Commit graph

1,403,262 commits

Author SHA1 Message Date
Thomas Zimmermann
6af7e16c4d drm/ast: astdp: Look up mode index from table
Replace the large switch statement with a look-up table when selecting
the mode index. Makes the code easier to read. The table is sorted by
resolutions; if run-time overhead from traversal becomes significant,
binary search would be a possible optimization.

The mode index requires a refresh-rate index to be added or subtracted,
which still requires a minimal switch. In the original code, some of
the indices did not contain this computation. Those cases would have been
equivalent to adding 0, so they are now all subsumed in the switch's
default branch.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Suggested-by: Jocelyn Falempe <jfalempe@redhat.com>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>

v3:
- explain the semantics of the new switch statement (Jocelyn)
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-15-tzimmermann@suse.de
2025-02-03 14:01:15 +01:00
Thomas Zimmermann
d86f6beed0 drm/ast: astdp: Rework display-mode setting
ASTDP requires a mode index, depending on the resolution. Move the
look-up code from ast_dp_set_mode() into a separate helper. Inline
the rest of the function into its only caller. Rename the variable
names and register constants to match the programming manual.

As before, the mode-index lookup still happens during the update's
atomic commit. Right now, there's no way of doing it during the atomic
check. The lookup requires the VBIOS mode, which is not available at
the atomic check's invocation. At least warn now if the mode index
could not be found.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-14-tzimmermann@suse.de
2025-02-03 14:01:14 +01:00
Thomas Zimmermann
c337d4c1de drm/ast: Inline ast_get_vbios_mode_info()
The helper ast_get_vbios_mode_info() retrieves hardware-specific format
and display-mode information. Inline the function into its only caller.

While at it, also replace a use of struct drm_format_info.cpp with the
correct DRM 4CC codes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-13-tzimmermann@suse.de
2025-02-03 14:01:14 +01:00
Thomas Zimmermann
2e90349bf4 drm/ast: Validate DRM display modes against VBIOS modes
Test DRM display modes against the list of modes supported by the
VBIOS. The helper will respect the supported-modes flags in struct
ast_device. Hence only DRM display modes supported by the VBIOS
will be reported; without the current duplication of this information.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-12-tzimmermann@suse.de
2025-02-03 14:01:13 +01:00
Thomas Zimmermann
517e28d40c drm/ast: Add helpers for VBIOS mode lookup
Mode lines are independent from hardware Gen or TX chip, so hide all
VBIOS mode tables in ast_vbios.c.

Move the look-up code for VBIOS modes from ast_vbios_get_mode_info()
to ast_vbios_find_mode(). The new look-up function respects the
supported-mode flags in struct ast_device. For example, if a device
does not have struct ast_device.support_fullhd set, the helper does
not return a valid mode for 1920x1080. Taking the supported-mode flags
into account allows for making the VBIOS tables the single reference
for validating and setting display modes against hardware capabilities.

v2:
- replace mode switch with look-up table (Jocelyn)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-11-tzimmermann@suse.de
2025-02-03 14:01:12 +01:00
Thomas Zimmermann
192cc9f2b7 drm/ast: Add empty initializer for VBIOS modes
VBIOS mode tables are terminated by an entry with a refresh rate of
0xff. The code is hard to read and fragile to use. Therefore create
an empty entry with AST_VBIOS_MODE_INVALID to terminate each mode list.

Stop at the invalid entry when searching for modes in the tables.
Instead of testing for refresh == 0xff, test with a helper function if
the mode's size and refresh have meaningful values.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-10-tzimmermann@suse.de
2025-02-03 14:01:11 +01:00
Thomas Zimmermann
c81202906b drm/ast: Find VBIOS mode from regular display size
The ast driver looks up supplied display modes from an internal list of
display modes supported by the VBIOS.

Do not use the crtc_-prefixed display values from struct drm_display_mode
for looking up the VBIOS mode. The fields contain raw values that the
driver programs to hardware. They are affected by display settings like
double-scan or interlace.

Instead use the regular vdisplay and hdisplay fields for lookup. As the
programmed values can now differ from the values used for lookup, set
struct drm_display_mode.crtc_vdisplay and .crtc_hdisplay from the VBIOS
mode.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-9-tzimmermann@suse.de
2025-02-03 14:01:11 +01:00
Thomas Zimmermann
255b3ff65d drm/ast: Always validate H/V sync flags
The ast driver matches DRM display modes against an internal list of
modes supported by the VBIOS. Matching H/V sync flags between modes is
preferred, but optional. If sync flags are not matching, the driver
would program the VBIOS settings to hardware and let the display handle
the difference.

