| Age | Commit message (Collapse) | Author |
|---|
|
Convert hashtab.c from raw_spinlock to rqspinlock, and drop the hashed
per-cpu counter crud from the code base which is no longer necessary.
Closes: https://lore.kernel.org/bpf/675302fd.050a0220.2477f.0004.GAE@google.com
Closes: https://lore.kernel.org/bpf/000000000000b3e63e061eed3f6b@google.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-20-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Introduce locktorture support for rqspinlock using the newly added
macros as the first in-kernel user and consumer. Guard the code with
CONFIG_BPF_SYSCALL ifdef since rqspinlock is not available otherwise.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-19-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Ensure that the rqspinlock code is only built when the BPF subsystem is
compiled in. Depending on queued spinlock support, we may or may not end
up building the queued spinlock slowpath, and instead fallback to the
test-and-set implementation. Also add entries to MAINTAINERS file.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-18-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
We ripped out PV and virtualization related bits from rqspinlock in an
earlier commit, however, a fair lock performs poorly within a virtual
machine when the lock holder is preempted. As such, retain the
virt_spin_lock fallback to test and set lock, but with timeout and
deadlock detection. We can do this by simply depending on the
resilient_tas_spin_lock implementation from the previous patch.
We don't integrate support for CONFIG_PARAVIRT_SPINLOCKS yet, as that
requires more involved algorithmic changes and introduces more
complexity. It can be done when the need arises in the future.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-15-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Include a test-and-set fallback when queued spinlock support is not
available. Introduce a rqspinlock type to act as a fallback when
qspinlock support is absent.
Include ifdef guards to ensure the slow path in this file is only
compiled when CONFIG_QUEUED_SPINLOCKS=y. Subsequent patches will add
further logic to ensure fallback to the test-and-set implementation
when queued spinlock support is unavailable on an architecture.
Unlike other waiting loops in rqspinlock code, the one for test-and-set
has no theoretical upper bound under contention, therefore we need a
longer timeout than usual. Bump it up to a second in this case.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-14-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
While the timeout logic provides guarantees for the waiter's forward
progress, the time until a stalling waiter unblocks can still be long.
The default timeout of 1/4 sec can be excessively long for some use
cases. Additionally, custom timeouts may exacerbate recovery time.
Introduce logic to detect common cases of deadlocks and perform quicker
recovery. This is done by dividing the time from entry into the locking
slow path until the timeout into intervals of 1 ms. Then, after each
interval elapses, deadlock detection is performed, while also polling
the lock word to ensure we can quickly break out of the detection logic
and proceed with lock acquisition.
A 'held_locks' table is maintained per-CPU where the entry at the bottom
denotes a lock being waited for or already taken. Entries coming before
it denote locks that are already held. The current CPU's table can thus
be looked at to detect AA deadlocks. The tables from other CPUs can be
looked at to discover ABBA situations. Finally, when a matching entry
for the lock being taken on the current CPU is found on some other CPU,
a deadlock situation is detected. This function can take a long time,
therefore the lock word is constantly polled in each loop iteration to
ensure we can preempt detection and proceed with lock acquisition, using
the is_lock_released check.
We set 'spin' member of rqspinlock_timeout struct to 0 to trigger
deadlock checks immediately to perform faster recovery.
Note: Extending lock word size by 4 bytes to record owner CPU can allow
faster detection for ABBA. It is typically the owner which participates
in a ABBA situation. However, to keep compatibility with existing lock
words in the kernel (struct qspinlock), and given deadlocks are a rare
event triggered by bugs, we choose to favor compatibility over faster
detection.
The release_held_lock_entry function requires an smp_wmb, while the
release store on unlock will provide the necessary ordering for us. Add
comments to document the subtleties of why this is correct. It is
possible for stores to be reordered still, but in the context of the
deadlock detection algorithm, a release barrier is sufficient and
needn't be stronger for unlock's case.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-13-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When we run out of maximum rqnodes, the original queued spin lock slow
path falls back to a try lock. In such a case, we are again susceptible
to stalls in case the lock owner fails to make progress. We use the
timeout as a fallback to break out of this loop and return to the
caller. This is a fallback for an extreme edge case, when on the same
CPU we run out of all 4 qnodes. When could this happen? We are in slow
path in task context, we get interrupted by an IRQ, which while in the
slow path gets interrupted by an NMI, whcih in the slow path gets
another nested NMI, which enters the slow path. All of the interruptions
happen after node->count++.
We use RES_DEF_TIMEOUT as our spinning duration, but in the case of this
fallback, no fairness is guaranteed, so the duration may be too small
for contended cases, as the waiting time is not bounded. Since this is
an extreme corner case, let's just prefer timing out instead of
attempting to spin for longer.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-12-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Implement the wait queue cleanup algorithm for rqspinlock. There are
three forms of waiters in the original queued spin lock algorithm. The
first is the waiter which acquires the pending bit and spins on the lock
word without forming a wait queue. The second is the head waiter that is
the first waiter heading the wait queue. The third form is of all the
non-head waiters queued behind the head, waiting to be signalled through
their MCS node to overtake the responsibility of the head.
