| Age | Commit message (Collapse) | Author |
|---|
|
The recent commit: 90b5363acd47 ("sched: Clean up scheduler_ipi()")
got smp_call_function_single_async() subtly wrong. Even though it will
return -EBUSY when trying to re-use a csd, that condition is not
atomic and still requires external serialization.
The change in ttwu_queue_remote() got this wrong.
While on first reading ttwu_queue_remote() has an atomic test-and-set
that appears to serialize the use, the matching 'release' is not in
the right place to actually guarantee this serialization.
The actual race is vs the sched_ttwu_pending() call in the idle loop;
that can run the wakeup-list without consuming the CSD.
Instead of trying to chain the lists, merge them.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200526161908.129371594@infradead.org
|
|
In preparation of removing rq->wake_list, replace the
!list_empty(rq->wake_list) with rq->ttwu_pending. This is not fully
equivalent as this new variable is racy.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200526161908.070399698@infradead.org
|
|
Currently irq_work_queue_on() will issue an unconditional
arch_send_call_function_single_ipi() and has the handler do
irq_work_run().
This is unfortunate in that it makes the IPI handler look at a second
cacheline and it misses the opportunity to avoid the IPI. Instead note
that struct irq_work and struct __call_single_data are very similar in
layout, so use a few bits in the flags word to encode a type and stick
the irq_work on the call_single_queue list.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200526161908.011635912@infradead.org
|
|
Just like the ttwu_queue_remote() IPI, make use of _TIF_POLLING_NRFLAG
to avoid sending IPIs to idle CPUs.
[ mingo: Fix UP build bug. ]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200526161907.953304789@infradead.org
|
|
This ensures flush_smp_call_function_queue() is strictly about
call_single_queue.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200526161907.895109676@infradead.org
|
|
The call_single_queue can contain (two) different callbacks,
synchronous and asynchronous. The current interrupt handler runs them
in-order, which means that remote CPUs that are waiting for their
synchronous call can be delayed by running asynchronous callbacks.
Rework the interrupt handler to first run the synchonous callbacks.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20200526161907.836818381@infradead.org
|
|
The recent commit: 90b5363acd47 ("sched: Clean up scheduler_ipi()")
got smp_call_function_single_async() subtly wrong. Even though it will
return -EBUSY when trying to re-use a csd, that condition is not
atomic and still requires external serialization.
The change in kick_ilb() got this wrong.
While on first reading kick_ilb() has an atomic test-and-set that
appears to serialize the use, the matching 'release' is not in the
right place to actually guarantee this serialization.
Rework the nohz_idle_balance() trigger so that the release is in the
IPI callback and thus guarantees the required serialization for the
CSD.
Fixes: 90b5363acd47 ("sched: Clean up scheduler_ipi()")
Reported-by: Qian Cai <cai@lca.pw>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Cc: mgorman@techsingularity.net
Link: https://lore.kernel.org/r/20200526161907.778543557@infradead.org
|
|
We are going to rely on the loosening of RCU callback semantics,
introduced by this commit:
806f04e9fd2c: ("rcu: Allow for smp_call_function() running callbacks from idle")
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Current RCU hard relies on smp_call_function() callbacks running from
interrupt context. A pending optimization is going to break that, it
will allow idle CPUs to run the callbacks from the idle loop. This
avoids raising the IPI on the requesting CPU and avoids handling an
exception on the receiving CPU.
Change rcu_is_cpu_rrupt_from_idle() to also accept task context,
provided it is the idle task.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Link: https://lore.kernel.org/r/20200527171236.GC706495@hirez.programming.kicks-ass.net
|
|
Stefano reported a crash with using SQPOLL with io_uring:
BUG: kernel NULL pointer dereference, address: 00000000000003b0
CPU: 2 PID: 1307 Comm: io_uring-sq Not tainted 5.7.0-rc7 #11
RIP: 0010:task_numa_work+0x4f/0x2c0
Call Trace:
task_work_run+0x68/0xa0
io_sq_thread+0x252/0x3d0
kthread+0xf9/0x130
ret_from_fork+0x35/0x40
which is task_numa_work() oopsing on current->mm being NULL.
The task work is queued by task_tick_numa(), which checks if current->mm is
NULL at the time of the call. But this state isn't necessarily persistent,
if the kthread is using use_mm() to temporarily adopt the mm of a task.
Change the task_tick_numa() check to exclude kernel threads in general,
as it doesn't make sense to attempt ot balance for kthreads anyway.
Reported-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/865de121-8190-5d30-ece5-3b097dc74431@kernel.dk
|
|
The previous commit:
c6e7bd7afaeb: ("sched/core: Optimize ttwu() spinning on p->on_cpu")
avoids spinning on p->on_rq when the task is descheduling, but only if the
wakee is on a CPU that does not share cache with the waker.