DRM modes are generated from attached displays or standard mode lines.
Therefore differences to the VBIOS modes are not just cosmetical, but
signal possible incompatibility with the display hardware.

Hence make matching H/V sync flags mandatory. If the VBIOS does not
support a certain mode, we should report it as unsupported. Note that
the VBIOS mode tables all appear to refer to standard modes.

(If sync flags really make no difference to the VBIOS, the ast driver
shouldn't match them in the first place.)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-8-tzimmermann@suse.de
2025-02-03 14:01:08 +01:00
Thomas Zimmermann
454bdaccca drm/ast: Add support_wuxga flag to struct ast_device
Detect support for 1920x1200 (WUXGA) in ast_detect_widescreen(). The
flag is cleared by default. The test logic has been taken from existing
code in ast_crtc_helper_mode_valid(). The code in that function is being
replaced by the new flag.

v2:
- move shared detection code into helper (Jocelyn)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-7-tzimmermann@suse.de
2025-02-03 14:01:07 +01:00
Thomas Zimmermann
7a74caabe1 drm/ast: Add support_fullhd flag to struct ast_device
Detect support for 1920x1080 (FullHD) in ast_detect_widescreen(). The
flag is cleared by default. The test logic has been taken from existing
code in ast_crtc_helper_mode_valid(). The code in that function is being
replaced by the new flag.

For Gen3, a new branch duplicates the Gen2 logic and adds a test for
AST2200. Gen2 adds a test for AST2100.

v2:
- use fullhd flag for setting max width/height

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-6-tzimmermann@suse.de
2025-02-03 14:01:06 +01:00
Thomas Zimmermann
1bb3f70c41 drm/ast: Reorganize widescreen test around hardware Gens
Testing for support of widescreen modes mixes up various hardware
Gens. First branch by hardware Gen, then do specific tests for each
Gen. By default, widesscreen support is disabled.

Later patches will add more specific tests for each Gen.

v2:
- move shared detection code into helper (Jocelyn)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-5-tzimmermann@suse.de
2025-02-03 14:01:06 +01:00
Thomas Zimmermann
219c6a4a6f drm/ast: Rename support_wide_screen to support_wsxga_p
The flag support_wide_screen in struct ast_device selects devices that
support resolutions of at least 1680x1050 (WSXGA+). Name it accordingly.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-4-tzimmermann@suse.de
2025-02-03 14:01:05 +01:00
Thomas Zimmermann
fdebbb2cec drm/ast: Align naming in widescreen detection code to manual
Rename variables and register constants to align with the programming
manual. Add new constants where necessary.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-3-tzimmermann@suse.de
2025-02-03 14:01:04 +01:00
Thomas Zimmermann
707e860119 drm/ast: Remove 1152x864 from list of widescreen resolutions
The resolution 1152x864 has a ratio of 4:3 and is already handled
as such. Remove it from the list of widescreen resolutions.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250131092257.115596-2-tzimmermann@suse.de
2025-02-03 14:01:04 +01:00
Thomas Zimmermann
1bcfd16b12
platform/x86: thinkpad-acpi: Do not include <linux/fb.h>
Nothing in this source file requires <linux/fb.h>. Remove the include
statement.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Henrique de Moraes Holschuh <hmh@hmh.eng.br>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20250131095235.125533-3-tzimmermann@suse.de
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
2025-02-03 14:59:05 +02:00
Thomas Zimmermann
e87af94c07
platform/x86: compal-laptop: Do not include <linux/fb.h>
Nothing in this source file requires <linux/fb.h>. Remove the include
statement.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Cezary Jackiewicz <cezary.jackiewicz@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20250131095235.125533-2-tzimmermann@suse.de
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
2025-02-03 14:59:04 +02:00
Xiangrong Li
5d40a85775
mlxbf-bootctl: Support sysfs entries for RTC battery status
This patch extends the mlxbf-bootctl driver's sysfs entries
to support read access for the board's RTC battery status.
A successful read from this attribute returns the status of
the board's RTC battery. The RTC battery status register is
also cleared upon successful read operation.