In this commit, we are concerned with the second and third kind. First,
we augment the waiting loop of the head of the wait queue with a
timeout. When this timeout happens, all waiters part of the wait queue
will abort their lock acquisition attempts. This happens in three steps.
First, the head breaks out of its loop waiting for pending and locked
bits to turn to 0, and non-head waiters break out of their MCS node spin
(more on that later). Next, every waiter (head or non-head) attempts to
check whether they are also the tail waiter, in such a case they attempt
to zero out the tail word and allow a new queue to be built up for this
lock. If they succeed, they have no one to signal next in the queue to
stop spinning. Otherwise, they signal the MCS node of the next waiter to
break out of its spin and try resetting the tail word back to 0. This
goes on until the tail waiter is found. In case of races, the new tail
will be responsible for performing the same task, as the old tail will
then fail to reset the tail word and wait for its next pointer to be
updated before it signals the new tail to do the same.
We terminate the whole wait queue because of two main reasons. Firstly,
we eschew per-waiter timeouts with one applied at the head of the wait
queue. This allows everyone to break out faster once we've seen the
owner / pending waiter not responding for the timeout duration from the
head. Secondly, it avoids complicated synchronization, because when not
leaving in FIFO order, prev's next pointer needs to be fixed up etc.
Lastly, all of these waiters release the rqnode and return to the
caller. This patch underscores the point that rqspinlock's timeout does
not apply to each waiter individually, and cannot be relied upon as an
upper bound. It is possible for the rqspinlock waiters to return early
from a failed lock acquisition attempt as soon as stalls are detected.
The head waiter cannot directly WRITE_ONCE the tail to zero, as it may
race with a concurrent xchg and a non-head waiter linking its MCS node
to the head's MCS node through 'prev->next' assignment.
One notable thing is that we must use RES_DEF_TIMEOUT * 2 as our maximum
duration for the waiting loop (for the wait queue head), since we may
have both the owner and pending bit waiter ahead of us, and in the worst
case, need to span their maximum permitted critical section lengths.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-11-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
The pending bit is used to avoid queueing in case the lock is
uncontended, and has demonstrated benefits for the 2 contender scenario,
esp. on x86. In case the pending bit is acquired and we wait for the
locked bit to disappear, we may get stuck due to the lock owner not
making progress. Hence, this waiting loop must be protected with a
timeout check.
To perform a graceful recovery once we decide to abort our lock
acquisition attempt in this case, we must unset the pending bit since we
own it. All waiters undoing their changes and exiting gracefully allows
the lock word to be restored to the unlocked state once all participants
(owner, waiters) have been recovered, and the lock remains usable.
Hence, set the pending bit back to zero before returning to the caller.
Introduce a lockevent (rqspinlock_lock_timeout) to capture timeout
event statistics.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-10-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Currently, for rqspinlock usage, the implementation of
smp_cond_load_acquire (and thus, atomic_cond_read_acquire) are
susceptible to stalls on arm64, because they do not guarantee that the
conditional expression will be repeatedly invoked if the address being
loaded from is not written to by other CPUs. When support for
event-streams is absent (which unblocks stuck WFE-based loops every
~100us), we may end up being stuck forever.
This causes a problem for us, as we need to repeatedly invoke the
RES_CHECK_TIMEOUT in the spin loop to break out when the timeout
expires.
Let us import the smp_cond_load_acquire_timewait implementation Ankur is
proposing in [0], and then fallback to it once it is merged.
While we rely on the implementation to amortize the cost of sampling
check_timeout for us, it will not happen when event stream support is
unavailable. This is not the common case, and it would be difficult to
fit our logic in the time_expr_ns >= time_limit_ns comparison, hence
just let it be.
[0]: https://lore.kernel.org/lkml/20250203214911.898276-1-ankur.a.arora@oracle.com
Cc: Ankur Arora <ankur.a.arora@oracle.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-9-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Introduce policy macro RES_CHECK_TIMEOUT which can be used to detect
when the timeout has expired for the slow path to return an error. It
depends on being passed two variables initialized to 0: ts, ret. The
'ts' parameter is of type rqspinlock_timeout.
This macro resolves to the (ret) expression so that it can be used in
statements like smp_cond_load_acquire to break the waiting loop
condition.
The 'spin' member is used to amortize the cost of checking time by
dispatching to the implementation every 64k iterations. The
'timeout_end' member is used to keep track of the timestamp that denotes
the end of the waiting period. The 'ret' parameter denotes the status of
the timeout, and can be checked in the slow path to detect timeouts
after waiting loops.
The 'duration' member is used to store the timeout duration for each
waiting loop. The default timeout value defined in the header
(RES_DEF_TIMEOUT) is 0.25 seconds.
This macro will be used as a condition for waiting loops in the slow
path. Since each waiting loop applies a fresh timeout using the same
rqspinlock_timeout, we add a new RES_RESET_TIMEOUT as well to ensure the
values can be easily reinitialized to the default state.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-8-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Changes to rqspinlock in subsequent commits will be algorithmic
modifications, which won't remain in agreement with the implementations
of paravirt spin lock and virt_spin_lock support. These future changes
include measures for terminating waiting loops in slow path after a
certain point. While using a fair lock like qspinlock directly inside
virtual machines leads to suboptimal performance under certain
conditions, we cannot use the existing virtualization support before we
make it resilient as well. Therefore, drop it for now.