This patch offloads the activation of the wakee to the CPU that is about to
go idle if the task is the only one on the runqueue. This potentially allows
the waker task to continue making progress when the wakeup is not strictly
synchronous.
This is very obvious with netperf UDP_STREAM running on localhost. The
waker is sending packets as quickly as possible without waiting for any
reply. It frequently wakes the server for the processing of packets and
when netserver is using local memory, it quickly completes the processing
and goes back to idle. The waker often observes that netserver is on_rq
and spins excessively leading to a drop in throughput.
This is a comparison of 5.7-rc6 against "sched: Optimize ttwu() spinning
on p->on_cpu" and against this patch labeled vanilla, optttwu-v1r1 and
localwakelist-v1r2 respectively.
5.7.0-rc6 5.7.0-rc6 5.7.0-rc6
vanilla optttwu-v1r1 localwakelist-v1r2
Hmean send-64 251.49 ( 0.00%) 258.05 * 2.61%* 305.59 * 21.51%*
Hmean send-128 497.86 ( 0.00%) 519.89 * 4.43%* 600.25 * 20.57%*
Hmean send-256 944.90 ( 0.00%) 997.45 * 5.56%* 1140.19 * 20.67%*
Hmean send-1024 3779.03 ( 0.00%) 3859.18 * 2.12%* 4518.19 * 19.56%*
Hmean send-2048 7030.81 ( 0.00%) 7315.99 * 4.06%* 8683.01 * 23.50%*
Hmean send-3312 10847.44 ( 0.00%) 11149.43 * 2.78%* 12896.71 * 18.89%*
Hmean send-4096 13436.19 ( 0.00%) 13614.09 ( 1.32%) 15041.09 * 11.94%*
Hmean send-8192 22624.49 ( 0.00%) 23265.32 * 2.83%* 24534.96 * 8.44%*
Hmean send-16384 34441.87 ( 0.00%) 36457.15 * 5.85%* 35986.21 * 4.48%*
Note that this benefit is not universal to all wakeups, it only applies
to the case where the waker often spins on p->on_rq.
The impact can be seen from a "perf sched latency" report generated from
a single iteration of one packet size:
-----------------------------------------------------------------------------------------------------------------
Task | Runtime ms | Switches | Average delay ms | Maximum delay ms | Maximum delay at |
-----------------------------------------------------------------------------------------------------------------
vanilla
netperf:4337 | 21709.193 ms | 2932 | avg: 0.002 ms | max: 0.041 ms | max at: 112.154512 s
netserver:4338 | 14629.459 ms | 5146990 | avg: 0.001 ms | max: 1615.864 ms | max at: 140.134496 s
localwakelist-v1r2
netperf:4339 | 29789.717 ms | 2460 | avg: 0.002 ms | max: 0.059 ms | max at: 138.205389 s
netserver:4340 | 18858.767 ms | 7279005 | avg: 0.001 ms | max: 0.362 ms | max at: 135.709683 s
-----------------------------------------------------------------------------------------------------------------
Note that the average wakeup delay is quite small on both the vanilla
kernel and with the two patches applied. However, there are significant
outliers with the vanilla kernel with the maximum one measured as 1615
milliseconds with a vanilla kernel but never worse than 0.362 ms with
both patches applied and a much higher rate of context switching.
Similarly a separate profile of cycles showed that 2.83% of all cycles
were spent in try_to_wake_up() with almost half of the cycles spent
on spinning on p->on_rq. With the two patches, the percentage of cycles
spent in try_to_wake_up() drops to 1.13%
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Jirka Hladky <jhladky@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: valentin.schneider@arm.com
Cc: Hillf Danton <hdanton@sina.com>
Cc: Rik van Riel <riel@surriel.com>
Link: https://lore.kernel.org/r/20200524202956.27665-3-mgorman@techsingularity.net
|
|
Both Rik and Mel reported seeing ttwu() spend significant time on:
smp_cond_load_acquire(&p->on_cpu, !VAL);
Attempt to avoid this by queueing the wakeup on the CPU that owns the
p->on_cpu value. This will then allow the ttwu() to complete without
further waiting.
Since we run schedule() with interrupts disabled, the IPI is
guaranteed to happen after p->on_cpu is cleared, this is what makes it
safe to queue early.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Jirka Hladky <jhladky@redhat.com>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: valentin.schneider@arm.com
Cc: Hillf Danton <hdanton@sina.com>
Cc: Rik van Riel <riel@surriel.com>
Link: https://lore.kernel.org/r/20200524202956.27665-2-mgorman@techsingularity.net
|
|
When users write some huge number into cpu.cfs_quota_us or
cpu.rt_runtime_us, overflow might happen during to_ratio() shifts of
schedulable checks.
to_ratio() could be altered to avoid unnecessary internal overflow, but
min_cfs_quota_period is less than 1 << BW_SHIFT, so a cutoff would still
be needed. Set a cap MAX_BW for cfs_quota_us and rt_runtime_us to
prevent overflow.
