Commit graph

1,398,251 commits

Author SHA1 Message Date
Jakub Kicinski
574583c30c Merge branch 'add-noinline_for_tracing-and-apply-it-to-tcp_drop_reason'
Yafang Shao says:

====================
Add noinline_for_tracing and apply it to tcp_drop_reason

This patchset introduces a new compiler annotation, noinline_for_tracing,
designed to prevent specific functions from being inlined to facilitate
tracing. In Patch #2, this annotation is applied to the tcp_drop_reason().
====================

Link: https://patch.msgid.link/20241024093742.87681-1-laoar.shao@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 09:02:35 -08:00
Yafang Shao
dbd5e2e79e net: tcp: Add noinline_for_tracing annotation for tcp_drop_reason()
We previously hooked the tcp_drop_reason() function using BPF to monitor
TCP drop reasons. However, after upgrading our compiler from GCC 9 to GCC
11, tcp_drop_reason() is now inlined, preventing us from hooking into it.
To address this, it would be beneficial to make noinline explicitly for
tracing.

Link: https://lore.kernel.org/netdev/CANn89iJuShCmidCi_ZkYABtmscwbVjhuDta1MS5LxV_4H9tKOA@mail.gmail.com/
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Cc: Menglong Dong <menglong8.dong@gmail.com>
Link: https://patch.msgid.link/20241024093742.87681-3-laoar.shao@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 09:02:32 -08:00
Yafang Shao
a8f80673ca compiler_types: Add noinline_for_tracing annotation
Kernel functions that are not inlined can be easily hooked with BPF for
tracing. However, functions intended for tracing may still be inlined
unexpectedly. For example, in our case, after upgrading the compiler from
GCC 9 to GCC 11, the tcp_drop_reason() function was inlined, which broke
our monitoring tools. To prevent this, we need to ensure that the function
remains non-inlined.

The noinline_for_tracing annotation is introduced as a general solution for
preventing inlining of kernel functions that need to be traced. This
approach avoids the need for adding individual noinline comments to each
function and provides a more consistent way to maintain traceability.

Link: https://lore.kernel.org/netdev/CANn89iKvr44ipuRYFaPTpzwz=B_+pgA94jsggQ946mjwreV6Aw@mail.gmail.com/
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://patch.msgid.link/20241024093742.87681-2-laoar.shao@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 09:02:32 -08:00
Jakub Kicinski
f07a6e6ceb Merge branch 'dpll-expose-clock-quality-level'
Jiri Pirko says:

====================
dpll: expose clock quality level

Some device driver might know the quality of the clock it is running.
In order to expose the information to the user, introduce new netlink
attribute and dpll device op. Implement the op in mlx5 driver.

Example:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/dpll.yaml --dump device-get
[{'clock-id': 13316852727532664826,
  'clock-quality-level': ['itu-opt1-eeec'],    <<<<<<<<<<<<<<<<<
  'id': 0,
  'lock-status': 'unlocked',
  'lock-status-error': 'none',
  'mode': 'manual',
  'mode-supported': ['manual'],
  'module-name': 'mlx5_dpll',
  'type': 'eec'}]
====================

Link: https://patch.msgid.link/20241030081157.966604-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 08:39:12 -08:00
Jiri Pirko
e2017f27b6 net/mlx5: DPLL, Add clock quality level op implementation
Use MSECQ register to query clock quality from firmware. Implement the
dpll op and fill-up the quality level value properly.

Reviewed-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com>
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20241030081157.966604-3-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 08:39:07 -08:00
Jiri Pirko
a1afb959ad dpll: add clock quality level attribute and op
In order to allow driver expose quality level of the clock it is
running, introduce a new netlink attr with enum to carry it to the
userspace. Also, introduce an op the dpll netlink code calls into the
driver to obtain the value.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20241030081157.966604-2-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 08:39:07 -08:00
Dmitry Baryshkov
86313a9cd1 drm/msm/dpu: rework documentation comments
Unfortunately the tooling doesn't check documents placed before funciton
prototypes. Such comments frequently become outdated, miss several
params, etc. Move documentation for the functions to be placed before
the actual function body, allowing 'make W=1' to actually check these
comments and report an error.

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/622690/
Link: https://lore.kernel.org/r/20241102-dpu-docs-rework-v1-1-d735853fd6db@linaro.org
2024-11-03 18:21:39 +02:00
Christophe JAILLET
aceee63a3a RDMA/bnxt_re: Remove some dead code
If the probe succeeds, then auxiliary_get_drvdata() can't return a NULL
pointer.