Note that we need to drop qspinlock_stat.h, as it's only relevant in
case of CONFIG_PARAVIRT_SPINLOCKS=y, but we need to keep lock_events.h
in the includes, which was indirectly pulled in before.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-7-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
This header contains the public declarations usable in the rest of the
kernel for rqspinlock.
Let's also type alias qspinlock to rqspinlock_t to ensure consistent use
of the new lock type. We want to remove dependence on the qspinlock type
in later patches as we need to provide a test-and-set fallback, hence
begin abstracting away from now onwards.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-6-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
In preparation for introducing a new lock implementation, Resilient
Queued Spin Lock, or rqspinlock, we first begin our modifications by
using the existing qspinlock.c code as the base. Simply copy the code to
a new file and rename functions and variables from 'queued' to
'resilient_queued'.
Since we place the file in kernel/bpf, include needs to be relative.
This helps each subsequent commit in clearly showing how and where the
code is being changed. The only change after a literal copy in this
commit is renaming the functions where necessary, and rename qnodes to
rqnodes. Let's also use EXPORT_SYMBOL_GPL for rqspinlock slowpath.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-5-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
To support upcoming changes that require inspecting the return value
once the conditional waiting loop in arch_mcs_spin_lock_contended
terminates, modify the macro to preserve the result of
smp_cond_load_acquire. This enables checking the return value as needed,
which will help disambiguate the MCS node’s locked state in future
patches.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-4-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Move qspinlock helper functions that encode, decode tail word, set and
clear the pending and locked bits, and other miscellaneous definitions
and macros to a private header. To this end, create a qspinlock.h header
file in kernel/locking. Subsequent commits will introduce a modified
qspinlock slow path function, thus moving shared code to a private
header will help minimize unnecessary code duplication.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-3-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When we currently create a pidfd we check that the task hasn't been
reaped right before we create the pidfd. But it is of course possible
that by the time we return the pidfd to userspace the task has already
been reaped since we don't check again after having created a dentry for
it.
This was fine until now because that race was meaningless. But now that
we provide PIDFD_INFO_EXIT it is a problem because it is possible that
the kernel returns a reaped pidfd and it depends on the race whether
PIDFD_INFO_EXIT information is available. This depends on if the task
gets reaped before or after a dentry has been attached to struct pid.
Make this consistent and only returned pidfds for reaped tasks if
PIDFD_INFO_EXIT information is available. This is done by performing
another check whether the task has been reaped right after we attached a
dentry to struct pid.
Since pidfs_exit() is called before struct pid's task linkage is removed
the case where the task got reaped but a dentry was already attached to
struct pid and exit information was recorded and published can be
handled correctly. In that case we do return a pidfd for a reaped task
like we would've before.
Link: https://lore.kernel.org/r/20250316-kabel-fehden-66bdb6a83436@brauner
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Christian Brauner <brauner@kernel.org>
|
|
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
The current verifier error message states that tail_calls are not
allowed in non-JITed programs with BPF-to-BPF calls. While this is
accurate, it is not the only scenario where this restriction applies.
Some architectures do not support this feature combination even when
programs are JITed. This update improves the error message to better
reflect these limitations.
Suggested-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Signed-off-by: Andrea Terzolo <andreaterzolo3@gmail.com>
Acked-by: Shung-Hsi Yu <shung-hsi.yu@suse.com>
Link: https://lore.kernel.org/r/20250318083551.8192-1-andreaterzolo3@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
If we attach fexit/fmod_ret to __noreturn functions, it will cause an
issue that the bpf trampoline image will be left over even if the bpf
link has been destroyed. Take attaching do_exit() with fexit for example.
The fexit works as follows,
bpf_trampoline
+ __bpf_tramp_enter
+ percpu_ref_get(&tr->pcref);
+ call do_exit()
+ __bpf_tramp_exit
+ percpu_ref_put(&tr->pcref);
Since do_exit() never returns, the refcnt of the trampoline image is
never decremented, preventing it from being freed. That can be verified
with as follows,
$ bpftool link show <<<< nothing output
$ grep "bpf_trampoline_[0-9]" /proc/kallsyms
ffffffffc04cb000 t bpf_trampoline_6442526459 [bpf] <<<< leftover
In this patch, all functions annotated with __noreturn are rejected, except
for the following cases:
- Functions that result in a system reboot, such as panic,
machine_real_restart and rust_begin_unwind
- Functions that are never executed by tasks, such as rest_init and
cpu_startup_entry
- Functions implemented in assembly, such as rewind_stack_and_make_dead and
xen_cpu_bringup_again, lack an associated BTF ID.
With this change, attaching fexit probes to functions like do_exit() will
be rejected.
$ ./fexit
libbpf: prog 'fexit': BPF program load failed: -EINVAL
libbpf: prog 'fexit': -- BEGIN PROG LOAD LOG --
Attaching fexit/fmod_ret to __noreturn functions is rejected.