Signed-off-by: Huaixin Chang <changhuaixin@linux.alibaba.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ben Segall <bsegall@google.com>
Link: https://lkml.kernel.org/r/20200425105248.60093-1-changhuaixin@linux.alibaba.com
|
|
The user_mode(task_pt_regs(tsk)) always return true for
user thread, and false for kernel thread. So it means that
the cpuacct.usage_sys is the time that kernel thread uses
not the time that thread uses in the kernel mode. We can
try get_irq_regs() first, if it is NULL, then we can fall
back to task_pt_regs().
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200420070453.76815-1-songmuchun@bytedance.com
|
|
The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by
this change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200507192141.GA16183@embeddedor
|
|
update_tg_cfs_*() propagate the impact of the attach/detach of an entity
down into the cfs_rq hierarchy and must keep the sync with the current pelt
window.
Even if we can't sync child cfs_rq and its group se, we can sync the group
se and its parent cfs_rq with current position in the PELT window. In fact,
we must keep them sync in order to stay also synced with others entities
and group entities that are already attached to the cfs_rq.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200506155301.14288-1-vincent.guittot@linaro.org
|
|
The cpuacct_charge() and cpuacct_account_field() are called with
rq->lock held, and this means preemption(and IRQs) are indeed
disabled, so it is safe to use __this_cpu_*() to allow for better
code-generation.
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200507031039.32615-1-songmuchun@bytedance.com
|
|
enqueue_task_fair jumps to enqueue_throttle label when cfs_rq_of(se) is
throttled which means that se can't be NULL in such case and we can move
the label after the if (!se) statement. Futhermore, the latter can be
removed because se is always NULL when reaching this point.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Phil Auld <pauld@redhat.com>
Link: https://lkml.kernel.org/r/20200513135502.4672-1-vincent.guittot@linaro.org
|
|
|
|
Although not exactly identical, unthrottle_cfs_rq() and enqueue_task_fair()
are quite close and follow the same sequence for enqueuing an entity in the
cfs hierarchy. Modify unthrottle_cfs_rq() to use the same pattern as
enqueue_task_fair(). This fixes a problem already faced with the latter and
add an optimization in the last for_each_sched_entity loop.
Fixes: fe61468b2cb (sched/fair: Fix enqueue_task_fair warning)
Reported-by Tao Zhou <zohooouoto@zoho.com.cn>
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Phil Auld <pauld@redhat.com>
Reviewed-by: Ben Segall <bsegall@google.com>
Link: https://lkml.kernel.org/r/20200513135528.4742-1-vincent.guittot@linaro.org
|
|
The intention of commit 96e74ebf8d59 ("sched/debug: Add task uclamp
values to SCHED_DEBUG procfs") was to print requested and effective
task uclamp values. The requested values printed are read from p->uclamp,
which holds the last effective values. Fix this by printing the values
from p->uclamp_req.
Fixes: 96e74ebf8d59 ("sched/debug: Add task uclamp values to SCHED_DEBUG procfs")
Signed-off-by: Pavankumar Kondeti <pkondeti@codeaurora.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Tested-by: Valentin Schneider <valentin.schneider@arm.com>
Link: https://lkml.kernel.org/r/1589115401-26391-1-git-send-email-pkondeti@codeaurora.org
|
|
sched/fair: Fix enqueue_task_fair warning some more
The recent patch, fe61468b2cb (sched/fair: Fix enqueue_task_fair warning)
did not fully resolve the issues with the rq->tmp_alone_branch !=
&rq->leaf_cfs_rq_list warning in enqueue_task_fair. There is a case where
the first for_each_sched_entity loop exits due to on_rq, having incompletely
updated the list. In this case the second for_each_sched_entity loop can
further modify se. The later code to fix up the list management fails to do
what is needed because se does not point to the sched_entity which broke out
of the first loop. The list is not fixed up because the throttled parent was
already added back to the list by a task enqueue in a parallel child hierarchy.
Address this by calling list_add_leaf_cfs_rq if there are throttled parents
while doing the second for_each_sched_entity loop.
Fixes: fe61468b2cb ("sched/fair: Fix enqueue_task_fair warning")
Suggested-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Phil Auld <pauld@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/20200512135222.GC2201@lorien.usersys.redhat.com
|
|
Same as rcu_is_watching() but without the preempt_disable/enable() pair
inside the function. It is merked noinstr so it ends up in the
non-instrumentable text section.
