Jakub Kicinski [Fri, 21 Feb 2025 02:51:41 +0000 (18:51 -0800)]
selftests: drv-net: test XDP, HDS auto and the ioctl path
Test XDP and HDS interaction. While at it add a test for using the IOCTL,
as that turned out to be the real culprit.
Testing bnxt:
# NETIF=eth0 ./ksft-net-drv/drivers/net/hds.py
KTAP version 1
1..12
ok 1 hds.get_hds
ok 2 hds.get_hds_thresh
ok 3 hds.set_hds_disable # SKIP disabling of HDS not supported by the device
ok 4 hds.set_hds_enable
ok 5 hds.set_hds_thresh_zero
ok 6 hds.set_hds_thresh_max
ok 7 hds.set_hds_thresh_gt
ok 8 hds.set_xdp
ok 9 hds.enabled_set_xdp
ok 10 hds.ioctl
ok 11 hds.ioctl_set_xdp
ok 12 hds.ioctl_enabled_set_xdp
# Totals: pass:11 fail:0 xfail:0 xpass:0 skip:1 error:0
and netdevsim:
# ./ksft-net-drv/drivers/net/hds.py
KTAP version 1
1..12
ok 1 hds.get_hds
ok 2 hds.get_hds_thresh
ok 3 hds.set_hds_disable
ok 4 hds.set_hds_enable
ok 5 hds.set_hds_thresh_zero
ok 6 hds.set_hds_thresh_max
ok 7 hds.set_hds_thresh_gt
ok 8 hds.set_xdp
ok 9 hds.enabled_set_xdp
ok 10 hds.ioctl
ok 11 hds.ioctl_set_xdp
ok 12 hds.ioctl_enabled_set_xdp
# Totals: pass:12 fail:0 xfail:0 xpass:0 skip:0 error:0
Netdevsim needs a sane default for tx/rx ring size.
ethtool 6.11 is needed for the --disable-netlink option.
Jakub Kicinski [Fri, 21 Feb 2025 02:51:40 +0000 (18:51 -0800)]
net: ethtool: fix ioctl confusing drivers about desired HDS user config
The legacy ioctl path does not have support for extended attributes.
So we issue a GET to fetch the current settings from the driver,
in an attempt to keep them unchanged. HDS is a bit "special" as
the GET only returns on/off while the SET takes a "ternary" argument
(on/off/default). If the driver was in the "default" setting -
executing the ioctl path binds it to on or off, even tho the user
did not intend to change HDS config.
Factor the relevant logic out of the netlink code and reuse it.
Fixes: 87c8f8496a05 ("bnxt_en: add support for tcp-data-split ethtool command") Acked-by: Stanislav Fomichev <sdf@fomichev.me> Tested-by: Daniel Xu <dxu@dxuuu.xyz> Tested-by: Taehee Yoo <ap420073@gmail.com> Link: https://patch.msgid.link/20250221025141.1132944-1-kuba@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ingo Molnar [Mon, 24 Feb 2025 19:40:17 +0000 (20:40 +0100)]
Merge branch into tip/master: 'perf/urgent'
# New commits in perf/urgent: bddf10d26e6e ("uprobes: Reject the shared zeropage in uprobe_write_opcode()") 2016066c6619 ("perf/core: Order the PMU list to fix warning about unordered pmu_ctx_list") 0fe8813baf4b ("perf/core: Add RCU read lock protection to perf_iterate_ctx()")
The cause is that zero pfn is set to the PTE without increasing the RSS
count in mfill_atomic_pte_zeropage() and the refcount of zero folio does
not increase accordingly. Then, the operation on the same pfn is performed
in uprobe_write_opcode()->__replace_page() to unconditional decrease the
RSS count and old_folio's refcount.
Therefore, two bugs are introduced:
1. The RSS count is incorrect, when process exit, the check_mm() report
error "Bad rss-count".
2. The reserved folio (zero folio) is freed when folio->refcount is zero,
then free_pages_prepare->free_page_is_bad() report error
"Bad page state".
There is more, the following warning could also theoretically be triggered:
Luo Gengkun [Wed, 22 Jan 2025 07:33:56 +0000 (07:33 +0000)]
perf/core: Order the PMU list to fix warning about unordered pmu_ctx_list
Syskaller triggers a warning due to prev_epc->pmu != next_epc->pmu in
perf_event_swap_task_ctx_data(). vmcore shows that two lists have the same
perf_event_pmu_context, but not in the same order.
The problem is that the order of pmu_ctx_list for the parent is impacted by
the time when an event/PMU is added. While the order for a child is
impacted by the event order in the pinned_groups and flexible_groups. So
the order of pmu_ctx_list in the parent and child may be different.
To fix this problem, insert the perf_event_pmu_context to its proper place
after iteration of the pmu_ctx_list.
The follow testcase can trigger above warning:
# perf record -e cycles --call-graph lbr -- taskset -c 3 ./a.out &
# perf stat -e cpu-clock,cs -p xxx // xxx is the pid of a.out
pid = fork();
if (pid == -1) {
printf("fork error\n");
return;
}
if (pid == 0) {
while (1) {
count++;
}
} else {
while (1) {
count++;
}
}
}
The testcase first opens an LBR event, so it will allocate task_ctx_data,
and then open tracepoint and software events, so the parent context will
have 3 different perf_event_pmu_contexts. On inheritance, child ctx will
insert the perf_event_pmu_context in another order and the warning will
trigger.
Paolo Bonzini [Mon, 24 Feb 2025 18:20:30 +0000 (13:20 -0500)]
Merge tag 'kvm-riscv-fixes-6.14-1' of https://github.com/kvm-riscv/linux into HEAD
KVM/riscv fixes for 6.14, take #1
- Fix hart status check in SBI HSM extension
- Fix hart suspend_type usage in SBI HSM extension
- Fix error returned by SBI IPI and TIME extensions for
unsupported function IDs
- Fix suspend_type usage in SBI SUSP extension
- Remove unnecessary vcpu kick after injecting interrupt
via IMSIC guest file
Paolo Bonzini [Mon, 24 Feb 2025 18:20:00 +0000 (13:20 -0500)]
Merge tag 'kvmarm-fixes-6.14-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD
KVM/arm64 fixes for 6.14, take #3
- Fix TCR_EL2 configuration to not use the ASID in TTBR1_EL2
and not mess-up T1SZ/PS by using the HCR_EL2.E2H==0 layout.