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://lore.kernel.org/r/20250318114447.75484-2-laoar.shao@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
storage
The current cgrp storage has a percpu counter, bpf_cgrp_storage_busy,
to detect potential deadlock at a spin_lock that the local storage
acquires during new storage creation.
There are false positives. It turns out to be too noisy in
production. For example, a bpf prog may be doing a
bpf_cgrp_storage_get on map_a. An IRQ comes in and triggers
another bpf_cgrp_storage_get on a different map_b. It will then
trigger the false positive deadlock check in the percpu counter.
On top of that, both are doing lookup only and no need to create
new storage, so practically it does not need to acquire
the spin_lock.
The bpf_task_storage_get already has a strategy to minimize this
false positive by only failing if the bpf_task_storage_get needs
to create a new storage and the percpu counter is busy. Creating
a new storage is the only time it must acquire the spin_lock.
This patch borrows the same idea. Unlike task storage that
has a separate variant for tracing (_recur) and non-tracing, this
patch stays with one bpf_cgrp_storage_get helper to keep it simple
for now in light of the upcoming res_spin_lock.
The variable could potentially use a better name noTbusy instead
of nobusy. This patch follows the same naming in
bpf_task_storage_get for now.
I have tested it by temporarily adding noinline to
the cgroup_storage_lookup(), traced it by fentry, and the fentry
program succeeded in calling bpf_cgrp_storage_get().
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20250318182759.3676094-1-martin.lau@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Move the definition of the struct mcs_spinlock from the private
mcs_spinlock.h header in kernel/locking to the mcs_spinlock.h
asm-generic header, since we will need to reference it from the
qspinlock.h header in subsequent commits.
Reviewed-by: Barret Rhoden <brho@google.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20250316040541.108729-2-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
The perf_event_read_event_output helper is currently only available to
tracing protrams, but is useful for other BPF programs like sched_ext
schedulers. When the helper is available, provide its bpf_func_proto
directly from the bpf base_proto.
Signed-off-by: Emil Tsalapatis (Meta) <emil@etsalapatis.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20250318030753.10949-1-emil@etsalapatis.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Move s390 sysctls (spin_retry and userprocess_debug) into their own
files under arch/s390. Create two new sysctl tables
(2390_{fault,spin}_sysctl_table) which will be initialized with
arch_initcall placing them after their original place in proc_root_init.
This is part of a greater effort to move ctl tables into their
respective subsystems which will reduce the merge conflicts in
kernel/sysctl.c.
Signed-off-by: joel granados <joel.granados@kernel.org>
Acked-by: Heiko Carstens <hca@linux.ibm.com>
Link: https://lore.kernel.org/r/20250306-jag-mv_ctltables-v2-6-71b243c8d3f8@kernel.org
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
|
|
No functional changes.
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/r/20250313142744.1323281-1-mjguzik@gmail.com
Signed-off-by: Christian Brauner <brauner@kernel.org>
|
|
For small folios, we traditionally use the mapcount to decide whether it
was "certainly mapped exclusively" by a single MM (mapcount == 1) or
whether it "maybe mapped shared" by multiple MMs (mapcount > 1). For
PMD-sized folios that were PMD-mapped, we were able to use a similar
mechanism (single PMD mapping), but for PTE-mapped folios and in the
future folios that span multiple PMDs, this does not work.
So we need a different mechanism to handle large folios. Let's add a new
mechanism to detect whether a large folio is "certainly mapped
exclusively", or whether it is "maybe mapped shared".
We'll use this information next to optimize CoW reuse for PTE-mapped
anonymous THP, and to convert folio_likely_mapped_shared() to
folio_maybe_mapped_shared(), independent of per-page mapcounts.
For each large folio, we'll have two slots, whereby a slot stores:
(1) an MM id: unique id assigned to each MM
(2) a per-MM mapcount
If a slot is unoccupied, it can be taken by the next MM that maps folio
page.
In addition, we'll remember the current state -- "mapped exclusively" vs.
"maybe mapped shared" -- and use a bit spinlock to sync on updates and to
reduce the total number of atomic accesses on updates. In the future, it
might be possible to squeeze a proper spinlock into "struct folio". For
now, keep it simple, as we require the whole thing with THP only, that is
incompatible with RT.
As we have to squeeze this information into the "struct folio" of even
folios of order-1 (2 pages), and we generally want to reduce the required
metadata, we'll assign each MM a unique ID that can fit into an int. In
total, we can squeeze everything into 4x int (2x long) on 64bit.
32bit support is a bit challenging, because we only have 2x long == 2x int
in order-1 folios. But we can make it work for now, because we neither
expect many MMs nor very large folios on 32bit.
We will reliably detect folios as "mapped exclusively" vs. "mapped
shared" as long as only two MMs map pages of a folio at one point in time
-- for example with fork() and short-lived child processes, or with apps
that hand over state from one instance to another.
As soon as three MMs are involved at the same time, we might detect "maybe
mapped shared" although the folio is "mapped exclusively".
Example 1:
(1) App1 faults in a (shmem/file-backed) folio page -> Tracked as MM0
(2) App2 faults in a folio page -> Tracked as MM1
(4) App1 unmaps all folio pages
-> We will detect "mapped exclusively".