So several NULL checks can be removed to simplify code.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://patch.msgid.link/f02eb630734ee530315dce9f60b078f631ae93d0.1730477345.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2024-11-03 11:02:02 -05:00
Christophe JAILLET
cf90a4d1b9 RDMA/bnxt_re: Fix some error handling paths in bnxt_re_probe()
If bnxt_re_add_device() fails, 'en_info' still needs to be freed, as
already done in the .remove() function.

The commit in Fixes incorrectly removed this call, certainly because it
was expecting the .remove() function was called anyway. But if the probe
fails, the remove function is not called.

There is no need to call bnxt_re_remove() as it was done before, kfree()
is enough.

Fixes: a5e099e0c4 ("RDMA/bnxt_re: Fix an error path in bnxt_re_add_device")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://patch.msgid.link/9e48ff955ae55fc39a9eb1eb590d374539eab5ba.1730477345.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Leon Romanovsky <leon@kernel.org>
2024-11-03 11:02:02 -05:00
Jinjie Ruan
3b557be89f net: wwan: t7xx: Fix off-by-one error in t7xx_dpmaif_rx_buf_alloc()
The error path in t7xx_dpmaif_rx_buf_alloc(), free and unmap the already
allocated and mapped skb in a loop, but the loop condition terminates when
the index reaches zero, which fails to free the first allocated skb at
index zero.

Check with i-- so that skb at index 0 is freed as well.

Cc: stable@vger.kernel.org
Fixes: d642b012df ("net: wwan: t7xx: Add data path interface")
Acked-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Signed-off-by: Jinjie Ruan <ruanjinjie@huawei.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Link: https://patch.msgid.link/20241101025316.3234023-1-ruanjinjie@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2024-11-03 07:59:30 -08:00
Masahiro Yamada
77dc55a978 modpost: fix input MODULE_DEVICE_TABLE() built for 64-bit on 32-bit host
When building a 64-bit kernel on a 32-bit build host, incorrect
input MODULE_ALIAS() entries may be generated.

For example, when compiling a 64-bit kernel with CONFIG_INPUT_MOUSEDEV=m
on a 64-bit build machine, you will get the correct output:

  $ grep MODULE_ALIAS drivers/input/mousedev.mod.c
  MODULE_ALIAS("input:b*v*p*e*-e*1,*2,*k*110,*r*0,*1,*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*2,*k*r*8,*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*3,*k*14A,*r*a*0,*1,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*3,*k*145,*r*a*0,*1,*18,*1C,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*3,*k*110,*r*a*0,*1,*m*l*s*f*w*");

However, building the same kernel on a 32-bit machine results in
incorrect output:

  $ grep MODULE_ALIAS drivers/input/mousedev.mod.c
  MODULE_ALIAS("input:b*v*p*e*-e*1,*2,*k*110,*130,*r*0,*1,*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*2,*k*r*8,*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*3,*k*14A,*16A,*r*a*0,*1,*20,*21,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*3,*k*145,*165,*r*a*0,*1,*18,*1C,*20,*21,*38,*3C,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*3,*k*110,*130,*r*a*0,*1,*20,*21,*m*l*s*f*w*");

A similar issue occurs with CONFIG_INPUT_JOYDEV=m. On a 64-bit build
machine, the output is:

  $ grep MODULE_ALIAS drivers/input/joydev.mod.c
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*0,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*2,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*8,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*6,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*k*120,*r*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*k*130,*r*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*k*2C0,*r*a*m*l*s*f*w*");

However, on a 32-bit machine, the output is incorrect:

  $ grep MODULE_ALIAS drivers/input/joydev.mod.c
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*0,*20,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*2,*22,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*8,*28,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*3,*k*r*a*6,*26,*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*k*11F,*13F,*r*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*k*11F,*13F,*r*a*m*l*s*f*w*");
  MODULE_ALIAS("input:b*v*p*e*-e*1,*k*2C0,*2E0,*r*a*m*l*s*f*w*");

When building a 64-bit kernel, BITS_PER_LONG is defined as 64. However,
on a 32-bit build machine, the constant 1L is a signed 32-bit value.
Left-shifting it beyond 32 bits causes wraparound, and shifting by 31
or 63 bits makes it a negative value.

The fix in commit e0e9263271 ("[PATCH] PATCH: 1 line 2.6.18 bugfix:
modpost-64bit-fix.patch") is incorrect; it only addresses cases where
a 64-bit kernel is built on a 64-bit build machine, overlooking cases
on a 32-bit build machine.