Signed-off-by: Xiangrong Li <xiangrongl@nvidia.com>
Reviewed-by: David Thompson <davthompson@nvidia.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://lore.kernel.org/r/20250124144655.48564-1-xiangrongl@nvidia.com
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
2025-02-03 14:57:27 +02:00
H. Nikolaus Schaller
c9ccb88f53 Input: ads7846 - fix gpiod allocation
commit 767d83361a ("Input: ads7846 - Convert to use software nodes")

has simplified the code but accidentially converted a devm_gpiod_get()
to gpiod_get(). This leaves the gpio reserved on module remove and the
driver can no longer be loaded again.

Fixes: 767d83361a ("Input: ads7846 - Convert to use software nodes")
Cc: stable@vger.kernel.org
Signed-off-by: H. Nikolaus Schaller <hns@goldelico.com>
Link: https://lore.kernel.org/r/6e9b143f19cdfda835711a8a7a3966e5a2494cff.1738410204.git.hns@goldelico.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
2025-02-03 04:37:57 -08:00
Andy-ld Lu
3e68abf2b9 mmc: mtk-sd: Fix register settings for hs400(es) mode
For hs400(es) mode, the 'hs400-ds-delay' is typically configured in the
dts. However, some projects may only define 'mediatek,hs400-ds-dly3',
which can lead to initialization failures in hs400es mode. CMD13 reported
response crc error in the mmc_switch_status() just after switching to
hs400es mode.

[    1.914038][   T82] mmc0: mmc_select_hs400es failed, error -84
[    1.914954][   T82] mmc0: error -84 whilst initialising MMC card

Currently, the hs400_ds_dly3 value is set within the tuning function. This
means that the PAD_DS_DLY3 field is not configured before tuning process,
which is the reason for the above-mentioned CMD13 response crc error.

Move the PAD_DS_DLY3 field configuration into msdc_prepare_hs400_tuning(),
and add a value check of hs400_ds_delay to prevent overwriting by zero when
the 'hs400-ds-delay' is not set in the dts. In addition, since hs400(es)
only tune the PAD_DS_DLY1, the PAD_DS_DLY2_SEL bit should be cleared to
bypass it.

Fixes: c4ac38c653 ("mmc: mtk-sd: Add HS400 online tuning support")
Signed-off-by: Andy-ld Lu <andy-ld.lu@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20250123092644.7359-1-andy-ld.lu@mediatek.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2025-02-03 13:34:50 +01:00
Josua Mayer
ac5a41b472 Revert "mmc: sdhci_am654: Add sdhci_am654_start_signal_voltage_switch"
This reverts commit 941a7abd46.

This commit uses presence of device-tree properties vmmc-supply and
vqmmc-supply for deciding whether to enable a quirk affecting timing of
clock and data.
The intention was to address issues observed with eMMC and SD on AM62
platforms.

This new quirk is however also enabled for AM64 breaking microSD access
on the SolidRun HimmingBoard-T which is supported in-tree since v6.11,
causing a regression. During boot microSD initialization now fails with
the error below:

[    2.008520] mmc1: SDHCI controller on fa00000.mmc [fa00000.mmc] using ADMA 64-bit
[    2.115348] mmc1: error -110 whilst initialising SD card

The heuristics for enabling the quirk are clearly not correct as they
break at least one but potentially many existing boards.

Revert the change and restore original behaviour until a more
appropriate method of selecting the quirk is derived.

Fixes: 941a7abd46 ("mmc: sdhci_am654: Add sdhci_am654_start_signal_voltage_switch")
Closes: https://lore.kernel.org/linux-mmc/a70fc9fc-186f-4165-a652-3de50733763a@solid-run.com/
Cc: stable@vger.kernel.org
Signed-off-by: Josua Mayer <josua@solid-run.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20250127-am654-mmc-regression-v2-1-9bb39fb12810@solid-run.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
2025-02-03 13:33:58 +01:00
Niklas Cassel
deca423213 ata: libata-core: Add 'external' to the libata.force kernel parameter
Commit ae1f3db006 ("ata: ahci: do not enable LPM on external ports")
changed so that LPM is not enabled on external ports (hotplug-capable or
eSATA ports).

This is because hotplug and LPM are mutually exclusive, see 7.3.1 Hot Plug
Removal Detection and Power Management Interaction in AHCI 1.3.1.

This does require that firmware has set the appropate bits (HPCP or ESP)
in PxCMD (which is a per port register in the AHCI controller).