This is useful for non-preemptible code especially in the low level entry
section. Using rcu_is_watching() there results in a call to the
preempt_schedule_notrace() thunk which triggers noinstr section warnings in
objtool.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200512213810.518709291@linutronix.de
|
|
Interrupts and exceptions invoke rcu_irq_enter() on entry and need to
invoke rcu_irq_exit() before they either return to the interrupted code or
invoke the scheduler due to preemption.
The general assumption is that RCU idle code has to have preemption
disabled so that a return from interrupt cannot schedule. So the return
from interrupt code invokes rcu_irq_exit() and preempt_schedule_irq().
If there is any imbalance in the rcu_irq/nmi* invocations or RCU idle code
had preemption enabled then this goes unnoticed until the CPU goes idle or
some other RCU check is executed.
Provide rcu_irq_exit_preempt() which can be invoked from the
interrupt/exception return code in case that preemption is enabled. It
invokes rcu_irq_exit() and contains a few sanity checks in case that
CONFIG_PROVE_RCU is enabled to catch such issues directly.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Paul E. McKenney <paulmck@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200505134904.364456424@linutronix.de
|
|
The rcu_nmi_enter_common() and rcu_nmi_exit_common() functions take an
"irq" parameter that indicates whether these functions have been invoked from
an irq handler (irq==true) or an NMI handler (irq==false).
However, recent changes have applied notrace to a few critical functions
such that rcu_nmi_enter_common() and rcu_nmi_exit_common() many now rely on
in_nmi(). Note that in_nmi() works no differently than before, but rather
that tracing is now prohibited in code regions where in_nmi() would
incorrectly report NMI state.
Therefore remove the "irq" parameter and inline rcu_nmi_enter_common() and
rcu_nmi_exit_common() into rcu_nmi_enter() and rcu_nmi_exit(),
respectively.
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Link: https://lkml.kernel.org/r/20200505134101.617130349@linutronix.de
|
|
These functions are invoked from context tracking and other places in the
low level entry code. Move them into the .noinstr.text section to exclude
them from instrumentation.
Mark the places which are safe to invoke traceable functions with
instrumentation_begin/end() so objtool won't complain.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Acked-by: Paul E. McKenney <paulmck@kernel.org>
Link: https://lkml.kernel.org/r/20200505134100.575356107@linutronix.de
|
|
SuperH is the last remaining user of arch_ftrace_nmi_{enter,exit}(),
remove it from the generic code and into the SuperH code.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Rich Felker <dalias@libc.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Link: https://lkml.kernel.org/r/20200505134101.248881738@linutronix.de
|
|
These functions are called {early,late} in nmi_{enter,exit} and should
not be traced or probed. They are also puny, so 'inline' them.
Reported-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Link: https://lkml.kernel.org/r/20200505134101.048523500@linutronix.de
|
|
It happens early in nmi_enter(), no tracing, probing or other funnies
allowed. Specifically as nmi_enter() will be used in do_debug(), which
would cause recursive exceptions when kprobed.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Link: https://lkml.kernel.org/r/20200505134101.139720912@linutronix.de
|
|
There is plenty of space in the printk_context variable. Reserve one byte
there for the NMI context to be on the safe side.
It should never overflow. The BUG_ON(in_nmi() == NMI_MASK) in nmi_enter()
will trigger much earlier.
Signed-off-by: Petr Mladek <pmladek@suse.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Link: https://lkml.kernel.org/r/20200505134100.681374113@linutronix.de
|
|
Get the noinstr section and annotation markers to base the RCU parts on.
|
|
Pull networking fixes from David Miller:
1) Fix sk_psock reference count leak on receive, from Xiyu Yang.
2) CONFIG_HNS should be invisible, from Geert Uytterhoeven.
3) Don't allow locking route MTUs in ipv6, RFCs actually forbid this,
from Maciej Żenczykowski.
4) ipv4 route redirect backoff wasn't actually enforced, from Paolo
Abeni.
5) Fix netprio cgroup v2 leak, from Zefan Li.
6) Fix infinite loop on rmmod in conntrack, from Florian Westphal.
7) Fix tcp SO_RCVLOWAT hangs, from Eric Dumazet.
8) Various bpf probe handling fixes, from Daniel Borkmann.