Using 1ULL ensures a 64-bit width on both 32-bit and 64-bit machines,
avoiding the wraparound issue.

Fixes: e0e9263271 ("[PATCH] PATCH: 1 line 2.6.18 bugfix: modpost-64bit-fix.patch")
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2024-11-03 23:58:56 +09:00
Masahiro Yamada
2e766a1f5f modpost: fix acpi MODULE_DEVICE_TABLE built with mismatched endianness
When CONFIG_SATA_AHCI_PLATFORM=m, modpost outputs incorect acpi
MODULE_ALIAS() if the endianness of the target and the build machine
do not match.

When the endianness of the target kernel and the build machine match,
the output is correct:

  $ grep 'MODULE_ALIAS("acpi' drivers/ata/ahci_platform.mod.c
  MODULE_ALIAS("acpi*:APMC0D33:*");
  MODULE_ALIAS("acpi*:010601:*");

However, when building a little-endian kernel on a big-endian machine
(or vice versa), the output is incorrect:

  $ grep 'MODULE_ALIAS("acpi' drivers/ata/ahci_platform.mod.c
  MODULE_ALIAS("acpi*:APMC0D33:*");
  MODULE_ALIAS("acpi*:0601??:*");

The 'cls' and 'cls_msk' fields are 32-bit.

DEF_FIELD() must be used instead of DEF_FIELD_ADDR() to correctly handle
endianness of these 32-bit fields.

The check 'if (cls)' was unnecessary; it never became NULL, as it was
the pointer to 'symval' plus the offset to the 'cls' field.

Fixes: 26095a01d3 ("ACPI / scan: Add support for ACPI _CLS device matching")
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2024-11-03 22:52:12 +09:00
Claudiu Beznea
c520bbb523 arm64: defconfig: Enable VBATTB clock and Renesas RTCA-3
Enable the Renesas VBATTB clock and RTCA-3 RTC drivers.  These are
available on the Renesas RZ/G3S SoC.  VBATTB is the clock provider for
the RTC counter.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Link: https://lore.kernel.org/20241101095720.2247815-10-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:33:11 +01:00
Claudiu Beznea
0cd647cd53 arm64: dts: renesas: rzg3s-smarc-som: Enable RTC
Enable RTC.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Link: https://lore.kernel.org/20241101095720.2247815-9-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:29:51 +01:00
Claudiu Beznea
ac948eb8ea arm64: dts: renesas: rzg3s-smarc-som: Enable VBATTB
Enable the VBATTB controller. It provides the clock for RTC.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Link: https://lore.kernel.org/20241101095720.2247815-8-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:29:51 +01:00
Claudiu Beznea
2d768aee9f arm64: dts: renesas: r9a08g045: Add RTC node
Add the DT node for the RTC IP available on the Renesas RZ/G3S SoC.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Link: https://lore.kernel.org/20241101095720.2247815-7-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:29:51 +01:00
Claudiu Beznea
23c44956bc arm64: dts: renesas: r9a08g045: Add VBATTB node
Add the DT node for the VBATTB IP along with DT bindings for the clock
it provides.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Link: https://lore.kernel.org/20241101095720.2247815-6-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:29:51 +01:00
Geert Uytterhoeven
75f3dec9a4 Renesas RZ/G3S DT Binding Definitions
VBATTB clock definitions for the Renesas RZ/G3S (R9A08G045) SoC, shared
 by driver and DT source files.
 -----BEGIN PGP SIGNATURE-----
 
 iHUEABYIAB0WIQQ9qaHoIs/1I4cXmEiKwlD9ZEnxcAUCZydUqwAKCRCKwlD9ZEnx
 cCzrAPoD4Kl7JyfZw7thlvx0SBEr4a1bxY/GKE1ee+FUA81OpgD5AQd88yI6pvQ7
 kT3ZBdNC1HkmZV204UsQauZtBfMoOw4=
 =jzFZ
 -----END PGP SIGNATURE-----

Merge tag 'renesas-r9a08g045-dt-binding-defs-tag3' into renesas-dts-for-v6.13

Renesas RZ/G3S DT Binding Definitions

VBATTB clock definitions for the Renesas RZ/G3S (R9A08G045) SoC, shared
by driver and DT source files.
2024-11-03 12:29:40 +01:00
Geert Uytterhoeven
8219a455ef arm64: dts: renesas: white-hawk-cpu-common: Add pin control for DSI-eDP IRQ
When the DSI to eDP bridge was added, pin control for the IRQ pin was
left out, because the pin controller did not support INTC-EX pins yet.