- Bring back the VMID allocation to the vcpu_load phase, ensuring
that we only setup VTTBR_EL2 once on VHE. This cures an ugly
race that would lead to running with an unallocated VMID.
Breno Leitao [Fri, 17 Jan 2025 14:41:07 +0000 (06:41 -0800)]
perf/core: Add RCU read lock protection to perf_iterate_ctx()
The perf_iterate_ctx() function performs RCU list traversal but
currently lacks RCU read lock protection. This causes lockdep warnings
when running perf probe with unshare(1) under CONFIG_PROVE_RCU_LIST=y:
WARNING: suspicious RCU usage
kernel/events/core.c:8168 RCU-list traversed in non-reader section!!
This protection was previously present but was removed in commit bd2756811766 ("perf: Rewrite core context handling"). Add back the
necessary rcu_read_lock()/rcu_read_unlock() pair around
perf_iterate_ctx() call in perf_event_exec().
[ mingo: Use scoped_guard() as suggested by Peter ]
The ES8328 codec driver, which is also used for the ES8388 chip that
appears to have an identical register map, claims that the output can
either take the route from DAC->Mixer->Output or through DAC->Output
directly. To the best of what I could find, this is not true, and
creates problems.
Without DACCONTROL17 bit index 7 set for the left channel, as well as
DACCONTROL20 bit index 7 set for the right channel, I cannot get any
analog audio out on Left Out 2 and Right Out 2 respectively, despite the
DAPM routes claiming that this should be possible. Furthermore, the same
is the case for Left Out 1 and Right Out 1, showing that those two don't
have a direct route from DAC to output bypassing the mixer either.
Those control bits toggle whether the DACs are fed (stale bread?) into
their respective mixers. If one "unmutes" the mixer controls in
alsamixer, then sure, the audio output works, but if it doesn't work
without the mixer being fed the DAC input then evidently it's not a
direct output from the DAC.
ES8328/ES8388 are seemingly not alone in this. ES8323, which uses a
separate driver for what appears to be a very similar register map,
simply flips those two bits on in its probe function, and then pretends
there is no power management whatsoever for the individual controls.
Fair enough.
My theory as to why nobody has noticed this up to this point is that
everyone just assumes it's their fault when they had to unmute an
additional control in ALSA.
Fix this in the es8328 driver by removing the erroneous direct route,
then get rid of the playback switch controls and have those bits tied to
the mixer's widget instead, which until now had no register to play
with.
Linus Walleij [Thu, 20 Feb 2025 18:48:15 +0000 (19:48 +0100)]
net: dsa: rtl8366rb: Fix compilation problem
When the kernel is compiled without LED framework support the
rtl8366rb fails to build like this:
rtl8366rb.o: in function `rtl8366rb_setup_led':
rtl8366rb.c:953:(.text.unlikely.rtl8366rb_setup_led+0xe8):
undefined reference to `led_init_default_state_get'
rtl8366rb.c:980:(.text.unlikely.rtl8366rb_setup_led+0x240):
undefined reference to `devm_led_classdev_register_ext'
As this is constantly coming up in different randconfig builds,
bite the bullet and create a separate file for the offending
code, split out a header with all stuff needed both in the
core driver and the leds code.
Add a new bool Kconfig option for the LED compile target, such
that it depends on LEDS_CLASS=y || LEDS_CLASS=RTL8366RB
which make LED support always available when LEDS_CLASS is
compiled into the kernel and enforce that if the LEDS_CLASS
is a module, then the RTL8366RB driver needs to be a module
as well so that modprobe can resolve the dependencies.
Fixes: 32d617005475 ("net: dsa: realtek: add LED drivers for rtl8366rb") Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202502070525.xMUImayb-lkp@intel.com/ Signed-off-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Vladimir Oltean <olteanv@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ard Biesheuvel [Sun, 23 Feb 2025 15:48:54 +0000 (16:48 +0100)]
efivarfs: Defer PM notifier registration until .fill_super
syzbot reports an issue that turns out to be caused by the fact that the
efivarfs PM notifier may be invoked before the efivarfs_fs_info::sb
field is populated, resulting in a NULL deference.
So defer the registration until efivarfs_fill_super() is invoked.
Patrick Rudolph [Fri, 21 Feb 2025 11:15:16 +0000 (12:15 +0100)]
efi/cper: Fix cper_arm_ctx_info alignment
According to the UEFI Common Platform Error Record appendix, the
processor context information structure is a variable length structure,
but "is padded with zeros if the size is not a multiple of 16 bytes".
Currently this isn't honoured, causing all but the first structure to
be garbage when printed. Thus align the size to be a multiple of 16.
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Patrick Rudolph [Fri, 21 Feb 2025 08:12:42 +0000 (09:12 +0100)]
efi/cper: Fix cper_ia_proc_ctx alignment
According to the UEFI Common Platform Error Record appendix, the
IA32/X64 Processor Context Information Structure is a variable length
structure, but "is padded with zeros if the size is not a multiple
of 16 bytes".
Currently this isn't honoured, causing all but the first structure to
be garbage when printed. Thus align the size to be a multiple of 16.
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
When there is a failure during bind QP, the cleanup flow destroys the
counter regardless if it is the one that created it or not, which is
problematic since if it isn't the one that created it, that counter could
still be in use.
Fix that by destroying the counter only if it was created during this call.
Brendan Jackman [Thu, 20 Feb 2025 12:23:40 +0000 (12:23 +0000)]
scripts/gdb: add $lx_per_cpu_ptr()
We currently have $lx_per_cpu() which works fine for stuff that kernel
code would access via per_cpu(). But this doesn't work for stuff that
kernel code accesses via per_cpu_ptr():
(gdb) p $lx_per_cpu(node_data[1].node_zones[2]->per_cpu_pageset)
Cannot access memory at address 0xffff11105fbd6c28
This is because we take the address of the pointer and use that as the
offset, instead of using the stored value.
Add a GDB version that mirrors the kernel API, which uses the pointer
value.