Example 2:
(1) App1 faults in a (shmem/file-backed) folio page -> Tracked as MM0
(2) App2 faults in a folio page -> Tracked as MM1
(3) App3 faults in a folio page -> No slot available, tracked as "unknown"
(4) App1 and App2 unmap all folio pages
-> We will detect "maybe mapped shared".
Make use of __always_inline to keep possible performance degradation when
(un)mapping large folios to a minimum.
Note: by squeezing the two flags into the "unsigned long" that stores the
MM ids, we can use non-atomic __bit_spin_unlock() and non-atomic
setting/clearing of the "maybe mapped shared" bit, effectively not adding
any new atomics on the hot path when updating the large mapcount + new
metadata, which further helps reduce the runtime overhead in
micro-benchmarks.
Link: https://lkml.kernel.org/r/20250303163014.1128035-13-david@redhat.com
Signed-off-by: David Hildenbrand <david@redhat.com>
Cc: Andy Lutomirks^H^Hski <luto@kernel.org>
Cc: Borislav Betkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jann Horn <jannh@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Lance Yang <ioworker0@gmail.com>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Matthew Wilcow (Oracle) <willy@infradead.org>
Cc: Michal Koutn <mkoutny@suse.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: tejun heo <tj@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zefan Li <lizefan.x@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull probes fixes from Masami Hiramatsu:
- Clean up tprobe correctly when module unload
Tracepoint probes do not set TRACEPOINT_STUB on the 'tpoint' pointer
when unloading a module, thus they show as a normal 'fprobe' instead
of 'tprobe' and never come back
- Fix leakage of tprobe module refcount
When a tprobe's target module is loaded, it gets the module's
refcount in the module notifier but forgot to put it after
registering the probe on it.
Fix it by getting the refcount only when registering tprobe.
* tag 'probes-fixes-v6.14-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing: tprobe-events: Fix leakage of module refcount
tracing: tprobe-events: Fix to clean up tprobe correctly when module unload
|
|
Fixed some formatting specifiers errors, such as using %d for int and %u
for unsigned int, as well as other byte-length types.
Perform type cast using the type derived from the data type itself, for
example, if it's originally an int, it will be cast to unsigned int if
forced to unsigned.
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20250311112809.81901-3-jiayuan.chen@linux.dev
|
|
Return prog's btf_id from bpf_prog_get_info_by_fd regardless of capable
check. This patch enables scenario, when freplace program, running
from user namespace, requires to query target prog's btf.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Link: https://lore.kernel.org/bpf/20250317174039.161275-3-mykyta.yatsenko5@gmail.com
|
|
Currently BPF_BTF_GET_FD_BY_ID requires CAP_SYS_ADMIN, which does not
allow running it from user namespace. This creates a problem when
freplace program running from user namespace needs to query target
program BTF.
This patch relaxes capable check from CAP_SYS_ADMIN to CAP_BPF and adds
support for BPF token that can be passed in attributes to syscall.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20250317174039.161275-2-mykyta.yatsenko5@gmail.com
|
|
The UEFI Specification version 2.9 introduces the concept of memory
acceptance: some Virtual Machine platforms, such as Intel TDX or AMD
SEV-SNP, require memory to be accepted before it can be used by the guest.
Accepting memory is expensive. The memory must be allocated by the VMM
and then brought to a known safe state: cache must be flushed, memory must
be zeroed with the guest's encryption key, and associated metadata must be
manipulated. These operations must be performed from a trusted
environment (firmware or TDX module). Switching context to and from it
also takes time.
This cost adds up. On large confidential VMs, memory acceptance alone can
take minutes. It is better to delay memory acceptance until the memory is
actually needed.
The kernel accepts memory when it is allocated from buddy allocator for
the first time. This reduces boot time and decreases memory overhead as
the VMM can allocate memory as needed.
It does not work when the guest attempts to kexec into a new kernel.
The kexec segments' destination addresses are not allocated by the buddy
allocator. Instead, they are searched from normal system RAM (top-down or
bottom-up) and exclude driver-managed memory, ACPI, persistent, and
reserved memory. Unaccepted memory is normal system RAM from kernel point
of view and kexec can place segments there.
Kexec bypasses the code path in buddy allocator where memory gets accepted
and it leads to a crash when kexec accesses segments' memory.
Accept the destination addresses during the kexec load, immediately after
they pass sanity checks. This ensures the code is located in a common
place shared by both the kexec_load and kexec_file_load system calls.
This will not conflict with the accounting in try_to_accept_memory_one()
since the accounting is set during kernel boot and decremented when pages
are moved to the freelists. There is no harm in invoking accept_memory()
on a page before making it available to the buddy allocator.
No need to worry about re-accepting memory since accept_memory() checks
the unaccepted bitmap before accepting a memory page.
Although a user may perform kexec loading without ever triggering the
jump, it doesn't impact much since kexec loading is not in a
performance-critical path. Additionally, the destination addresses are
always searched and found in the same location on a given system.