If the firmware has failed to mark a port as hotplug-capable or eSATA in
PxCMD, then there is currently not much a user can do.

If LPM is enabled on the port, hotplug insertions and removals will not be
detected on that port.

In order to allow a user to fix up broken firmware, add 'external' to the
libata.force kernel parameter.

libata.force can be specified either on the kernel command line, or as a
kernel module parameter.

For more information, see Documentation/admin-guide/kernel-parameters.txt.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>
Link: https://lore.kernel.org/r/20250130133544.219297-4-cassel@kernel.org
Signed-off-by: Niklas Cassel <cassel@kernel.org>
2025-02-03 13:09:24 +01:00
Thomas Zimmermann
1fdadc8418 ARM: s3c: Do not include <linux/fb.h>
These ARM architecture's source files do not require <linux/fb.h>.
Remove the include statements.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://lore.kernel.org/r/20250203102818.56747-3-tzimmermann@suse.de
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
2025-02-03 12:58:24 +01:00
Krzysztof Kozlowski
00c1fda73d soc: samsung: Use syscon_regmap_lookup_by_phandle_args
Use syscon_regmap_lookup_by_phandle_args() which is a wrapper over
syscon_regmap_lookup_by_phandle() combined with getting the syscon
argument.  Except simpler code this annotates within one line that given
phandle has arguments, so grepping for code would be easier.

Link: https://lore.kernel.org/r/20250111185402.183793-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
2025-02-03 12:56:34 +01:00
Konstantin Taranov
6e1b8bdcd0 RDMA/mana_ib: implement get_dma_mr
Implement allocation of DMA-mapped memory regions.

Signed-off-by: Konstantin Taranov <kotaranov@microsoft.com>
Link: https://patch.msgid.link/1737394039-28772-3-git-send-email-kotaranov@linux.microsoft.com
Reviewed-by: Shiraz Saleem <shirazsaleem@microsoft.com>
Reviewed-by: Long Li <longli@microsoft.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:56:19 -05:00
Konstantin Taranov
78683c25c8 RDMA/mana_ib: Allow registration of DMA-mapped memory in PDs
Allow the HW to register DMA-mapped memory for kernel-level PDs.

Signed-off-by: Konstantin Taranov <kotaranov@microsoft.com>
Link: https://patch.msgid.link/1737394039-28772-2-git-send-email-kotaranov@linux.microsoft.com
Reviewed-by: Shiraz Saleem <shirazsaleem@microsoft.com>
Reviewed-by: Long Li <longli@microsoft.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:56:19 -05:00
Huisong Li
33c1452978 w1: w1_therm: w1: Use HWMON_CHANNEL_INFO macro to simplify code
Use HWMON_CHANNEL_INFO macro to simplify code.

Signed-off-by: Huisong Li <lihuisong@huawei.com>
Link: https://lore.kernel.org/r/20250124022635.16647-9-lihuisong@huawei.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
2025-02-03 12:54:18 +01:00
Chenyuan Yang
0dd6770a72 w1: fix NULL pointer dereference in probe
The w1_uart_probe() function calls w1_uart_serdev_open() (which includes
devm_serdev_device_open()) before setting the client ops via
serdev_device_set_client_ops(). This ordering can trigger a NULL pointer
dereference in the serdev controller's receive_buf handler, as it assumes
serdev->ops is valid when SERPORT_ACTIVE is set.

This is similar to the issue fixed in commit 5e700b384e
("platform/chrome: cros_ec_uart: properly fix race condition") where
devm_serdev_device_open() was called before fully initializing the
device.

Fix the race by ensuring client ops are set before enabling the port via
w1_uart_serdev_open().

Fixes: a3c0880436 ("w1: add UART w1 bus driver")
Signed-off-by: Chenyuan Yang <chenyuan0y@gmail.com>
Acked-by: Christoph Winklhofer <cj.winklhofer@gmail.com>
Link: https://lore.kernel.org/r/20250111181803.2283611-1-chenyuan0y@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
2025-02-03 12:52:49 +01:00
Krzysztof Kozlowski
2f76a890fd memory: tegra20-emc: Drop redundant platform_get_irq() error printk
platform_get_irq() already prints error message, so duplicating it is
redundant.  The message about "need of updating DT" makes no sense,
because this code was there since beginning.

