linux-stable.git/kernel, branch v5.7.17

sched/uclamp: Fix a deadlock when enabling uclamp static key

2020-08-21T11:07:37+00:00

[ Upstream commit e65855a52b479f98674998cb23b21ef5a8144b04 ]

The following splat was caught when setting uclamp value of a task:

  BUG: sleeping function called from invalid context at ./include/linux/percpu-rwsem.h:49

   cpus_read_lock+0x68/0x130
   static_key_enable+0x1c/0x38
   __sched_setscheduler+0x900/0xad8

Fix by ensuring we enable the key outside of the critical section in
__sched_setscheduler()

Fixes: 46609ce22703 ("sched/uclamp: Protect uclamp fast path code with static key")
Signed-off-by: Qais Yousef 
Signed-off-by: Peter Zijlstra (Intel) 
Link: https://lkml.kernel.org/r/20200716110347.19553-4-qais.yousef@arm.com
Signed-off-by: Sasha Levin

sched/uclamp: Protect uclamp fast path code with static key

2020-08-21T11:07:33+00:00

[ Upstream commit 46609ce227039fd192e0ecc7d940bed587fd2c78 ]

There is a report that when uclamp is enabled, a netperf UDP test
regresses compared to a kernel compiled without uclamp.

https://lore.kernel.org/lkml/20200529100806.GA3070@suse.de/

While investigating the root cause, there were no sign that the uclamp
code is doing anything particularly expensive but could suffer from bad
cache behavior under certain circumstances that are yet to be
understood.

https://lore.kernel.org/lkml/20200616110824.dgkkbyapn3io6wik@e107158-lin/

To reduce the pressure on the fast path anyway, add a static key that is
by default will skip executing uclamp logic in the
enqueue/dequeue_task() fast path until it's needed.

As soon as the user start using util clamp by:

	1. Changing uclamp value of a task with sched_setattr()
	2. Modifying the default sysctl_sched_util_clamp_{min, max}
	3. Modifying the default cpu.uclamp.{min, max} value in cgroup

We flip the static key now that the user has opted to use util clamp.
Effectively re-introducing uclamp logic in the enqueue/dequeue_task()
fast path. It stays on from that point forward until the next reboot.

This should help minimize the effect of util clamp on workloads that
don't need it but still allow distros to ship their kernels with uclamp
compiled in by default.

SCHED_WARN_ON() in uclamp_rq_dec_id() was removed since now we can end
up with unbalanced call to uclamp_rq_dec_id() if we flip the key while
a task is running in the rq. Since we know it is harmless we just
quietly return if we attempt a uclamp_rq_dec_id() when
rq->uclamp[].bucket[].tasks is 0.

In schedutil, we introduce a new uclamp_is_enabled() helper which takes
the static key into account to ensure RT boosting behavior is retained.

The following results demonstrates how this helps on 2 Sockets Xeon E5
2x10-Cores system.

                                   nouclamp                 uclamp      uclamp-static-key
Hmean     send-64         162.43 (   0.00%)      157.84 *  -2.82%*      163.39 *   0.59%*
Hmean     send-128        324.71 (   0.00%)      314.78 *  -3.06%*      326.18 *   0.45%*
Hmean     send-256        641.55 (   0.00%)      628.67 *  -2.01%*      648.12 *   1.02%*
Hmean     send-1024      2525.28 (   0.00%)     2448.26 *  -3.05%*     2543.73 *   0.73%*
Hmean     send-2048      4836.14 (   0.00%)     4712.08 *  -2.57%*     4867.69 *   0.65%*
Hmean     send-3312      7540.83 (   0.00%)     7425.45 *  -1.53%*     7621.06 *   1.06%*
Hmean     send-4096      9124.53 (   0.00%)     8948.82 *  -1.93%*     9276.25 *   1.66%*
Hmean     send-8192     15589.67 (   0.00%)    15486.35 *  -0.66%*    15819.98 *   1.48%*
Hmean     send-16384    26386.47 (   0.00%)    25752.25 *  -2.40%*    26773.74 *   1.47%*