To be consistent with per_cpu_ptr(), we need to return the pointer value
instead of dereferencing it for the user. Therefore, move the existing
dereference out of the per_cpu() Python helper and do that only in the
$lx_per_cpu() implementation.
Michal Koutný [Fri, 21 Feb 2025 17:02:49 +0000 (18:02 +0100)]
pid: optional first-fit pid allocation
Noone would need to use this allocation strategy (it's slower, pid
numbers collide sooner). Its primary purpose are pid namespaces in
conjunction with pids.max cgroup limit which keeps (virtual) pid numbers
below the given limit. This is for 32-bit userspace programs that may
not work well with pid numbers above 65536.
Michal Koutný [Fri, 21 Feb 2025 17:02:48 +0000 (18:02 +0100)]
Revert "pid: allow pid_max to be set per pid namespace"
Patch series "Alternative "pid_max" for 32-bit userspace".
pid_max is sort of a legacy limit (its value and partially the concept
too, given the existence of pids cgroup controller). It is tempting to
make the pid_max value part of a pid namespace to provide compat
environment for 32-bit applications [1]. On the other hand, it provides
yet another mechanism for limitation of task count. Even without
namespacing of pid_max value, the configuration of conscious limit is
confusing for users [2].
This series builds upon the idea of restricting the number (amount) of
tasks by pids controller and ensuring that number (pid) never exceeds the
amount of tasks. This would not currently work out of the box because
next-fit pid allocation would continue to assign numbers (pids) higher
than the actual amount (there would be gaps in the lower range of the
interval). The patch 2/2 implements this idea by extending semantics of
ns_last_pid knob to allow first-fit numbering. (The implementation has
clumsy ifdefery, which can might be dropped since it's too x86-centric.)
The patch 1/2 is a mere revert to simplify pid_max to one global limit
only.
It is already difficult for users to troubleshoot which of multiple pid
limits restricts their workload. I'm afraid making pid_max
per-(hierarchical-)NS will contribute to confusion. Also, the
implementation copies the limit upon creation from parent, this pattern
showed cumbersome with some attributes in legacy cgroup controllers --
it's subject to race condition between parent's limit modification and
children creation and once copied it must be changed in the descendant.
This is very similar to what pids.max of a cgroup (already) does that can
be used as an alternative.
Kuan-Wei Chiu [Sat, 15 Feb 2025 16:56:18 +0000 (00:56 +0800)]
lib min_heap: use size_t for array size and index variables
Replace the int type with size_t for variables representing array sizes
and indices in the min-heap implementation. Using size_t aligns with
standard practices for size-related variables and avoids potential issues
on platforms where int may be insufficient to represent all valid sizes or
indices.
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:52 +0000 (21:39 +0100)]
reboot: retire hw_protection_reboot and hw_protection_shutdown helpers
The hw_protection_reboot and hw_protection_shutdown functions mix
mechanism with policy: They let the driver requesting an emergency action
for hardware protection also decide how to deal with it.
This is inadequate in the general case as a driver reporting e.g. an
imminent power failure can't know whether a shutdown or a reboot would be
more appropriate for a given hardware platform.
With the addition of the hw_protection parameter, it's now possible to
configure at runtime the default emergency action and drivers are expected
to use hw_protection_trigger to have this parameter dictate policy.
As no current users of either hw_protection_shutdown or
hw_protection_shutdown helpers remain, remove them, as not to tempt driver
authors to call them.
Existing users now either defer to hw_protection_trigger or call
__hw_protection_trigger with a suitable argument directly when they have
inside knowledge on whether a reboot or shutdown would be more
appropriate.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-12-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:51 +0000 (21:39 +0100)]
thermal: core: allow user configuration of hardware protection action
In the general case, we don't know which of system shutdown or reboot is
the better action to take to protect hardware in an emergency situation.
We thus allow the policy to come from the device-tree in the form of an
optional critical-action OF property, but so far there was no way for the
end user to configure this.
With recent addition of the hw_protection parameter, the user can now
choose a default action for the case, where the driver isn't fully sure
what's the better course of action.
Let's make use of this by passing HWPROT_ACT_DEFAULT in absence of the
critical-action OF property.
As HWPROT_ACT_DEFAULT is shutdown by default, this introduces no
functional change for users, unless they start using the new parameter.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-11-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:50 +0000 (21:39 +0100)]
dt-bindings: thermal: give OS some leeway in absence of critical-action
An operating system may allow its user to configure the action to be
undertaken on critical overtemperature events.
However, the bindings currently mandate an absence of the critical-action
property to be equal to critical-action = "shutdown", which would mean any
differing user configuration would violate the bindings.
Resolve this by documenting the absence of the property to mean that the
OS gets to decide.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-10-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Acked-by: Rob Herring (Arm) <robh@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Cc: Tzung-Bi Shih <tzungbi@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:49 +0000 (21:39 +0100)]
platform/chrome: cros_ec_lpc: prepare for hw_protection_shutdown removal
In the general case, a driver doesn't know which of system shutdown or
reboot is the better action to take to protect hardware in an emergency
situation. For this reason, hw_protection_shutdown is going to be removed
in favor of hw_protection_trigger, which defaults to shutdown, but may be
configured at kernel runtime to be a reboot instead.
The ChromeOS EC situation is different as we do know that shutdown is the
correct action as the EC is programmed to force reset after the short
period, thus replace hw_protection_shutdown with __hw_protection_trigger
with HWPROT_ACT_SHUTDOWN as argument to maintain the same behavior.
No functional change.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-9-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Acked-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:48 +0000 (21:39 +0100)]
regulator: allow user configuration of hardware protection action
When the core detects permanent regulator hardware failure or imminent
power failure of a critical supply, it will call hw_protection_shutdown in
an attempt to do a limited orderly shutdown followed by powering off the
system.
This doesn't work out well for many unattended embedded systems that don't
have support for shutdown and that power on automatically when power is
supplied:
- A brief power cycle gets detected by the driver
- The kernel powers down the system and SoC goes into shutdown mode
- Power is restored
- The system remains oblivious to the restored power
- System needs to be manually power cycled for a duration long enough
to drain the capacitors
Allow users to fix this by calling the newly introduced
hw_protection_trigger() instead: This way the hw_protection commandline or
sysfs parameter is used to dictate the policy of dealing with the
regulator fault.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-8-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Reviewed-by: Matti Vaittinen <mazziesaccount@gmail.com> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:47 +0000 (21:39 +0100)]
reboot: add support for configuring emergency hardware protection action
We currently leave the decision of whether to shutdown or reboot to
protect hardware in an emergency situation to the individual drivers.