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (68 commits)
selftests: mptcp: pm: rm the right tmp file
dpaa2-eth: properly handle buffer size restrictions
bpf: Restrict bpf_trace_printk()'s %s usage and add %pks, %pus specifier
bpf: Add bpf_probe_read_{user, kernel}_str() to do_refine_retval_range
bpf: Restrict bpf_probe_read{, str}() only to archs where they work
MAINTAINERS: Mark networking drivers as Maintained.
ipmr: Add lockdep expression to ipmr_for_each_table macro
ipmr: Fix RCU list debugging warning
drivers: net: hamradio: Fix suspicious RCU usage warning in bpqether.c
net: phy: broadcom: fix BCM54XX_SHD_SCR3_TRDDAPD value for BCM54810
tcp: fix error recovery in tcp_zerocopy_receive()
MAINTAINERS: Add Jakub to networking drivers.
MAINTAINERS: another add of Karsten Graul for S390 networking
drivers: ipa: fix typos for ipa_smp2p structure doc
pppoe: only process PADT targeted at local interfaces
selftests/bpf: Enforce returning 0 for fentry/fexit programs
bpf: Enforce returning 0 for fentry/fexit progs
net: stmmac: fix num_por initialization
security: Fix the default value of secid_to_secctx hook
libbpf: Fix register naming in PT_REGS s390 macros
...
|
|
Usage of plain %s conversion specifier in bpf_trace_printk() suffers from the
very same issue as bpf_probe_read{,str}() helpers, that is, it is broken on
archs with overlapping address ranges.
While the helpers have been addressed through work in 6ae08ae3dea2 ("bpf: Add
probe_read_{user, kernel} and probe_read_{user, kernel}_str helpers"), we need
an option for bpf_trace_printk() as well to fix it.
Similarly as with the helpers, force users to make an explicit choice by adding
%pks and %pus specifier to bpf_trace_printk() which will then pick the corresponding
strncpy_from_unsafe*() variant to perform the access under KERNEL_DS or USER_DS.
The %pk* (kernel specifier) and %pu* (user specifier) can later also be extended
for other objects aside strings that are probed and printed under tracing, and
reused out of other facilities like bpf_seq_printf() or BTF based type printing.
Existing behavior of %s for current users is still kept working for archs where it
is not broken and therefore gated through CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE.
For archs not having this property we fall-back to pick probing under KERNEL_DS as
a sensible default.
Fixes: 8d3b7dce8622 ("bpf: add support for %s specifier to bpf_trace_printk()")
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Reported-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Link: https://lore.kernel.org/bpf/20200515101118.6508-4-daniel@iogearbox.net
|
|
Given bpf_probe_read{,str}() BPF helpers are now only available under
CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE, we need to add the drop-in
replacements of bpf_probe_read_{kernel,user}_str() to do_refine_retval_range()
as well to avoid hitting the same issue as in 849fa50662fbc ("bpf/verifier:
refine retval R0 state for bpf_get_stack helper").
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200515101118.6508-3-daniel@iogearbox.net
|
|
Given the legacy bpf_probe_read{,str}() BPF helpers are broken on archs
with overlapping address ranges, we should really take the next step to
disable them from BPF use there.
To generally fix the situation, we've recently added new helper variants
bpf_probe_read_{user,kernel}() and bpf_probe_read_{user,kernel}_str().
For details on them, see 6ae08ae3dea2 ("bpf: Add probe_read_{user, kernel}
and probe_read_{user,kernel}_str helpers").
Given bpf_probe_read{,str}() have been around for ~5 years by now, there
are plenty of users at least on x86 still relying on them today, so we
cannot remove them entirely w/o breaking the BPF tracing ecosystem.
However, their use should be restricted to archs with non-overlapping
address ranges where they are working in their current form. Therefore,
move this behind a CONFIG_ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE and
have x86, arm64, arm select it (other archs supporting it can follow-up
on it as well).
For the remaining archs, they can workaround easily by relying on the
feature probe from bpftool which spills out defines that can be used out
of BPF C code to implement the drop-in replacement for old/new kernels
via: bpftool feature probe macro
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org>
Acked-by: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/bpf/20200515101118.6508-2-daniel@iogearbox.net
|
|
Currently, tracing/fentry and tracing/fexit prog
return values are not enforced. In trampoline codes,
the fentry/fexit prog return values are ignored.
Let us enforce it to be 0 to avoid confusion and
allows potential future extension.
This patch also explicitly added return value
checking for tracing/raw_tp, tracing/fmod_ret,
and freplace programs such that these program
return values can be anything. The purpose are
two folds:
1. to make it explicit about return value expectations
for these programs in verifier.
2. for tracing prog_type, if a future attach type
is added, the default is -ENOTSUPP which will
enforce to specify return value ranges explicitly.
