfd_write()/lcd_write()/display_close() and imon_disconnect(), for
disconnect event can happen while file descriptors are in use.
Since "struct file"->private_data is set by display_open(), vfd_write()/
lcd_write()/display_close() can assume that "struct file"->private_data
is not NULL even after usb_set_intfdata(interface, NULL) was called.
Replace insufficiently held driver_lock with refcount_t based management.
Add a boolean flag for recording whether imon_disconnect() was already
called. Use RCU for accessing this boolean flag and refcount_t.
Since the boolean flag for imon_disconnect() is shared, disconnect event
on either intf0 or intf1 affects both interfaces. But I assume that this
change does not matter, for usually disconnect event would not happen
while interfaces are in use.
Link: https://syzkaller.appspot.com/bug?extid=c558267ad910fc494497 Reported-by: syzbot <syzbot+c558267ad910fc494497@syzkaller.appspotmail.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Tested-by: syzbot <syzbot+c558267ad910fc494497@syzkaller.appspotmail.com> Cc: Alan Stern <stern@rowland.harvard.edu> Signed-off-by: Sean Young <sean@mess.org> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Always set pps_cb_qp_offset and pps_cr_qp_offset values in Hantro/G2
register whatever is V4L2_HEVC_PPS_FLAG_PPS_SLICE_CHROMA_QP_OFFSETS_PRESENT
flag value.
The vendor code does the same to set these values.
This fixes conformance test CAINIT_G_SHARP_3.
Fluster HEVC score is increase by one with this patch.
Let VIDIOC_ENUM_FRAMEINTERVALS return -EINVAL if userspace queries
frame intervals for frame sizes unsupported by the encoder. Fixes the
following v4l2-compliance failure:
fail: v4l2-test-formats.cpp(123): found frame intervals for invalid size 47x16
fail: v4l2-test-formats.cpp(282): node->codec_mask & STATEFUL_ENCODER
test VIDIOC_ENUM_FMT/FRAMESIZES/FRAMEINTERVALS: FAIL
[hverkuil: drop incorrect 'For decoder devices, return -ENOTTY.' in the commit log]
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The driver needs to check if the format is 802.2 vs 802.3 in order to set
a tx descriptor flag. skb->protocol can't be used, since it may not be properly
initialized for packets coming in from a packet socket.
Fix misdetection by checking the ethertype from the skb data instead
Even though it's not possible to get into the SSIF_GETTING_MESSAGES and
SSIF_GETTING_EVENTS states without a valid message in the msg field,
it's probably best to be defensive here and check and print a log, since
that means something else went wrong.
Also add a default clause to that switch statement to release the lock
and print a log, in case the state variable gets messed up somehow.
ASUS B1400CEAE fails to resume from suspend to idle by default. This was
bisected back to commit 5a915aa99ecb ("nvme-pci: add support for ACPI
StorageD3Enable property") but this is a red herring to the problem.
Before this commit the system wasn't getting into deepest sleep state.
Presumably this commit is allowing entry into deepest sleep state as
advertised by firmware, but there are some other problems related to
the wakeup.
As it is confirmed the system works properly with S3, set the default for
this system to S3.
Reported-by: Jian-Hong Pan <jhp@endlessos.org> Link: https://bugzilla.kernel.org/show_bug.cgi?id=215742 Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Tested-by: Jian-Hong Pan <jhp@endlessos.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
We observed the error "cacheline tracking ENOMEM, dma-debug disabled"
during a light system load (copying some files). The reason for this error
is that the dma_active_cacheline radix tree uses GFP_NOWAIT allocation -
so it can't access the emergency memory reserves and it fails as soon as
anybody reaches the watermark.
This patch changes GFP_NOWAIT to GFP_ATOMIC, so that it can access the
emergency memory reserves.
projects / kernel.git / log