Commit 10544ec1b3 ("pinctrl: renesas: r8a779g0: Add INTC-EX
pins, groups, and function") added support for these pins, so add the
missing pin control description.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/89bab2008891be1f003a3c0dbcdf36af3b98da70.1729240573.git.geert+renesas@glider.be
2024-11-03 12:27:45 +01:00
Wolfram Sang
83eb988437 ARM: dts: renesas: r7s72100: Add DMA support to MMCIF
Add DMA properties to the device node for the MMC Host Interface.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/20241015224801.2535-6-wsa+renesas@sang-engineering.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:27:45 +01:00
Wolfram Sang
d823e397f3 ARM: dts: renesas: r7s72100: Add DMAC node
Add a device node for the Direct Memory Access Controller on the RZ/A1H
SoC.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/20241015224801.2535-5-wsa+renesas@sang-engineering.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:27:45 +01:00
Lad Prabhakar
9cc926e3fa arm64: dts: renesas: hihope: Drop #sound-dai-cells
"#sound-dai-cells" is required if the board is using "simple-card".
However, the HiHope board uses "audio-graph", thus remove the unneeded
`#sound-dai-cells`.

Commit 9e72606cd2 ("arm64: dts: renesas: #sound-dai-cells is used when
simple-card") updated the comment regarding usage of "#sound-dai-cells"
in the SoC DTSI but missed to remove "#sound-dai-cells" from board DTS
files.

Fixes: 9e72606cd2 ("arm64: dts: renesas: #sound-dai-cells is used when simple-card")
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/20241010135332.710648-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 12:27:21 +01:00
Biju Das
dabf72b85f clk: renesas: rzg2l: Fix FOUTPOSTDIV clk
While computing foutpostdiv_rate, the value of params->pl5_fracin
is discarded, which results in the wrong refresh rate. Fix the formula
for computing foutpostdiv_rate.

Fixes: 1561380ee7 ("clk: renesas: rzg2l: Add FOUTPOSTDIV clk support")
Signed-off-by: Hien Huynh <hien.huynh.px@renesas.com>
Signed-off-by: Biju Das <biju.das.jz@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/20241024134236.315289-1-biju.das.jz@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 11:48:39 +01:00
Claudiu Beznea
cdfd5daf90 dt-bindings: clock: renesas,r9a08g045-vbattb: Document VBATTB
The VBATTB IP of the Renesas RZ/G3S SoC controls the clock for RTC,
the tamper detector and a small general usage memory of 128B.

The VBATTB controller controls the clock for the RTC on the Renesas
RZ/G3S. The HW block diagram for the clock logic is as follows:

           +----------+ XC   `\
RTXIN  --->|          |----->| \       +----+  VBATTCLK
           | 32K clock|      |  |----->|gate|----------->
           | osc      | XBYP |  |      +----+
RTXOUT --->|          |----->| /
           +----------+      ,/

One could connect as input to this HW block either a crystal or
an external clock device. This is board specific.

After discussions w/ Stephen Boyd the clock tree associated with this
hardware block was exported in Linux as:

input-xtal
  xbyp
  xc
     mux
        vbattclk

where:
- input-xtal is the input clock (connected to RTXIN, RTXOUT pins)
- xc, xbyp are mux inputs
- mux is the internal mux
- vbattclk is the gate clock that feeds in the end the RTC

to allow selecting the input of the MUX though assigned-clock DT
properties, using the already existing clock drivers and avoid adding
other DT properties.

This allows select the input of the mux based on the type of the
connected input clock:
- if the 32768 crystal is connected as input for the VBATTB,
  the input of the mux should be xc
- if an external clock device is connected as input for the VBATTB the
  input of the mux should be xbyp

Add bindings for the VBATTB controller.

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
Link: https://lore.kernel.org/20241101095720.2247815-2-claudiu.beznea.uj@bp.renesas.com
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
2024-11-03 11:43:21 +01:00
Manivannan Sadhasivam
40e2125381
PCI: tegra194: Move controller cleanups to pex_ep_event_pex_rst_deassert()
Currently, the endpoint cleanup function dw_pcie_ep_cleanup() and EPF
deinit notify function pci_epc_deinit_notify() are called during the
execution of pex_ep_event_pex_rst_assert() i.e., when the host has asserted
PERST#. But quickly after this step, refclk will also be disabled by the
host.