The perf diff between nouclamp and uclamp-static-key when uclamp is
disabled in the fast path:

     8.73%     -1.55%  [kernel.kallsyms]        [k] try_to_wake_up
     0.07%     +0.04%  [kernel.kallsyms]        [k] deactivate_task
     0.13%     -0.02%  [kernel.kallsyms]        [k] activate_task

The diff between nouclamp and uclamp-static-key when uclamp is enabled
in the fast path:

     8.73%     -0.72%  [kernel.kallsyms]        [k] try_to_wake_up
     0.13%     +0.39%  [kernel.kallsyms]        [k] activate_task
     0.07%     +0.38%  [kernel.kallsyms]        [k] deactivate_task

Fixes: 69842cba9ace ("sched/uclamp: Add CPU's clamp buckets refcounting")
Reported-by: Mel Gorman 
Signed-off-by: Qais Yousef 
Signed-off-by: Peter Zijlstra (Intel) 
Tested-by: Lukasz Luba 
Link: https://lkml.kernel.org/r/20200630112123.12076-3-qais.yousef@arm.com
Signed-off-by: Sasha Levin

module: Correctly truncate sysfs sections output

2020-08-21T11:07:29+00:00

commit 11990a5bd7e558e9203c1070fc52fb6f0488e75b upstream.

The only-root-readable /sys/module/$module/sections/$section files
did not truncate their output to the available buffer size. While most
paths into the kernfs read handlers end up using PAGE_SIZE buffers,
it's possible to get there through other paths (e.g. splice, sendfile).
Actually limit the output to the "count" passed into the read function,
and report it back correctly. *sigh*

Reported-by: kernel test robot 
Link: https://lore.kernel.org/lkml/20200805002015.GE23458@shao2-debian
Fixes: ed66f991bb19 ("module: Refactor section attr into bin attribute")
Cc: stable@vger.kernel.org
Reviewed-by: Greg Kroah-Hartman 
Acked-by: Jessica Yu 
Signed-off-by: Kees Cook 
Signed-off-by: Greg Kroah-Hartman

tracing: Use trace_sched_process_free() instead of exit() for pid tracing

2020-08-21T11:07:28+00:00

commit afcab636657421f7ebfa0783a91f90256bba0091 upstream.

On exit, if a process is preempted after the trace_sched_process_exit()
tracepoint but before the process is done exiting, then when it gets
scheduled in, the function tracers will not filter it properly against the
function tracing pid filters.

That is because the function tracing pid filters hooks to the
sched_process_exit() tracepoint to remove the exiting task's pid from the
filter list. Because the filtering happens at the sched_switch tracepoint,
when the exiting task schedules back in to finish up the exit, it will no
longer be in the function pid filtering tables.

This was noticeable in the notrace self tests on a preemptable kernel, as
the tests would fail as it exits and preempted after being taken off the
notrace filter table and on scheduling back in it would not be in the
notrace list, and then the ending of the exit function would trace. The test
detected this and would fail.

Cc: stable@vger.kernel.org
Cc: Namhyung Kim 
Fixes: 1e10486ffee0a ("ftrace: Add 'function-fork' trace option")
Fixes: c37775d57830a ("tracing: Add infrastructure to allow set_event_pid to follow children"
Signed-off-by: Steven Rostedt (VMware) 
Signed-off-by: Greg Kroah-Hartman

tracing/hwlat: Honor the tracing_cpumask

2020-08-21T11:07:28+00:00

commit 96b4833b6827a62c295b149213c68b559514c929 upstream.

In calculation of the cpu mask for the hwlat kernel thread, the wrong
cpu mask is used instead of the tracing_cpumask, this causes the
tracing/tracing_cpumask useless for hwlat tracer. Fixes it.