Changes to the destination address searching logic to locate only memory in
either unaccepted or accepted status are unnecessary and complicated.
[kirill.shutemov@linux.intel.com: update the commit message]
Link: https://lkml.kernel.org/r/20250307084411.2150367-1-kirill.shutemov@linux.intel.com
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Ashish Kalra <Ashish.Kalra@amd.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Jianxiong Gao <jxgao@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Currently, up to 23 bytes of the source string are copied to the
destination buffer (including the comma and anything after it), only to
then manually NUL-terminate the destination buffer again at index 'len'
(where the comma was found).
Fix this by calling strscpy() with 'len' instead of the destination buffer
size to copy only as many bytes from the source string as needed.
Change the length check to allow 'len' to be less than or equal to the
destination buffer size to fill the whole buffer if needed.
Remove the if-check for the return value of strscpy(), because calling
strscpy() with 'len' always truncates the source string at the comma as
expected and NUL-terminates the destination buffer at the corresponding
index instead. Remove the manual NUL-termination.
No functional changes intended.
Link: https://lkml.kernel.org/r/20250313133004.36406-2-thorsten.blum@linux.dev
Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Cc: Song Liu <song@kernel.org>
Cc: Thomas Gleinxer <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
Clearing is an atomic op and the flag is not set most of the time.
When creating and destroying threads in the same process with the pthread
family, the primary bottleneck is calls to sigprocmask which take the
process-wide sighand lock.
Avoiding the atomic gives me a 2% bump in start/teardown rate at 24-core
scale.
[akpm@linux-foundation.org: add unlikely() as well]
Link: https://lkml.kernel.org/r/20250303134908.423242-1-mjguzik@gmail.com
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
|
|
The new option is CONFIG_NULL_TTY_DEFAULT_CONSOLE.
if enabled, and CONFIG_VT is disabled, ttynull will become the default
primary console device.
ttynull will be the only console device usually with this option enabled.
Some architectures do call add_preferred_console() which may add another
console though.
Motivation:
Many distributions ship with CONFIG_VT enabled. On tested desktop hardware
if CONFIG_VT is disabled, the default console device falls back to
/dev/ttyS0 instead of /dev/tty.
This could cause issues in user space, and hardware problems:
1. The user space issues include the case where /dev/ttyS0 is
disconnected, and the TCGETS ioctl, which some user space libraries use
as a probe to determine if a file is a tty, is called on /dev/console and
fails. Programs that call isatty() on /dev/console and get an incorrect
false value may skip expected logging to /dev/console.
2. The hardware issues include the case if a user has a science instrument
or other device connected to the /dev/ttyS0 port, and they were to upgrade
to a kernel that is disabling the CONFIG_VT option, kernel logs will then
be sent to the device connected to /dev/ttyS0 unless they edit their
kernel command line manually.
The new CONFIG_NULL_TTY_DEFAULT_CONSOLE option will give users and
distribution maintainers an option to avoid this. Disabling CONFIG_VT and
enabling CONFIG_NULL_TTY_DEFAULT_CONSOLE will ensure the default kernel
console behavior is not dependent on hardware configuration by default, and
avoid unexpected new behavior on devices connected to the /dev/ttyS0 serial
port.
Reviewed-by: Petr Mladek <pmladek@suse.com>
Tested-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Adam Simonelli <adamsimonelli@gmail.com>
Link: https://lore.kernel.org/r/20250314160749.3286153-2-adamsimonelli@gmail.com
[pmladek@suse.com: Fixed indentation of the commit message.]
Signed-off-by: Petr Mladek <pmladek@suse.com>
|
|
The are no callers of partition_sched_domains_locked() outside
topology.c.
Stop exposing such function.
Suggested-by: Waiman Long <llong@redhat.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Waiman Long <longman@redhat.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/Z9MSC96a8FcqWV3G@jlelli-thinkpadt14gen4.remote.csb
|
|
partition_and_rebuild_sched_domains() and partition_sched_domains() are
now equivalent.
Remove the former as a nice clean up.
Suggested-by: Waiman Long <llong@redhat.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Waiman Long <llong@redhat.com>
Reviewed-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Waiman Long <longman@redhat.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/Z9MR4ryNDJZDzsSG@jlelli-thinkpadt14gen4.remote.csb
|
|
We completely clean and restore root domains bandwidth accounting after
every root domains change, so the dl_clear_root_domain() call in
partition_sched_domains_locked() is redundant.
Remove it.
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Waiman Long <llong@redhat.com>
Reviewed-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Waiman Long <longman@redhat.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/Z9MRtcX4tz4tcLRR@jlelli-thinkpadt14gen4.remote.csb
|
|
Rebuilding of root domains accounting information (total_bw) is
currently broken on some cases, e.g. suspend/resume on aarch64. Problem
is that the way we keep track of domain changes and try to add bandwidth
back is convoluted and fragile.
Fix it by simplify things by making sure bandwidth accounting is cleared
and completely restored after root domains changes (after root domains
are again stable).
To be sure we always call dl_rebuild_rd_accounting while holding
cpuset_mutex we also add cpuset_reset_sched_domains() wrapper.