All of the tegra194 endpoint SoCs supported as of now depend on the refclk
from the host for keeping the controller operational. Due to this
limitation, any access to the hardware registers in the absence of refclk
will result in a whole endpoint crash. Unfortunately, most of the
controller cleanups require accessing the hardware registers (like eDMA
cleanup performed in dw_pcie_ep_cleanup(), etc...). So these cleanup
functions can cause the crash in the endpoint SoC once host asserts PERST#.

One way to address this issue is by generating the refclk in the endpoint
itself and not depending on the host. But that is not always possible as
some of the endpoint designs do require the endpoint to consume refclk from
the host.

Thus, fix this crash by moving the controller cleanups to the start of
the pex_ep_event_pex_rst_deassert() function. This function is called
whenever the host has deasserted PERST# and it is guaranteed that the
refclk would be active at this point. So at the start of this function
(after enabling resources) the controller cleanup can be performed. Once
finished, rest of the code execution for PERST# deassert can continue as
usual.

Fixes: 473b2cf9c4 ("PCI: endpoint: Introduce 'epc_deinit' event and notify the EPF drivers")
Fixes: 570d7715ee ("PCI: dwc: ep: Introduce dw_pcie_ep_cleanup() API for drivers supporting PERST#")
Link: https://lore.kernel.org/r/20240817-pci-qcom-ep-cleanup-v1-2-d6b958226559@linaro.org
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
Cc: Jonathan Hunter <jonathanh@nvidia.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Vidya Sagar <vidyas@nvidia.com>
Cc: linux-tegra@vger.kernel.org
2024-11-03 09:08:33 +00:00
Manivannan Sadhasivam
7d7cf89b11
PCI: qcom-ep: Move controller cleanups to qcom_pcie_perst_deassert()
Currently, the endpoint cleanup function dw_pcie_ep_cleanup() and EPF
deinit notify function pci_epc_deinit_notify() are called during the
execution of qcom_pcie_perst_assert() i.e., when the host has asserted
PERST#. But quickly after this step, refclk will also be disabled by the
host.

All of the Qcom endpoint SoCs supported as of now depend on the refclk from
the host for keeping the controller operational. Due to this limitation,
any access to the hardware registers in the absence of refclk will result
in a whole endpoint crash. Unfortunately, most of the controller cleanups
require accessing the hardware registers (like eDMA cleanup performed in
dw_pcie_ep_cleanup(), powering down MHI EPF etc...). So these cleanup
functions are currently causing the crash in the endpoint SoC once host
asserts PERST#.

One way to address this issue is by generating the refclk in the endpoint
itself and not depending on the host. But that is not always possible as
some of the endpoint designs do require the endpoint to consume refclk from
the host (as I was told by the Qcom engineers).

Thus, fix this crash by moving the controller cleanups to the start of
the qcom_pcie_perst_deassert() function. qcom_pcie_perst_deassert() is
called whenever the host has deasserted PERST# and it is guaranteed that
the refclk would be active at this point. So at the start of this function
(after enabling resources), the controller cleanup can be performed. Once
finished, rest of the code execution for PERST# deassert can continue as
usual.

Fixes: 473b2cf9c4 ("PCI: endpoint: Introduce 'epc_deinit' event and notify the EPF drivers")
Fixes: 570d7715ee ("PCI: dwc: ep: Introduce dw_pcie_ep_cleanup() API for drivers supporting PERST#")
Link: https://lore.kernel.org/r/20240817-pci-qcom-ep-cleanup-v1-1-d6b958226559@linaro.org
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org>
2024-11-03 09:03:30 +00:00
Al Viro
38052c2dd7 deal with the last remaing boolean uses of fd_file()
Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
457a654939 css_set_fork(): switch to CLASS(fd_raw, ...)
reference acquired there by fget_raw() is not stashed anywhere -
we could as well borrow instead.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
7133dd5ac6 memcg_write_event_control(): switch to CLASS(fd)
some reordering required - take both fdget() to the point before
the allocations, with matching move of fdput() to the very end
of failure exit(s); after that it converts trivially.

simplify the cleanups that involve css_put(), while we are at it...

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
66635b0776 assorted variants of irqfd setup: convert to CLASS(fd)
in all of those failure exits prior to fdget() are plain returns and
the only thing done after fdput() is (on failure exits) a kfree(),
which can be done before fdput() just fine.