This works out in some cases, where the driver detecting the critical
failure has inside knowledge: It binds to the system management controller
for example or is guided by hardware description that defines what to do.
In the general case, however, the driver detecting the issue can't know
what the appropriate course of action is and shouldn't be dictating the
policy of dealing with it.
Therefore, add a global hw_protection toggle that allows the user to
specify whether shutdown or reboot should be the default action when the
driver doesn't set policy.
This introduces no functional change yet as hw_protection_trigger() has no
callers, but these will be added in subsequent commits.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-7-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:46 +0000 (21:39 +0100)]
reboot: indicate whether it is a HARDWARE PROTECTION reboot or shutdown
It currently depends on the caller, whether we attempt a hardware
protection shutdown (poweroff) or a reboot. A follow-up commit will make
this partially user-configurable, so it's a good idea to have the
emergency message clearly state whether the kernel is going for a reboot
or a shutdown.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-6-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:45 +0000 (21:39 +0100)]
reboot: rename now misleading __hw_protection_shutdown symbols
The __hw_protection_shutdown function name has become misleading since it
can cause either a shutdown (poweroff) or a reboot depending on its
argument.
To avoid further confusion, let's rename it, so it doesn't suggest that a
poweroff is all it can do.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-5-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:43 +0000 (21:39 +0100)]
docs: thermal: sync hardware protection doc with code
Originally, the thermal framework's only hardware protection action was to
trigger a shutdown. This has been changed a little over a year ago to
also support rebooting as alternative hardware protection action.
Update the documentation to reflect this.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-3-e1c09b090c0c@pengutronix.de Fixes: 62e79e38b257 ("thermal/thermal_of: Allow rebooting after critical temp") Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Reviewed-by: Matti Vaittinen <mazziesaccount@gmail.com> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:42 +0000 (21:39 +0100)]
reboot: reboot, not shutdown, on hw_protection_reboot timeout
hw_protection_shutdown() will kick off an orderly shutdown and if that
takes longer than a configurable amount of time, an emergency shutdown
will occur.
Recently, hw_protection_reboot() was added for those systems that don't
implement a proper shutdown and are better served by rebooting and having
the boot firmware worry about doing something about the critical
condition.
On timeout of the orderly reboot of hw_protection_reboot(), the system
would go into shutdown, instead of reboot. This is not a good idea, as
going into shutdown was explicitly not asked for.
Fix this by always doing an emergency reboot if hw_protection_reboot() is
called and the orderly reboot takes too long.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-2-e1c09b090c0c@pengutronix.de Fixes: 79fa723ba84c ("reboot: Introduce thermal_zone_device_critical_reboot()") Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Reviewed-by: Matti Vaittinen <mazziesaccount@gmail.com> Cc: Benson Leung <bleung@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Mark Brown <broonie@kernel.org> Cc: Matteo Croce <teknoraver@meta.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring (Arm) <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Ahmad Fatoum [Mon, 17 Feb 2025 20:39:41 +0000 (21:39 +0100)]
reboot: replace __hw_protection_shutdown bool action parameter with an enum
Patch series "reboot: support runtime configuration of emergency
hw_protection action", v3.
We currently leave the decision of whether to shutdown or reboot to
protect hardware in an emergency situation to the individual drivers.
This works out in some cases, where the driver detecting the critical
failure has inside knowledge: It binds to the system management controller
for example or is guided by hardware description that defines what to do.
This is inadequate in the general case though as a driver reporting e.g.
an imminent power failure can't know whether a shutdown or a reboot would
be more appropriate for a given hardware platform.
To address this, this series adds a hw_protection kernel parameter and
sysfs toggle that can be used to change the action from the shutdown
default to reboot. A new hw_protection_trigger API then makes use of this
default action.
My particular use case is unattended embedded systems that don't have
support for shutdown and that power on automatically when power is
supplied:
- A brief power cycle gets detected by the driver
- The kernel powers down the system and SoC goes into shutdown mode
- Power is restored
- The system remains oblivious to the restored power
- System needs to be manually power cycled for a duration long enough
to drain the capacitors
With this series, such systems can configure the kernel with
hw_protection=reboot to have the boot firmware worry about critical
conditions.
This patch (of 12):
Currently __hw_protection_shutdown() either reboots or shuts down the
system according to its shutdown argument.
To make the logic easier to follow, both inside __hw_protection_shutdown
and at caller sites, lets replace the bool parameter with an enum.
This will be extra useful, when in a later commit, a third action is added
to the enumeration.
No functional change.
Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-0-e1c09b090c0c@pengutronix.de Link: https://lkml.kernel.org/r/20250217-hw_protection-reboot-v3-1-e1c09b090c0c@pengutronix.de Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de> Reviewed-by: Tzung-Bi Shih <tzungbi@kernel.org> Cc: Benson Leung <bleung@chromium.org> Cc: Mark Brown <broonie@kernel.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Fabio Estevam <festevam@denx.de> Cc: Guenter Roeck <groeck@chromium.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Liam Girdwood <lgirdwood@gmail.com> Cc: Lukasz Luba <lukasz.luba@arm.com> Cc: Matteo Croce <teknoraver@meta.com> Cc: Matti Vaittinen <mazziesaccount@gmail.com> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Rob Herring <robh@kernel.org> Cc: Rui Zhang <rui.zhang@intel.com> Cc: Sascha Hauer <kernel@pengutronix.de> Cc: "Serge E. Hallyn" <serge@hallyn.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Buffer heads are attached to folios, not to pages. Also
flush_dcache_page() is now deprecated in favour of flush_dcache_folio().
Link: https://lkml.kernel.org/r/20250213214533.2242224-2-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> Acked-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Joel Becker <jlbec@evilplan.org> Cc: Jun Piao <piaojun@huawei.com> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Mark Tinguely <mark.tinguely@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
ocfs2: use memcpy_to_folio() in ocfs2_symlink_get_block()
Replace use of kmap_atomic() with the higher-level construct
memcpy_to_folio(). This removes a use of b_page and supports large folios
as well as being easier to understand. It also removes the check for
kmap_atomic() failing (because it can't).