Fixes: 53916d5fd3c0 ("sched/deadline: Check bandwidth overflow earlier for hotplug")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Co-developed-by: Waiman Long <llong@redhat.com>
Signed-off-by: Waiman Long <llong@redhat.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/Z9MRfeJKJUOyUSto@jlelli-thinkpadt14gen4.remote.csb
|
|
Bandwidth checks and updates that work on root domains currently employ
a cookie mechanism for efficiency. This mechanism is very much tied to
when root domains are first created and initialized.
Generalize the cookie mechanism so that it can be used also later at
runtime while updating root domains. Also, additionally guard it with
sched_domains_mutex, since domains need to be stable while updating them
(and it will be required for further dynamic changes).
Fixes: 53916d5fd3c0 ("sched/deadline: Check bandwidth overflow earlier for hotplug")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Waiman Long <longman@redhat.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/Z9MQaiXPvEeW_v7x@jlelli-thinkpadt14gen4.remote.csb
|
|
Create wrappers for sched_domains_mutex so that it can transparently be
used on both CONFIG_SMP and !CONFIG_SMP, as some function will need to
do.
Fixes: 53916d5fd3c0 ("sched/deadline: Check bandwidth overflow earlier for hotplug")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Waiman Long <longman@redhat.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/Z9MP5Oq9RB8jBs3y@jlelli-thinkpadt14gen4.remote.csb
|
|
SCHED_DEADLINE special tasks get a fake bandwidth that is only used to
make sure sleeping and priority inheritance 'work', but it is ignored
for runtime enforcement and admission control.
Be consistent with it also when rebuilding root domains.
Fixes: 53916d5fd3c0 ("sched/deadline: Check bandwidth overflow earlier for hotplug")
Reported-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Juri Lelli <juri.lelli@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Waiman Long <longman@redhat.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/20250313170011.357208-2-juri.lelli@redhat.com
|
|
Use preempt_model_str() instead of manually conducting the preemption
model.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: "Steven Rostedt (Google)" <rostedt@goodmis.org>
Link: https://lore.kernel.org/r/20250314160810.2373416-10-bigeasy@linutronix.de
|
|
The pmu specific data is saved in task_struct now. Remove it from event
context structure.
Remove swap_task_ctx() as well.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250314172700.438923-7-kan.liang@linux.intel.com
|
|
The individual architectures often add the preemption model to the begin
of the backtrace. This is the case on X86 or ARM64 for the "die" case
but not for regular warning. With the addition of DYNAMIC_PREEMPT for
PREEMPT_RT we end up with CONFIG_PREEMPT and CONFIG_PREEMPT_RT set
simultaneously. That means that everyone who tried to add that piece of
information gets it wrong for PREEMPT_RT because PREEMPT is checked
first.
Provide a generic function which returns the current scheduling model
considering LAZY preempt and the current state of PREEMPT_DYNAMIC.
The resulting strings are:
┏━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━┓
┃ Model ┃ -RT -DYN ┃ +RT -DYN ┃ -RT +DYN ┃ +RT +DYN ┃
┡━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━┩
│NONE │ NONE │ n/a │ PREEMPT(none) │ n/a │
├───────────┼──────────────┼───────────────────┼────────────────────┼───────────────────┤
│VOLUNTARY │ VOLUNTARY │ n/a │ PREEMPT(voluntary) │ n/a │
├───────────┼──────────────┼───────────────────┼────────────────────┼───────────────────┤
│FULL │ PREEMPT │ PREEMPT_RT │ PREEMPT(full) │ PREEMPT_{RT,full} │
├───────────┼──────────────┼───────────────────┼────────────────────┼───────────────────┤
│LAZY │ PREEMPT_LAZY │ PREEMPT_{RT,LAZY} │ PREEMPT(lazy) │ PREEMPT_{RT,lazy} │
└───────────┴──────────────┴───────────────────┴────────────────────┴───────────────────┘
[ The dynamic building of the string can lead to an empty string if the
function is invoked simultaneously on two CPUs. ]
Co-developed-by: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Signed-off-by: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Co-developed-by: "Steven Rostedt (Google)" <rostedt@goodmis.org>
Signed-off-by: "Steven Rostedt (Google)" <rostedt@goodmis.org>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Shrikanth Hegde <sshegde@linux.ibm.com>
Link: https://lore.kernel.org/r/20250314160810.2373416-2-bigeasy@linutronix.de
|
|
To save/restore LBR call stack data in system-wide mode, the task_struct
information is required.
Extend the parameters of sched_task() to supply task_struct information.
When schedule in, the LBR call stack data for new task will be restored.
When schedule out, the LBR call stack data for old task will be saved.
Only need to pass the required task_struct information.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250314172700.438923-4-kan.liang@linux.intel.com
|
|
The LBR call stack data has to be saved/restored during context switch
to fix the shorter LBRs call stacks issue in the system-wide mode.
Allocate PMU specific data and attach them to the corresponding
task_struct during LBR call stack monitoring.
When a LBR call stack event is accounted, the perf_ctx_data for the
related tasks will be allocated/attached by attach_perf_ctx_data().
When a LBR call stack event is unaccounted, the perf_ctx_data for
related tasks will be detached/freed by detach_perf_ctx_data().