Link: https://lore.kernel.org/r/20250104141430.115031-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
2025-02-03 12:52:07 +01:00
Chris Morgan
98380110bd power: supply: axp20x_battery: Fix fault handling for AXP717
Correct the fault handling for the AXP717 by changing the i2c write
from regmap_update_bits() to regmap_write_bits(). The update bits
function does not work properly on a RW1C register where we must
write a 1 back to an existing register to clear it.

Additionally, as part of this testing I confirmed the behavior of
errors reappearing, so remove comment about assumptions.

Fixes: 6625767049 ("power: supply: axp20x_battery: add support for AXP717")
Signed-off-by: Chris Morgan <macromorgan@hotmail.com>
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Link: https://lore.kernel.org/r/20250131231455.153447-2-macroalpha82@gmail.com
Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com>
2025-02-03 12:41:18 +01:00
Zhu Yanjun
190797d47f RDMA/rxe: Make rping work with tun device
Because the type of tun device is ARPHRD_NONE, in cma, ARPHRD_NONE is
not recognized. To RXE device, just as SIW does, ARPHRD_NONE is added
to support.

Signed-off-by: Zhu Yanjun <yanjun.zhu@linux.dev>
Link: https://patch.msgid.link/20250119172831.3123110-4-yanjun.zhu@linux.dev
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:38:43 -05:00
Zhu Yanjun
93486fc96f RDMA/rxe: Add query_gid support
The query_gid is not implemented in RXE. After the raw_gid is added,
this query_gid should be implemented in RXE.

Signed-off-by: Zhu Yanjun <yanjun.zhu@linux.dev>
Link: https://patch.msgid.link/20250119172831.3123110-3-yanjun.zhu@linux.dev
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:38:43 -05:00
Zhu Yanjun
d34d0bdb50 RDMA/rxe: Replace netdev dev addr with raw_gid
Because TUN device does not have dev_addr, but a gid in rdma is needed,
as such, a raw_gid is generated to act as the gid. The similar commit is
in SIW. This commit learns from the similar commit bad5b6e34f
("RDMA/siw: Fabricate a GID on tun and loopback devices") in SIW.

Signed-off-by: Zhu Yanjun <yanjun.zhu@linux.dev>
Link: https://patch.msgid.link/20250119172831.3123110-2-yanjun.zhu@linux.dev
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:38:43 -05:00
Mark Zhang
12d044770e IB/mlx5: Set and get correct qp_num for a DCT QP
When a DCT QP is created on an active lag, it's dctc.port is assigned
in a round-robin way, which is from 1 to dev->lag_port. In this case
when querying this QP, we may get qp_attr.port_num > 2.
Fix this by setting qp->port when modifying a DCT QP, and read port_num
from qp->port instead of dctc.port when querying it.

Fixes: 7c4b1ab9f1 ("IB/mlx5: Add DCT RoCE LAG support")
Signed-off-by: Mark Zhang <markzhang@nvidia.com>
Reviewed-by: Maher Sanalla <msanalla@nvidia.com>
Link: https://patch.msgid.link/94c76bf0adbea997f87ffa27674e0a7118ad92a9.1737290358.git.leon@kernel.org
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:35:55 -05:00
Yishai Hadas
d97505baea RDMA/mlx5: Fix the recovery flow of the UMR QP
This patch addresses an issue in the recovery flow of the UMR QP,
ensuring tasks do not get stuck, as highlighted by the call trace [1].

During recovery, before transitioning the QP to the RESET state, the
software must wait for all outstanding WRs to complete.

Failing to do so can cause the firmware to skip sending some flushed
CQEs with errors and simply discard them upon the RESET, as per the IB
specification.

This race condition can result in lost CQEs and tasks becoming stuck.

To resolve this, the patch sends a final WR which serves only as a
barrier before moving the QP state to RESET.

Once a CQE is received for that final WR, it guarantees that no
outstanding WRs remain, making it safe to transition the QP to RESET and
subsequently back to RTS, restoring proper functionality.

Note:
For the barrier WR, we simply reuse the failed and ready WR.
Since the QP is in an error state, it will only receive
IB_WC_WR_FLUSH_ERR. However, as it serves only as a barrier we don't
care about its status.