Link: https://lkml.kernel.org/r/20250213214533.2242224-1-willy@infradead.org Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org> Acked-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Mark Tinguely <mark.tinguely@oracle.com> Cc: Mark Fasheh <mark@fasheh.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Jun Piao <piaojun@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Alban Kurti [Fri, 7 Feb 2025 18:39:06 +0000 (18:39 +0000)]
checkpatch: add warning for pr_* and dev_* macros without a trailing newline
Add a new check in scripts/checkpatch.pl to detect usage of pr_(level) and
dev_(level) macros (for both C and Rust) when the string literal does not
end with '\n'. Missing trailing newlines can lead to incomplete log lines
that do not appear properly in dmesg or in console output. To show an
example of this working after applying the patch we can run the script on
the commit that likely motivated this need/issue:
Also, the patch is able to handle correctly if there is a printing call
without a newline which then has a newline printed via pr_cont for both
Rust and C alike. If there is no newline printed and the patch ends or
there is another pr_* call before a newline with pr_cont is printed it
will show a warning. Not implemented for dev_cont because it is not clear
to me if that is used at all.
One false warning that will be generated due to this change is in case we
have a patch that modifies a `pr_* call without a newline` which has a
pr_cont with a newline following it. In this case there will be a warning
but because the patch does not include the following pr_cont it will warn
there is nothing creating a newline. I have modified the warning to be
softer due to this known problem.
I have tested with comments, whitespace, differen orders of pr_* calls and
pr_cont and the only case that I suspect to be a problem is the one
outlined above.
Link: https://lkml.kernel.org/r/20250207-checkpatch-newline2-v4-1-26d8e80d0059@invicto.ai Signed-off-by: Alban Kurti <kurti@invicto.ai> Suggested-by: Miguel Ojeda <ojeda@kernel.org> Closes: https://github.com/Rust-for-Linux/linux/issues/1140 Cc: Alex Gaynor <alex.gaynor@gmail.com> Cc: Alice Ryhl <aliceryhl@google.com> Cc: Andreas Hindborg <a.hindborg@kernel.org> Cc: Andy Whitcroft <apw@canonical.com> Cc: Benno Lossin <benno.lossin@proton.me> Cc: Björn Roy Baron <bjorn3_gh@protonmail.com> Cc: Boqun Feng <boqun.feng@gmail.com> Cc: Dwaipayan Ray <dwaipayanray1@gmail.com> Cc: Gary Guo <gary@garyguo.net> Cc: Joe Perches <joe@perches.com> Cc: Lukas Bulwahn <lukas.bulwahn@gmail.com> Cc: Trevor Gross <tmgross@umich.edu> Cc: Charalampos Mitrodimas <charmitro@posteo.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Use rcuref_t for reference counting. This eliminates the cmpxchg loop in
the get and put path. This also eliminates the need to acquire the lock
in the put path because once the final user returns the reference, it can
no longer be obtained anymore.
Use rcuref_t for reference counting.
Link: https://lkml.kernel.org/r/20250203150525.456525-5-bigeasy@linutronix.de Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Boqun Feng <boqun.feng@gmail.com> Cc: Joel Fernandes <joel@joelfernandes.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Lai jiangshan <jiangshanlai@gmail.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Mengen Sun <mengensun@tencent.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com> Cc: YueHong Wu <yuehongwu@tencent.com> Cc: Zqiang <qiang.zhang1211@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The ucounts element is looked up under ucounts_lock. This can be
optimized by using RCU for a lockless lookup and return and element if the
reference can be obtained.
Replace hlist_head with hlist_nulls_head which is RCU compatible. Let
find_ucounts() search for the required item within a RCU section and
return the item if a reference could be obtained. This means
alloc_ucounts() will always return an element (unless the memory
allocation failed). Let put_ucounts() RCU free the element if the
reference counter dropped to zero.
Link: https://lkml.kernel.org/r/20250203150525.456525-4-bigeasy@linutronix.de Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Boqun Feng <boqun.feng@gmail.com> Cc: Joel Fernandes <joel@joelfernandes.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Lai jiangshan <jiangshanlai@gmail.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Mengen Sun <mengensun@tencent.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com> Cc: YueHong Wu <yuehongwu@tencent.com> Cc: Zqiang <qiang.zhang1211@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
ucount: replace get_ucounts_or_wrap() with atomic_inc_not_zero()
get_ucounts_or_wrap() increments the counter and if the counter is
negative then it decrements it again in order to reset the previous
increment. This statement can be replaced with atomic_inc_not_zero() to
only increment the counter if it is not yet 0.
This simplifies the get function because the put (if the get failed) can
be removed. atomic_inc_not_zero() is implement as a cmpxchg() loop which
can be repeated several times if another get/put is performed in parallel.
This will be optimized later.
Increment the reference counter only if not yet dropped to zero.
Link: https://lkml.kernel.org/r/20250203150525.456525-3-bigeasy@linutronix.de Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Boqun Feng <boqun.feng@gmail.com> Cc: Joel Fernandes <joel@joelfernandes.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Lai jiangshan <jiangshanlai@gmail.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Mengen Sun <mengensun@tencent.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com> Cc: YueHong Wu <yuehongwu@tencent.com> Cc: Zqiang <qiang.zhang1211@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
rcu: provide a static initializer for hlist_nulls_head
Patch series "ucount: Simplify refcounting with rcuref_t".
I noticed that the atomic_dec_and_lock_irqsave() in put_ucounts() loops
sometimes even during boot. Something like 2-3 iterations but still.
This series replaces the refcounting with rcuref_t and adds a RCU lookup.
This allows a lockless lookup in alloc_ucounts() if the entry is available
and a cmpxchg()less put of the item.
This patch (of 4):
Provide a static initializer for hlist_nulls_head so that it can be used
in statically defined data structures.