Fixes: fec56f5890d9 ("bpf: Introduce BPF trampoline")
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Link: https://lore.kernel.org/bpf/20200514053206.1298415-1-yhs@fb.com
|
|
mmap() subsystem allows user-space application to memory-map region with
initial page offset. This wasn't taken into account in initial implementation
of BPF array memory-mapping. This would result in wrong pages, not taking into
account requested page shift, being memory-mmaped into user-space. This patch
fixes this gap and adds a test for such scenario.
Fixes: fc9702273e2e ("bpf: Add mmap() support for BPF_MAP_TYPE_ARRAY")
Signed-off-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/bpf/20200512235925.3817805-1-andriin@fb.com
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux
Pull thread fix from Christian Brauner:
"This contains a single fix for all exported legacy fork helpers to
block accidental access to clone3() features in the upper 32 bits of
their respective flags arguments.
I got Cced on a glibc issue where someone reported consistent failures
for the legacy clone() syscall on ppc64le when sign extension was
performed (since the clone() syscall in glibc exposes the flags
argument as an int whereas the kernel uses unsigned long).
The legacy clone() syscall is odd in a bunch of ways and here two
things interact:
- First, legacy clone's flag argument is word-size dependent, i.e.
it's an unsigned long whereas most system calls with flag arguments
use int or unsigned int.
- Second, legacy clone() ignores unknown and deprecated flags.
The two of them taken together means that users on 64bit systems can
pass garbage for the upper 32bit of the clone() syscall since forever
and things would just work fine.
The following program compiled on a 64bit kernel prior to v5.7-rc1
will succeed and will fail post v5.7-rc1 with EBADF:
int main(int argc, char *argv[])
{
pid_t pid;
/* Note that legacy clone() has different argument ordering on
* different architectures so this won't work everywhere.
*
* Only set the upper 32 bits.
*/
pid = syscall(__NR_clone, 0xffffffff00000000 | SIGCHLD,
NULL, NULL, NULL, NULL);
if (pid < 0)
exit(EXIT_FAILURE);
if (pid == 0)
exit(EXIT_SUCCESS);
if (wait(NULL) != pid)
exit(EXIT_FAILURE);
exit(EXIT_SUCCESS);
}
Since legacy clone() couldn't be extended this was not a problem so
far and nobody really noticed or cared since nothing in the kernel
ever bothered to look at the upper 32 bits.
But once we introduced clone3() and expanded the flag argument in
struct clone_args to 64 bit we opened this can of worms. With the
first flag-based extension to clone3() making use of the upper 32 bits
of the flag argument we've effectively made it possible for the legacy
clone() syscall to reach clone3() only flags on accident. The sign
extension scenario is just the odd corner-case that we needed to
figure this out.
The reason we just realized this now and not already when we
introduced CLONE_CLEAR_SIGHAND was that CLONE_INTO_CGROUP assumes that
a valid cgroup file descriptor has been given - whereas
CLONE_CLEAR_SIGHAND doesn't need to verify anything. It just silently
resets the signal handlers to SIG_DFL.
So the sign extension (or the user accidently passing garbage for the
upper 32 bits) caused the CLONE_INTO_CGROUP bit to be raised and the
kernel to error out when it didn't find a valid cgroup file
descriptor.
Note, I'm also capping kernel_thread()'s flag argument mainly because
none of the new features make sense for kernel_thread() and we
shouldn't risk them being accidently activated however unlikely. If we
wanted to, we could even make kernel_thread() yell when an unknown
flag has been set which it doesn't do right now. But it's not worth
risking this in a bugfix imho"
* tag 'for-linus-2020-05-13' of git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux:
fork: prevent accidental access to clone3 features
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull more tracing fixes from Steven Rostedt:
"Various tracing fixes:
- Fix a crash when having function tracing and function stack tracing
on the command line.
The ftrace trampolines are created as executable and read only. But
the stack tracer tries to modify them with text_poke() which
expects all kernel text to still be writable at boot. Keep the
trampolines writable at boot, and convert them to read-only with
the rest of the kernel.
- A selftest was triggering in the ring buffer iterator code, that is
no longer valid with the update of keeping the ring buffer writable
while a iterator is reading.
Just bail after three failed attempts to get an event and remove
the warning and disabling of the ring buffer.
- While modifying the ring buffer code, decided to remove all the
unnecessary BUG() calls"
* tag 'trace-v5.7-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
ring-buffer: Remove all BUG() calls
ring-buffer: Don't deactivate the ring buffer on failed iterator reads
x86/ftrace: Have ftrace trampolines turn read-only at the end of system boot up
|
|
There's a lot of checks to make sure the ring buffer is working, and if an
anomaly is detected, it safely shuts itself down. But there's a few cases
that it will call BUG(), which defeats the point of being safe (it crashes
the kernel when an anomaly is found!). There's no reason for them. Switch
them all to either WARN_ON_ONCE() (when no ring buffer descriptor is present),
or to RB_WARN_ON() (when a ring buffer descriptor is present).