[1]
INFO: task rdma_resource_l:1922 blocked for more than 120 seconds.
Tainted: G        W          6.12.0-rc7+ #1626
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
task:rdma_resource_l state:D stack:0  pid:1922 tgid:1922  ppid:1369
     flags:0x00004004
Call Trace:
<TASK>
__schedule+0x420/0xd30
schedule+0x47/0x130
schedule_timeout+0x280/0x300
? mark_held_locks+0x48/0x80
? lockdep_hardirqs_on_prepare+0xe5/0x1a0
wait_for_completion+0x75/0x130
mlx5r_umr_post_send_wait+0x3c2/0x5b0 [mlx5_ib]
? __pfx_mlx5r_umr_done+0x10/0x10 [mlx5_ib]
mlx5r_umr_revoke_mr+0x93/0xc0 [mlx5_ib]
__mlx5_ib_dereg_mr+0x299/0x520 [mlx5_ib]
? _raw_spin_unlock_irq+0x24/0x40
? wait_for_completion+0xfe/0x130
? rdma_restrack_put+0x63/0xe0 [ib_core]
ib_dereg_mr_user+0x5f/0x120 [ib_core]
? lock_release+0xc6/0x280
destroy_hw_idr_uobject+0x1d/0x60 [ib_uverbs]
uverbs_destroy_uobject+0x58/0x1d0 [ib_uverbs]
uobj_destroy+0x3f/0x70 [ib_uverbs]
ib_uverbs_cmd_verbs+0x3e4/0xbb0 [ib_uverbs]
? __pfx_uverbs_destroy_def_handler+0x10/0x10 [ib_uverbs]
? __lock_acquire+0x64e/0x2080
? mark_held_locks+0x48/0x80
? find_held_lock+0x2d/0xa0
? lock_acquire+0xc1/0x2f0
? ib_uverbs_ioctl+0xcb/0x170 [ib_uverbs]
? __fget_files+0xc3/0x1b0
ib_uverbs_ioctl+0xe7/0x170 [ib_uverbs]
? ib_uverbs_ioctl+0xcb/0x170 [ib_uverbs]
__x64_sys_ioctl+0x1b0/0xa70
do_syscall_64+0x6b/0x140
entry_SYSCALL_64_after_hwframe+0x76/0x7e
RIP: 0033:0x7f99c918b17b
RSP: 002b:00007ffc766d0468 EFLAGS: 00000246 ORIG_RAX:
     0000000000000010
RAX: ffffffffffffffda RBX: 00007ffc766d0578 RCX:
     00007f99c918b17b
RDX: 00007ffc766d0560 RSI: 00000000c0181b01 RDI:
     0000000000000003
RBP: 00007ffc766d0540 R08: 00007f99c8f99010 R09:
     000000000000bd7e
R10: 00007f99c94c1c70 R11: 0000000000000246 R12:
     00007ffc766d0530
R13: 000000000000001c R14: 0000000040246a80 R15:
     0000000000000000
</TASK>

Fixes: 158e71bb69 ("RDMA/mlx5: Add a umr recovery flow")
Signed-off-by: Yishai Hadas <yishaih@nvidia.com>
Reviewed-by: Michael Guralnik <michaelgur@nvidia.com>
Link: https://patch.msgid.link/27b51b92ec42dfb09d8096fcbd51878f397ce6ec.1737290141.git.leon@kernel.org
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2025-02-03 06:35:27 -05:00
Martin KaFai Lau
03f3aa4a6b Merge branch 'selftests-bpf-migrate-test_xdp_redirect_multi-sh-to-test_progs'
Bastien Curutchet says:

====================
This patch series continues the work to migrate the *.sh tests into
prog_tests framework.

test_xdp_redirect_multi.sh tests the XDP redirections done through
bpf_redirect_map().

This is already partly covered by test_xdp_veth.c that already tests
map redirections at XDP level. What isn't covered yet by test_xdp_veth is
the use of the broadcast flags (BPF_F_BROADCAST or BPF_F_EXCLUDE_INGRESS)
and XDP egress programs.

This series is the prep work that will be followed up adding test cases
to eventually cover the tests done in test_xdp_redirect_multi.sh:
 - PATCH 1 Add an helper to generate unique names
 - PATCH 2 to 9 rework test_xdp_veth to make it more generic and allow to
   configure different test cases
 - PATCH 10 adds test cases for 'classic' bpf_redirect_map()
====================

Link: https://patch.msgid.link/20250131-redirect-multi-v4-0-970b33678512@bootlin.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:34:26 -08:00
Bastien Curutchet (eBPF Foundation)
0c4ea7e347 selftests/bpf: test_xdp_veth: Add new test cases for XDP flags
The XDP redirection is tested without any flag provided to the
xdp_attach() function.