Link: https://lkml.kernel.org/r/20250203150525.456525-1-bigeasy@linutronix.de Link: https://lkml.kernel.org/r/20250203150525.456525-2-bigeasy@linutronix.de Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Reviewed-by: Paul E. McKenney <paulmck@kernel.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Boqun Feng <boqun.feng@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Joel Fernandes <joel@joelfernandes.org> Cc: Josh Triplett <josh@joshtriplett.org> Cc: Lai jiangshan <jiangshanlai@gmail.com> Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: Mengen Sun <mengensun@tencent.com> Cc: "Paul E . McKenney" <paulmck@kernel.org> Cc: "Uladzislau Rezki (Sony)" <urezki@gmail.com> Cc: YueHong Wu <yuehongwu@tencent.com> Cc: Zqiang <qiang.zhang1211@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Yury Norov [Wed, 5 Feb 2025 21:29:32 +0000 (16:29 -0500)]
lib/zlib: drop EQUAL macro
The macro is prehistoric, and only exists to help those readers who don't
know what memcmp() returns if memory areas differ. This is pretty well
documented, so the macro looks excessive.
Now that the only user of the macro depends on DEBUG_ZLIB config, GCC
warns about unused macro if the library is built with W=2 against
defconfig. So drop it for good.
Vlastimil Babka [Tue, 11 Feb 2025 15:16:11 +0000 (16:16 +0100)]
get_maintainer: add --substatus for reporting subsystem status - fix
The automatically enabled --substatus can break existing scripts that do
not disable --rolestats. Require that script output goes to a terminal to
enable it automatically.
Vlastimil Babka [Mon, 3 Feb 2025 11:13:16 +0000 (12:13 +0100)]
get_maintainer: add --substatus for reporting subsystem status
Patch series "get_maintainer: report subsystem status separately", v2.
The subsystem status (S: field) can inform a patch submitter if the
subsystem is well maintained or e.g. maintainers are missing. In
get_maintainer, it is currently reported with --role(stats) by adjusting
the maintainer role for any status different from Maintained. This has
two downsides:
- if a subsystem has only reviewers or mailing lists and no maintainers,
the status is not reported. For example Orphan subsystems typically
have no maintainers so there's nobody to report as orphan minder.
- the Supported status means that someone is paid for maintaining, but
it is reported as "supporter" for all the maintainers, which can be
incorrect (only some of them may be paid). People (including myself)
have been also confused about what "supporter" means.
The second point has been brought up in 2022 and the discussion in the end
resulted in adjusting documentation only [1]. I however agree with Ted's
points that it's misleading to take the subsystem status and apply it to
all maintainers [2].
The attempt to modify get_maintainer output was retracted after Joe
objected that the status becomes not reported at all [3]. This series
addresses that concern by reporting the status (unless it's the most
common Maintained one) on separate lines that follow the reported emails,
using a new --substatus parameter. Care is taken to reduce the noise to
minimum by not reporting the most common Maintained status, by default
require no opt-in that would need the users to discover the new parameter,
and at the same time not to break existing git --cc-cmd usage.
The subsystem status is currently reported with --role(stats) by adjusting
the maintainer role for any status different from Maintained. This has
two downsides:
- if a subsystem has only reviewers or mailing lists and no maintainers,
the status is not reported (i.e. typically, Orphan subsystems have no
maintainers)
- the Supported status means that someone is paid for maintaining, but
it is reported as "supporter" for all the maintainers, which can be
incorrect. People have been also confused about what "supporter"
means.
This patch introduces a new --substatus option and functionality aimed to
report the subsystem status separately, without adjusting the reported
maintainer role. After the e-mails are output, the status of subsystems
will follow, for example:
Sourabh Jain [Fri, 31 Jan 2025 11:38:30 +0000 (17:08 +0530)]
powerpc/crash: use generic crashkernel reservation
Commit 0ab97169aa05 ("crash_core: add generic function to do reservation")
added a generic function to reserve crashkernel memory. So let's use the
same function on powerpc and remove the architecture-specific code that
essentially does the same thing.
The generic crashkernel reservation also provides a way to split the
crashkernel reservation into high and low memory reservations, which can
be enabled for powerpc in the future.
Along with moving to the generic crashkernel reservation, the code related
to finding the base address for the crashkernel has been separated into
its own function name get_crash_base() for better readability and
maintainability.
Link: https://lkml.kernel.org/r/20250131113830.925179-8-sourabhjain@linux.ibm.com Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com> Reviewed-by: Mahesh Salgaonkar <mahesh@linux.ibm.com> Acked-by: Hari Bathini <hbathini@linux.ibm.com> Cc: Baoquan he <bhe@redhat.com> Cc: Madhavan Srinivasan <maddy@linux.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Sourabh Jain [Fri, 31 Jan 2025 11:38:29 +0000 (17:08 +0530)]
powerpc: insert System RAM resource to prevent crashkernel conflict
The next patch in the series with title "powerpc/crash: use generic
crashkernel reservation" enables powerpc to use generic crashkernel
reservation instead of custom implementation. This leads to exporting of
`Crash Kernel` memory to iomem_resource (/proc/iomem) via
insert_crashkernel_resources():kernel/crash_reserve.c or at another place
in the same file if HAVE_ARCH_ADD_CRASH_RES_TO_IOMEM_EARLY is set.
The add_system_ram_resources():arch/powerpc/mm/mem.c adds `System RAM` to
iomem_resource using request_resource(). This creates a conflict with
`Crash Kernel`, which is added by the generic crashkernel reservation
code. As a result, the kernel ultimately fails to add `System RAM` to
iomem_resource. Consequently, it does not appear in /proc/iomem.
There are multiple approches tried to avoid this:
1. Don't add Crash Kernel to iomem_resource:
- This has two issues.
First, it requires adding an architecture-specific hook in the
generic code. There are already two code paths to choose when to
add `Crash Kernel` to iomem_resource. This adds one more code path
to skip it.
Second, what if `Crash Kernel` is required in /proc/iomem in the
future? Many architectures do export it.
2. Don't add `System RAM` to iomem_resource by reverting commit c40dd2f766440 ("powerpc: Add System RAM to /proc/iomem"):
- It's not ideal to export `System RAM` via /proc/iomem, but since
it already done ealier and userspace tools like kdump and
kdump-utils rely on `System RAM` from /proc/iomem, removing it
will break userspace.