NOTE: in acrn_irqfd_assign() 'fail:' failure exit is wrong for
eventfd_ctx_fileget() failure (we only want fdput() there) and once
we stop doing that, it doesn't need to check if eventfd is NULL or
ERR_PTR(...) there.

NOTE: in privcmd we move fdget() up before the allocation - more
to the point, before the copy_from_user() attempt.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
8935989798 do_pollfd(): convert to CLASS(fd)
lift setting ->revents into the caller, so that failure exits (including
the early one) would be plain returns.

We need the scope of our struct fd to end before the store to ->revents,
since that's shared with the failure exits prior to the point where we
can do fdget().

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
d000e073ca convert do_select()
take the logics from fdget() to fdput() into an inlined helper - with existing
wait_key_set() subsumed into that.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
6b1a5ae9b5 convert vfs_dedupe_file_range().
fdput() is followed by checking fatal_signal_pending() (and aborting
the loop in such case).  fdput() is transposable with that check.
Yes, it'll probably end up with slightly fatter code (call after the
check has returned false + call on the almost never taken out-of-line path
instead of one call before the check), but it's not worth bothering with
explicit extra scope there (or dragging the check into the loop condition,
for that matter).

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
9bd812744d convert cifs_ioctl_copychunk()
fdput() moved past mnt_drop_file_write(); harmless, if somewhat cringeworthy.
Reordering could be avoided either by adding an explicit scope or by making
mnt_drop_file_write() called via __cleanup.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
44b11a56c3 convert media_request_get_by_fd()
the only thing done after fdput() (in failure cases) is a printk; safely
transposable with fdput()...

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
00ec41ac16 convert spu_run(2)
all failure exits prior to fdget() are returns, fdput() is immediately
followed by return.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
d8426e6972 switch spufs_calls_{get,put}() to CLASS() use
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
65c8941e7d convert cachestat(2)
fdput() can be transposed with copy_to_user()

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:07 -05:00
Al Viro
20d9eb3b87 convert do_preadv()/do_pwritev()
fdput() can be transposed with add_{r,w}char() and inc_sysc{r,w}();
it's the same story as with do_readv()/do_writev(), only with
fdput() instead of fdput_pos().

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
8152f82010 fdget(), more trivial conversions
all failure exits prior to fdget() leave the scope, all matching fdput()
are immediately followed by leaving the scope.

[xfs_ioc_commit_range() chunk moved here as well]

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
6348be02ee fdget(), trivial conversions
fdget() is the first thing done in scope, all matching fdput() are
immediately followed by leaving the scope.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
46b5e18ae8 privcmd_ioeventfd_assign(): don't open-code eventfd_ctx_fdget()
just call it, same as privcmd_ioeventfd_deassign() does...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
554ceb7a5e o2hb_region_dev_store(): avoid goto around fdget()/fdput()
Preparation for CLASS(fd) conversion.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
d7a9616ce0 introduce "fd_pos" class, convert fdget_pos() users to it.
fdget_pos() for constructor, fdput_pos() for cleanup, all users of
fd..._pos() converted trivially.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
048181992c fdget_raw() users: switch to CLASS(fd_raw)
Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
a6f46579d7 convert vmsplice() to CLASS(fd)
Irregularity here is fdput() not in the same scope as fdget(); we could
just lift it out vmsplice_type() in vmsplice(2), but there's no much
point keeping vmsplice_type() separate after that...

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
0d113fcbc2 simplify xfs_find_handle() a bit
XFS_IOC_FD_TO_HANDLE can grab a reference to copied ->f_path and
let the file go; results in simpler control flow - cleanup is
the same for both "by descriptor" and "by pathname" cases.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
54dac3dacc do_mq_notify(): switch to CLASS(fd)
The only failure exit before fdget() is a return, the only thing done
after fdput() is transposable with it.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
1aaf6a7e75 do_mq_notify(): saner skb freeing on failures
cleanup is convoluted enough as it is; it's easier to have early
failure outs do explicit kfree_skb(nc), rather than going to
contortions needed to reuse the cleanup from late failures.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00
Al Viro
f302edb9d8 switch netlink_getsockbyfilp() to taking descriptor
the only call site (in do_mq_notify()) obtains the argument
from an immediately preceding fdget() and it is immediately
followed by fdput(); might as well just replace it with
a variant that would take a descriptor instead of struct file *
and have file lookups handled inside that function.

Reviewed-by: Christian Brauner <brauner@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
2024-11-03 01:28:06 -05:00