Add two subtests that check the correct behaviour with
XDP_FLAGS_{DRV/SKB}_MODE flags

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-10-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:52 -08:00
Bastien Curutchet (eBPF Foundation)
29c7bb7d0f selftests/bpf: test_xdp_veth: Use unique names
The network namespaces and the veth used by the tests have hardcoded
names that can conflict with other tests during parallel runs.

Use the append_tid() helper to ensure the uniqueness of these names.
Use the static network configuration table as a template on which
thread IDs are appended in each test.
Set a fixed size to remote_addr field so the struct veth_configuration
can also have a fixed size.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-9-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:52 -08:00
Bastien Curutchet (eBPF Foundation)
450effe2da selftests/bpf: test_xdp_veth: Add XDP flags to prog_configuration
XDP flags are hardcoded to 0 at attachment.

Add flags attributes to the struct prog_configuration to allow flag
modifications for each test case.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-8-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:52 -08:00
Bastien Curutchet (eBPF Foundation)
edb996fae2 selftests/bpf: test_xdp_veth: Add prog_config[] table
The BPF program attached to each veth is hardcoded through the
use of the struct skeletons. It prevents from re-using the initialization
code in new test cases.

Replace the struct skeletons by a bpf_object table.
Add a struct prog_configuration that holds the name of BPF program to
load on a given veth pair.
Use bpf_object__find_program_by_name() / bpf_xdp_attach() API instead of
bpf_program__attach_xdp() to retrieve the BPF programs from their names.
Detach BPF progs in the cleanup() as it's not automatically done by this
API.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-7-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:52 -08:00
Martin KaFai Lau
12fdd29d5d bpf: Use kallsyms to find the function name of a struct_ops's stub function
In commit 1611603537 ("bpf: Create argument information for nullable arguments."),
it introduced a "__nullable" tagging at the argument name of a
stub function. Some background on the commit:
it requires to tag the stub function instead of directly tagging
the "ops" of a struct. This is because the btf func_proto of the "ops"
does not have the argument name and the "__nullable" is tagged at
the argument name.

To find the stub function of a "ops", it currently relies on a naming
convention on the stub function "st_ops__ops_name".
e.g. tcp_congestion_ops__ssthresh. However, the new kernel
sub system implementing bpf_struct_ops have missed this and
have been surprised that the "__nullable" and the to-be-landed
"__ref" tagging was not effective.

One option would be to give a warning whenever the stub function does
not follow the naming convention, regardless if it requires arg tagging
or not.

Instead, this patch uses the kallsyms_lookup approach and removes
the requirement on the naming convention. The st_ops->cfi_stubs has
all the stub function kernel addresses. kallsyms_lookup() is used to
lookup the function name. With the function name, BTF can be used to
find the BTF func_proto. The existing "__nullable" arg name searching
logic will then fall through.

One notable change is,
if it failed in kallsyms_lookup or it failed in looking up the stub
function name from the BTF, the bpf_struct_ops registration will fail.
This is different from the previous behavior that it silently ignored
the "st_ops__ops_name" function not found error.

The "tcp_congestion_ops", "sched_ext_ops", and "hid_bpf_ops" can still be
registered successfully after this patch. There is struct_ops_maybe_null
selftest to cover the "__nullable" tagging.

Other minor changes:
1. Removed the "%s__%s" format from the pr_warn because the naming
   convention is removed.
2. The existing bpf_struct_ops_supported() is also moved earlier
   because prepare_arg_info needs to use it to decide if the
   stub function is NULL before calling the prepare_arg_info.

Cc: Tejun Heo <tj@kernel.org>
Cc: Benjamin Tissoires <bentiss@kernel.org>
Cc: Yonghong Song <yonghong.song@linux.dev>
Cc: Amery Hung <ameryhung@gmail.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Reviewed-by: Amery Hung <ameryhung@gmail.com>
Link: https://lore.kernel.org/r/20250127222719.2544255-1-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Bastien Curutchet (eBPF Foundation)
7e9f3c875d selftests/bpf: test_xdp_veth: Rename config[]
The network topology is held by the config[] table. This 'config' name
is a bit too generic if we want to add other configuration variables.

Rename config[] to net_config[].