3. Add Crash Kernel along with System RAM to /proc/iomem:
This patch takes the third approach by updating add_system_ram_resources()
to use insert_resource() instead of the request_resource() API to add the
`System RAM` resource to iomem_resource. insert_resource() allows
inserting resources even if they overlap with existing ones. Since `Crash
Kernel` and `System RAM` resources are added to iomem_resource early in
the boot, any other conflict is not expected.
With the changes introduced here and in the next patch, "powerpc/crash:
use generic crashkernel reservation," /proc/iomem now exports `System RAM`
and `Crash Kernel` as shown below:
Commit 59d58189f3d9 ("crash: fix crash memory reserve exceed system memory
bug") fails crashkernel parsing if the crash size is found to be higher
than system RAM, which makes the memory_limit adjustment code ineffective
due to an early exit from reserve_crashkernel().
Regardless lets not violate the user-specified memory limit by adjusting
it. Remove this adjustment to ensure all reservations stay within the
limit. Commit f94f5ac07983 ("powerpc/fadump: Don't update the
user-specified memory limit") did the same for fadump.
Link: https://lkml.kernel.org/r/20250131113830.925179-6-sourabhjain@linux.ibm.com Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com> Reviewed-by: Mahesh Salgaonkar <mahesh@linux.ibm.com> Acked-by: Hari Bathini <hbathini@linux.ibm.com> Cc: Baoquan he <bhe@redhat.com> Cc: Madhavan Srinivasan <maddy@linux.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Sourabh Jain [Fri, 31 Jan 2025 11:38:27 +0000 (17:08 +0530)]
powerpc/crash: use generic APIs to locate memory hole for kdump
On PowerPC, the memory reserved for the crashkernel can contain components
like RTAS, TCE, OPAL, etc., which should be avoided when loading kexec
segments into crashkernel memory. Due to these special components,
PowerPC has its own set of APIs to locate holes in the crashkernel memory
for loading kexec segments for kdump. However, for loading kexec segments
in the kexec case, PowerPC already uses generic APIs to locate holes.
The previous patch in this series, titled "crash: Let arch decide usable
memory range in reserved area," introduced arch-specific hook to handle
such special regions in the crashkernel area. So, switch PowerPC to use
the generic APIs to locate memory holes for kdump and remove the redundant
PowerPC-specific APIs.
Link: https://lkml.kernel.org/r/20250131113830.925179-5-sourabhjain@linux.ibm.com Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com> Cc: Baoquan he <bhe@redhat.com> Cc: Hari Bathini <hbathini@linux.ibm.com> Cc: Madhavan Srinivasan <maddy@linux.ibm.com> Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Sourabh Jain [Fri, 31 Jan 2025 11:38:26 +0000 (17:08 +0530)]
crash: let arch decide usable memory range in reserved area
Although the crashkernel area is reserved, on architectures like PowerPC,
it is possible for the crashkernel reserved area to contain components
like RTAS, TCE, OPAL, etc. To avoid placing kexec segments over these
components, PowerPC has its own set of APIs to locate holes in the
crashkernel reserved area.
Add an arch hook in the generic locate mem hole APIs so that architectures
can handle such special regions in the crashkernel area while locating
memory holes for kexec segments using generic APIs. With this, a lot of
redundant arch-specific code can be removed, as it performs the exact same
job as the generic APIs.
To keep the generic and arch-specific changes separate, the changes
related to moving PowerPC to use the generic APIs and the removal of
PowerPC-specific APIs for memory hole allocation are done in a subsequent
patch titled "powerpc/crash: Use generic APIs to locate memory hole for
kdump.
Link: https://lkml.kernel.org/r/20250131113830.925179-4-sourabhjain@linux.ibm.com Signed-off-by: Sourabh Jain <sourabhjain@linux.ibm.com> Acked-by: Baoquan He <bhe@redhat.com> Cc: Hari Bathini <hbathini@linux.ibm.com> Cc: Madhavan Srinivasan <maddy@linux.ibm.com> Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Sourabh Jain [Fri, 31 Jan 2025 11:38:24 +0000 (17:08 +0530)]
kexec: initialize ELF lowest address to ULONG_MAX
Patch series "powerpc/crash: use generic crashkernel reservation", v3.
Commit 0ab97169aa05 ("crash_core: add generic function to do reservation")
added a generic function to reserve crashkernel memory. So let's use the
same function on powerpc and remove the architecture-specific code that
essentially does the same thing.
The generic crashkernel reservation also provides a way to split the
crashkernel reservation into high and low memory reservations, which can
be enabled for powerpc in the future.
Additionally move powerpc to use generic APIs to locate memory hole for
kexec segments while loading kdump kernel.
This patch (of 7):
kexec_elf_load() loads an ELF executable and sets the address of the
lowest PT_LOAD section to the address held by the lowest_load_addr
function argument.
To determine the lowest PT_LOAD address, a local variable lowest_addr
(type unsigned long) is initialized to UINT_MAX. After loading each
PT_LOAD, its address is compared to lowest_addr. If a loaded PT_LOAD
address is lower, lowest_addr is updated. However, setting lowest_addr to
UINT_MAX won't work when the kernel image is loaded above 4G, as the
returned lowest PT_LOAD address would be invalid. This is resolved by
initializing lowest_addr to ULONG_MAX instead.
This issue was discovered while implementing crashkernel high/low
reservation on the PowerPC architecture.
I Hsin Cheng [Sun, 19 Jan 2025 06:24:08 +0000 (14:24 +0800)]
lib/plist.c: add shortcut for plist_requeue()
In the operation of plist_requeue(), "node" is deleted from the list
before queueing it back to the list again, which involves looping to find
the tail of same-prio entries.
If "node" is the head of same-prio entries which means its prio_list is on
the priority list, then "node_next" can be retrieve immediately by the
next entry of prio_list, instead of looping nodes on node_list.
The shortcut implementation can benefit plist_requeue() running the below
test, and the test result is shown in the following table.
One can observe from the test result that when the number of nodes of
same-prio entries is smaller, then the probability of hitting the shortcut
can be bigger, thus the benefit can be more significant.
While it tends to behave almost the same for long same-prio entries, since
the probability of taking the shortcut is much smaller.