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
|
|
If the function tracer is running and the trace file is read (which uses the
ring buffer iterator), the iterator can get in sync with the writes, and
caues it to fail to find a page with content it can read three times. This
causes a warning and deactivation of the ring buffer code.
Looking at the other cases of failure to get an event, it appears that
there's a chance that the writer could cause them too. Since the iterator is
a "best effort" to read the ring buffer if there's an active writer (the
consumer reader is made for this case "see trace_pipe"), if it fails to get
an event after three tries, simply give up and return NULL. Don't warn, nor
disable the ring buffer on this failure.
Link: https://lore.kernel.org/r/20200429090508.GG5770@shao2-debian
Reported-by: kernel test robot <lkp@intel.com>
Fixes: ff84c50cfb4b ("ring-buffer: Do not die if rb_iter_peek() fails more than thrice")
Tested-by: Sven Schnelle <svens@linux.ibm.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
|
|
Booting one of my machines, it triggered the following crash:
Kernel/User page tables isolation: enabled
ftrace: allocating 36577 entries in 143 pages
Starting tracer 'function'
BUG: unable to handle page fault for address: ffffffffa000005c
#PF: supervisor write access in kernel mode
#PF: error_code(0x0003) - permissions violation
PGD 2014067 P4D 2014067 PUD 2015063 PMD 7b253067 PTE 7b252061
Oops: 0003 [#1] PREEMPT SMP PTI
CPU: 0 PID: 0 Comm: swapper Not tainted 5.4.0-test+ #24
Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./To be filled by O.E.M., BIOS SDBLI944.86P 05/08/2007
RIP: 0010:text_poke_early+0x4a/0x58
Code: 34 24 48 89 54 24 08 e8 bf 72 0b 00 48 8b 34 24 48 8b 4c 24 08 84 c0 74 0b 48 89 df f3 a4 48 83 c4 10 5b c3 9c 58 fa 48 89 df <f3> a4 50 9d 48 83 c4 10 5b e9 d6 f9 ff ff
0 41 57 49
RSP: 0000:ffffffff82003d38 EFLAGS: 00010046
RAX: 0000000000000046 RBX: ffffffffa000005c RCX: 0000000000000005
RDX: 0000000000000005 RSI: ffffffff825b9a90 RDI: ffffffffa000005c
RBP: ffffffffa000005c R08: 0000000000000000 R09: ffffffff8206e6e0
R10: ffff88807b01f4c0 R11: ffffffff8176c106 R12: ffffffff8206e6e0
R13: ffffffff824f2440 R14: 0000000000000000 R15: ffffffff8206eac0
FS: 0000000000000000(0000) GS:ffff88807d400000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffffffffa000005c CR3: 0000000002012000 CR4: 00000000000006b0
Call Trace:
text_poke_bp+0x27/0x64
? mutex_lock+0x36/0x5d
arch_ftrace_update_trampoline+0x287/0x2d5
? ftrace_replace_code+0x14b/0x160
? ftrace_update_ftrace_func+0x65/0x6c
__register_ftrace_function+0x6d/0x81
ftrace_startup+0x23/0xc1
register_ftrace_function+0x20/0x37
func_set_flag+0x59/0x77
__set_tracer_option.isra.19+0x20/0x3e
trace_set_options+0xd6/0x13e
apply_trace_boot_options+0x44/0x6d
register_tracer+0x19e/0x1ac
early_trace_init+0x21b/0x2c9
start_kernel+0x241/0x518
? load_ucode_intel_bsp+0x21/0x52
secondary_startup_64+0xa4/0xb0
I was able to trigger it on other machines, when I added to the kernel
command line of both "ftrace=function" and "trace_options=func_stack_trace".
The cause is the "ftrace=function" would register the function tracer
and create a trampoline, and it will set it as executable and
read-only. Then the "trace_options=func_stack_trace" would then update
the same trampoline to include the stack tracer version of the function
tracer. But since the trampoline already exists, it updates it with
text_poke_bp(). The problem is that text_poke_bp() called while
system_state == SYSTEM_BOOTING, it will simply do a memcpy() and not
the page mapping, as it would think that the text is still read-write.
But in this case it is not, and we take a fault and crash.
Instead, lets keep the ftrace trampolines read-write during boot up,
and then when the kernel executable text is set to read-only, the
ftrace trampolines get set to read-only as well.
Link: https://lkml.kernel.org/r/20200430202147.4dc6e2de@oasis.local.home
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: stable@vger.kernel.org
Fixes: 768ae4406a5c ("x86/ftrace: Use text_poke()")
Acked-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing fixes from Steven Rostedt:
"Fixes to previous fixes.