The LBR call stack event could be a per-task event or a system-wide
event.
- For a per-task event, perf only allocates the perf_ctx_data for the
current task. If the allocation fails, perf will error out.
- For a system-wide event, perf has to allocate the perf_ctx_data for
both the existing tasks and the upcoming tasks.
The allocation for the existing tasks is done in perf_event_alloc().
If any allocation fails, perf will error out.
The allocation for the new tasks will be done in perf_event_fork().
A global reader/writer semaphore, global_ctx_data_rwsem, is added to
address the global race.
- The perf_ctx_data only be freed by the last LBR call stack event.
The number of the per-task events is tracked by refcount of each task.
Since the system-wide events impact all tasks, it's not practical to
go through the whole task list to update the refcount for each
system-wide event. The number of system-wide events is tracked by a
global variable global_ctx_data_ref.
Suggested-by: "Peter Zijlstra (Intel)" <peterz@infradead.org>
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20250314172700.438923-3-kan.liang@linux.intel.com
|
|
Some PMU specific data has to be saved/restored during context switch,
e.g. LBR call stack data. Currently, the data is saved in event context
structure, but only for per-process event. For system-wide event,
because of missing the LBR call stack data after context switch, LBR
callstacks are always shorter in comparison to per-process mode.
For example,
Per-process mode:
$perf record --call-graph lbr -- taskset -c 0 ./tchain_edit
- 99.90% 99.86% tchain_edit tchain_edit [.] f3
99.86% _start
__libc_start_main
generic_start_main
main
f1
- f2
f3
System-wide mode:
$perf record --call-graph lbr -a -- taskset -c 0 ./tchain_edit
- 99.88% 99.82% tchain_edit tchain_edit [.] f3
- 62.02% main
f1
f2
f3
- 28.83% f1
- f2
f3
- 28.83% f1
- f2
f3
- 8.88% generic_start_main
main
f1
f2
f3
It isn't practical to simply allocate the data for system-wide event in
CPU context structure for all tasks. We have no idea which CPU a task
will be scheduled to. The duplicated LBR data has to be maintained on
every CPU context structure. That's a huge waste. Otherwise, the LBR
data still lost if the task is scheduled to another CPU.
Save the pmu specific data in task_struct. The size of pmu specific data
is 788 bytes for LBR call stack. Usually, the overall amount of threads
doesn't exceed a few thousands. For 10K threads, keeping LBR data would
consume additional ~8MB. The additional space will only be allocated
during LBR call stack monitoring. It will be released when the
monitoring is finished.
Furthermore, moving task_ctx_data from perf_event_context to task_struct
can reduce complexity and make things clearer. E.g. perf doesn't need to
swap task_ctx_data on optimized context switch path.
This patch set is just the first step. There could be other
optimization/extension on top of this patch set. E.g. for cgroup
profiling, perf just needs to save/store the LBR call stack information
for tasks in specific cgroup. That could reduce the additional space.
Also, the LBR call stack can be available for software events, or allow
even debugging use cases, like LBRs on crash later.
Because of the alignment requirement of Intel Arch LBR, the Kmem cache
is used to allocate the PMU specific data. It's required when child task
allocates the space. Save it in struct perf_ctx_data.
The refcount in struct perf_ctx_data is used to track the users of pmu
specific data.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Alexey Budankov <alexey.budankov@linux.intel.com>
Link: https://lore.kernel.org/r/20250314172700.438923-1-kan.liang@linux.intel.com
|
|
Initially in commit 6891c4509c79 memset() was required to clear a variable
allocated on stack. Commit 2482097c6c0f removed the on stack variable and
retained the memset() despite the fact that the memory is allocated via
kmem_cache_zalloc() and therefore zereoed already.
Drop the redundant memset().
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/all/Z9ctVxwaYOV4A2g4@grain
|
|
The poll man page says POLLRDNORM is equivalent to POLLIN. For poll(),
it seems that if user sets pollfd with POLLRDNORM in userspace, perf_poll
will not return until timeout even if perf_output_wakeup called,
whereas POLLIN returns.
Fixes: 76369139ceb9 ("perf: Split up buffer handling from core code")
Signed-off-by: Tao Chen <chen.dylane@linux.dev>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20250314030036.2543180-1-chen.dylane@linux.dev
|
|
Pinned performance events can enter an error state when they fail to be
scheduled in the context due to a failed constraint or some other conflict
or condition.
In error state these events won't generate any samples anymore and are
silently ignored until they are recovered by PERF_EVENT_IOC_ENABLE,
or the condition can also change so that they can be scheduled in.
Tooling should be allowed to know about the state change, but
currently there's no mechanism to notify tooling when events enter
an error state.
One way to do this is to issue a POLLHUP event to poll(2) to handle this.
Reading events in an error state would return 0 (EOF) and it matches to
the behavior of POLLHUP according to the man page.
Tooling should remove the fd of the event from pollfd after getting
POLLHUP, otherwise it'll be returned repeatedly.
[ mingo: Clarified the changelog ]
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20250317061745.1777584-1-namhyung@kernel.org
|