The test is done on x86_64 architecture with v6.9 kernel and
Intel(R) Core(TM) i7-2600 CPU @ 3.40GHz.
Test script( executed in kernel module mode ):
int init_module(void)
{
unsigned int test_data[test_size];
/* Split the list into 10 different priority
* , when test_size is larger, the number of
* nodes within each priority is larger.
*/
for (i = 0; i < ARRAY_SIZE(test_data); i++) {
test_data[i] = i % 10;
}
for (i = 0; i < ARRAY_SIZE(test_node_local); i++) {
plist_node_init(test_node_local + i, 0);
test_node_local[i].prio = test_data[i];
}
for (i = 0; i < ARRAY_SIZE(test_node_local); i++) {
if (plist_node_empty(test_node_local + i)) {
plist_add(test_node_local + i, &test_head_local);
}
}
for (i = 0; i < ARRAY_SIZE(test_node_local); i += 1) {
start = ktime_get();
plist_requeue(test_node_local + i, &test_head_local);
end = ktime_get();
time_elapsed += (end - start);
}
Add a paragraph explaining what sort of capabilities a process would need
to read procfs data for some other process. Also mention that reading
data for its own process doesn't require any extra permissions.
scripts: add script to extract built-in firmware blobs
There is currently no tool to extract a firmware blob that is built-in
on vmlinux to the best of my knowledge. So if we have a kernel image
containing the blobs, and we want to rebuild the kernel with some debug
patches for example (and given that the image also has IKCONFIG=y), we
currently can't do that for the same versions for all the firmware
blobs, _unless_ we have exact commits of linux-firmware for the
specific versions for each firmware included.
Through the options CONFIG_EXTRA_FIRMWARE{_DIR} one is able to build a
kernel including firmware blobs in a built-in fashion. This is usually
the case of built-in drivers that require some blobs in order to work
properly, for example, like in non-initrd based systems.
Add hereby a script to extract these blobs from a non-stripped vmlinux,
similar to the idea of "extract-ikconfig". The firmware loader interface
saves such built-in blobs as rodata entries, having a field for the FW
name as "_fw_<module_name>_<firmware_name>_bin"; the tool extracts files
named "<module_name>_<firmware_name>" for each rodata firmware entry
detected. It makes use of awk, bash, dd and readelf, pretty standard
tooling for Linux development.
With this tool, we can blindly extract the FWs and easily re-add them
in the new debug kernel build, allowing a more deterministic testing
without the burden of "hunting down" the proper version of each
firmware binary.
Link: https://lkml.kernel.org/r/20250120190436.127578-1-gpiccoli@igalia.com Signed-off-by: Guilherme G. Piccoli <gpiccoli@igalia.com> Suggested-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com> Reviewed-by: Thadeu Lima de Souza Cascardo <cascardo@igalia.com> Cc: Danilo Krummrich <dakr@kernel.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Luis Chamberalin <mcgrof@kernel.org> Cc: Masahiro Yamada <masahiroy@kernel.org> Cc: Nathan Chancellor <nathan@kernel.org> Cc: Nicolas Schier <nicolas@fjasle.eu> Cc: "Rafael J. Wysocki" <rafael@kernel.org> Cc: Russ Weight <russ.weight@linux.dev> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Yang Yang [Fri, 17 Jan 2025 14:20:13 +0000 (22:20 +0800)]
MAINTAINERS: add Yang Yang as a co-maintainer of PER-TASK DELAY ACCOUNTING
Balbir Singh is the unique maintainer of PER-TASK DELAY ACCOUNTING, and he
had started work on cgroupstats a long time back, this subsystem then is
not growing at a very rapid pace. With their excellent work delay
accounting is still very useful for observing and optimizing system delay,
but still needs continuous improvement. Yang Yang with his team had
worked for most of the recent patches of the subsystem, and he has a
strong willing to help, Balbir Singh is glad to see that, so add him as a
co-maintainer.
Andrii Nakryiko [Mon, 27 Jan 2025 22:21:14 +0000 (14:21 -0800)]
mm,procfs: allow read-only remote mm access under CAP_PERFMON
It's very common for various tracing and profiling toolis to need to
access /proc/PID/maps contents for stack symbolization needs to learn
which shared libraries are mapped in memory, at which file offset, etc.
Currently, access to /proc/PID/maps requires CAP_SYS_PTRACE (unless we are
looking at data for our own process, which is a trivial case not too
relevant for profilers use cases).
Unfortunately, CAP_SYS_PTRACE implies way more than just ability to
discover memory layout of another process: it allows to fully control
arbitrary other processes. This is problematic from security POV for
applications that only need read-only /proc/PID/maps (and other similar
read-only data) access, and in large production settings CAP_SYS_PTRACE is
frowned upon even for the system-wide profilers.
On the other hand, it's already possible to access similar kind of
information (and more) with just CAP_PERFMON capability. E.g., setting up
PERF_RECORD_MMAP collection through perf_event_open() would give one
similar information to what /proc/PID/maps provides.
CAP_PERFMON, together with CAP_BPF, is already a very common combination
for system-wide profiling and observability application. As such, it's
reasonable and convenient to be able to access /proc/PID/maps with
CAP_PERFMON capabilities instead of CAP_SYS_PTRACE.
For procfs, these permissions are checked through common mm_access()
helper, and so we augment that with cap_perfmon() check *only* if
requested mode is PTRACE_MODE_READ. I.e., PTRACE_MODE_ATTACH wouldn't be
permitted by CAP_PERFMON. So /proc/PID/mem, which uses
PTRACE_MODE_ATTACH, won't be permitted by CAP_PERFMON, but /proc/PID/maps,
/proc/PID/environ, and a bunch of other read-only contents will be
allowable under CAP_PERFMON.
Besides procfs itself, mm_access() is used by process_madvise() and
process_vm_{readv,writev}() syscalls. The former one uses
PTRACE_MODE_READ to avoid leaking ASLR metadata, and as such CAP_PERFMON
seems like a meaningful allowable capability as well.
process_vm_{readv,writev} currently assume PTRACE_MODE_ATTACH level of
permissions (though for readv PTRACE_MODE_READ seems more reasonable, but
that's outside the scope of this change), and as such won't be affected by
this patch.
gentwo / linux / .git / log