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-6-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Viktor Malik
0053f7d39d bpftool: Fix readlink usage in get_fd_type
The `readlink(path, buf, sizeof(buf))` call reads at most sizeof(buf)
bytes and *does not* append null-terminator to buf. With respect to
that, fix two pieces in get_fd_type:

1. Change the truncation check to contain sizeof(buf) rather than
   sizeof(path).
2. Append null-terminator to buf.

Reported by Coverity.

Signed-off-by: Viktor Malik <vmalik@redhat.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/20250129071857.75182-1-vmalik@redhat.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Bastien Curutchet (eBPF Foundation)
3c32cbbbcd selftests/bpf: test_xdp_veth: Split network configuration
configure_network() does two things : it first creates the network
topology and then configures the BPF maps to fit the test needs. This
isn't convenient if we want to re-use the same network topology for
different test cases.

Rename configure_network() create_network().
Move the BPF configuration to the test itself.
Split the test description in two parts, first the description of the
network topology, then the description of the test case.
Remove the veth indexes from the ASCII art as dynamic ones are used

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-5-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Tony Ambardar
cb3ade5678 selftests/bpf: Fix runqslower cross-endian build
The runqslower binary from a cross-endian build currently fails to run
because the included skeleton has host endianness. Fix this by passing the
target BPF endianness to the runqslower sub-make.

Fixes: 5a63c33d6f ("selftests/bpf: Support cross-endian building")
Signed-off-by: Tony Ambardar <tony.ambardar@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20250125071423.2603588-1-itugrok@yahoo.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Bastien Curutchet (eBPF Foundation)
71e0b1cc72 selftests/bpf: test_xdp_veth: Use int to describe next veth
In the struct veth_configuration, the next_veth string is used to tell
the next virtual interface to which packets must be redirected to. So it
has to match the local_veth string of an other veth_configuration.

Change next_veth type to int to avoid handling two identical strings.
This integer is used as an offset in the network configuration table.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-4-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Tony Ambardar
0a7c2a8435 libbpf: Fix accessing BTF.ext core_relo header
Update btf_ext_parse_info() to ensure the core_relo header is present
before reading its fields. This avoids a potential buffer read overflow
reported by the OSS Fuzz project.

Fixes: cf579164e9 ("libbpf: Support BTF.ext loading and output in either endianness")
Signed-off-by: Tony Ambardar <tony.ambardar@gmail.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://issues.oss-fuzz.com/issues/388905046
Link: https://lore.kernel.org/bpf/20250125065236.2603346-1-itugrok@yahoo.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Bastien Curutchet (eBPF Foundation)
0f5bab8dff selftests/bpf: test_xdp_veth: Remove unecessarry check_ping()
check_ping() directly returns a SYS_NOFAIL without any previous
treatment. It's called only once in the file and hardcodes the used
namespace and ip address.

Replace check_ping() with a direct call of SYS_NOFAIL in the test.

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-3-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Tengda Wu
a63a631c9b selftests/bpf: Fix freplace_link segfault in tailcalls prog test
There are two bpf_link__destroy(freplace_link) calls in
test_tailcall_bpf2bpf_freplace(). After the first bpf_link__destroy()
is called, if the following bpf_map_{update,delete}_elem() throws an
exception, it will jump to the "out" label and call bpf_link__destroy()
again, causing double free and eventually leading to a segfault.

Fix it by directly resetting freplace_link to NULL after the first
bpf_link__destroy() call.

Fixes: 021611d33e ("selftests/bpf: Add test to verify tailcall and freplace restrictions")
Signed-off-by: Tengda Wu <wutengda@huaweicloud.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Leon Hwang <leon.hwang@linux.dev>
Link: https://lore.kernel.org/bpf/20250122022838.1079157-1-wutengda@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Bastien Curutchet (eBPF Foundation)
6d34f5b728 selftests/bpf: test_xdp_veth: Remove unused defines
IP_CMD_MAX_LEN and NS_SUFFIX_LEN aren't used anywhere.

Remove these unused defines

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-2-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00
Bastien Curutchet (eBPF Foundation)
723f1b9ce3 selftests/bpf: helpers: Add append_tid()
Some tests can't be run in parallel because they use same namespace
names or veth names.

Create an helper that appends the thread ID to a given string. 8
characters are used for it (7 digits + '\0')

Signed-off-by: Bastien Curutchet (eBPF Foundation) <bastien.curutchet@bootlin.com>
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://patch.msgid.link/20250131-redirect-multi-v4-1-970b33678512@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
2025-02-03 03:33:51 -08:00