Unfortunately, the last set of fixes introduced some minor bugs:
- The bootconfig apply_xbc() leak fix caused the application to
return a positive number on success, when it should have returned
zero.
- The preempt_irq_delay_thread fix to make the creation code wait for
the kthread to finish to prevent it from executing after module
unload, can now cause the kthread to exit before it even executes
(preventing it to run its tests).
- The fix to the bootconfig that fixed the initrd to remove the
bootconfig from causing the kernel to panic, now prints a warning
that the bootconfig is not found, even when bootconfig is not on
the command line"
* tag 'trace-v5.7-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
bootconfig: Fix to prevent warning message if no bootconfig option
tracing: Wait for preempt irq delay thread to execute
tools/bootconfig: Fix apply_xbc() to return zero on success
|
|
Now that the scheduler IPI is trivial and simple again there is no point to
have the little function out of line. This simplifies the effort of
constraining the instrumentation nicely.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200505134058.453581595@linutronix.de
|
|
The scheduler IPI has grown weird and wonderful over the years, time
for spring cleaning.
Move all the non-trivial stuff out of it and into a regular smp function
call IPI. This then reduces the schedule_ipi() to most of it's former NOP
glory and ensures to keep the interrupt vector lean and mean.
Aside of that avoiding the full irq_enter() in the x86 IPI implementation
is incorrect as scheduler_ipi() can be instrumented. To work around that
scheduler_ipi() had an irq_enter/exit() hack when heavy work was
pending. This is gone now.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Link: https://lkml.kernel.org/r/20200505134058.361859938@linutronix.de
|
|
A bug report was posted that running the preempt irq delay module on a slow
machine, and removing it quickly could lead to the thread created by the
modlue to execute after the module is removed, and this could cause the
kernel to crash. The fix for this was to call kthread_stop() after creating
the thread to make sure it finishes before allowing the module to be
removed.
Now this caused the opposite problem on fast machines. What now happens is
the kthread_stop() can cause the kthread never to execute and the test never
to run. To fix this, add a completion and wait for the kthread to execute,
then wait for it to end.
This issue caused the ftracetest selftests to fail on the preemptirq tests.
Link: https://lore.kernel.org/r/20200510114210.15d9e4af@oasis.local.home
Cc: stable@vger.kernel.org
Fixes: d16a8c31077e ("tracing: Wait for preempt irq delay thread to finish")
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
|
|
We have some rather random rules about when we accept the
"maybe-initialized" warnings, and when we don't.
For example, we consider it unreliable for gcc versions < 4.9, but also
if -O3 is enabled, or if optimizing for size. And then various kernel
config options disabled it, because they know that they trigger that
warning by confusing gcc sufficiently (ie PROFILE_ALL_BRANCHES).
And now gcc-10 seems to be introducing a lot of those warnings too, so
it falls under the same heading as 4.9 did.
At the same time, we have a very straightforward way to _enable_ that
warning when wanted: use "W=2" to enable more warnings.
So stop playing these ad-hoc games, and just disable that warning by
default, with the known and straight-forward "if you want to work on the
extra compiler warnings, use W=123".
Would it be great to have code that is always so obvious that it never
confuses the compiler whether a variable is used initialized or not?
Yes, it would. In a perfect world, the compilers would be smarter, and
our source code would be simpler.
That's currently not the world we live in, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
If a UMH process created by fork_usermode_blob() fails to execute,
a pair of struct file allocated by umh_pipe_setup() will leak.
Under normal conditions, the caller (like bpfilter) needs to manage the
lifetime of the UMH and its two pipes. But when fork_usermode_blob()
fails, the caller doesn't really have a way to know what needs to be
done. It seems better to do the cleanup ourselves in this case.
Fixes: 449325b52b7a ("umh: introduce fork_usermode_blob() helper")
Signed-off-by: Vincent Minet <v.minet@criteo.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Daniel Borkmann says:
====================
pull-request: bpf 2020-05-09
The following pull-request contains BPF updates for your *net* tree.
We've added 4 non-merge commits during the last 9 day(s) which contain
a total of 4 files changed, 11 insertions(+), 6 deletions(-).
The main changes are:
1) Fix msg_pop_data() helper incorrectly setting an sge length in some
cases as well as fixing bpf_tcp_ingress() wrongly accounting bytes
in sg.size, from John Fastabend.
2) Fix to return an -EFAULT error when copy_to_user() of the value
fails in map_lookup_and_delete_elem(), from Wei Yongjun.
3) Fix sk_psock refcnt leak in tcp_bpf_recvmsg(), from Xiyu Yang.
====================
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