Link: https://lkml.kernel.org/r/20200730082318.42584-2-haokexin@gmail.com

Cc: Ingo Molnar 
Cc: stable@vger.kernel.org
Fixes: 0330f7aa8ee6 ("tracing: Have hwlat trace migrate across tracing_cpumask CPUs")
Signed-off-by: Kevin Hao 
Signed-off-by: Steven Rostedt (VMware) 
Signed-off-by: Greg Kroah-Hartman

kprobes: Fix NULL pointer dereference at kprobe_ftrace_handler

2020-08-21T11:07:28+00:00

commit 0cb2f1372baa60af8456388a574af6133edd7d80 upstream.

We found a case of kernel panic on our server. The stack trace is as
follows(omit some irrelevant information):

  BUG: kernel NULL pointer dereference, address: 0000000000000080
  RIP: 0010:kprobe_ftrace_handler+0x5e/0xe0
  RSP: 0018:ffffb512c6550998 EFLAGS: 00010282
  RAX: 0000000000000000 RBX: ffff8e9d16eea018 RCX: 0000000000000000
  RDX: ffffffffbe1179c0 RSI: ffffffffc0535564 RDI: ffffffffc0534ec0
  RBP: ffffffffc0534ec1 R08: ffff8e9d1bbb0f00 R09: 0000000000000004
  R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
  R13: ffff8e9d1f797060 R14: 000000000000bacc R15: ffff8e9ce13eca00
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 0000000000000080 CR3: 00000008453d0005 CR4: 00000000003606e0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  Call Trace:
   
   ftrace_ops_assist_func+0x56/0xe0
   ftrace_call+0x5/0x34
   tcpa_statistic_send+0x5/0x130 [ttcp_engine]

The tcpa_statistic_send is the function being kprobed. After analysis,
the root cause is that the fourth parameter regs of kprobe_ftrace_handler
is NULL. Why regs is NULL? We use the crash tool to analyze the kdump.

  crash> dis tcpa_statistic_send -r
         : callq 0xffffffffbd8018c0 

The tcpa_statistic_send calls ftrace_caller instead of ftrace_regs_caller.
So it is reasonable that the fourth parameter regs of kprobe_ftrace_handler
is NULL. In theory, we should call the ftrace_regs_caller instead of the
ftrace_caller. After in-depth analysis, we found a reproducible path.

  Writing a simple kernel module which starts a periodic timer. The
  timer's handler is named 'kprobe_test_timer_handler'. The module
  name is kprobe_test.ko.

  1) insmod kprobe_test.ko
  2) bpftrace -e 'kretprobe:kprobe_test_timer_handler {}'
  3) echo 0 > /proc/sys/kernel/ftrace_enabled
  4) rmmod kprobe_test
  5) stop step 2) kprobe
  6) insmod kprobe_test.ko
  7) bpftrace -e 'kretprobe:kprobe_test_timer_handler {}'

We mark the kprobe as GONE but not disarm the kprobe in the step 4).
The step 5) also do not disarm the kprobe when unregister kprobe. So
we do not remove the ip from the filter. In this case, when the module
loads again in the step 6), we will replace the code to ftrace_caller
via the ftrace_module_enable(). When we register kprobe again, we will
not replace ftrace_caller to ftrace_regs_caller because the ftrace is
disabled in the step 3). So the step 7) will trigger kernel panic. Fix
this problem by disarming the kprobe when the module is going away.

Link: https://lkml.kernel.org/r/20200728064536.24405-1-songmuchun@bytedance.com

Cc: stable@vger.kernel.org
Fixes: ae6aa16fdc16 ("kprobes: introduce ftrace based optimization")
Acked-by: Masami Hiramatsu 
Signed-off-by: Muchun Song 
Co-developed-by: Chengming Zhou 
Signed-off-by: Chengming Zhou 
Signed-off-by: Steven Rostedt (VMware) 
Signed-off-by: Greg Kroah-Hartman

ftrace: Setup correct FTRACE_FL_REGS flags for module

2020-08-21T11:07:28+00:00

commit 8a224ffb3f52b0027f6b7279854c71a31c48fc97 upstream.

When module loaded and enabled, we will use __ftrace_replace_code
for module if any ftrace_ops referenced it found. But we will get
wrong ftrace_addr for module rec in ftrace_get_addr_new, because
rec->flags has not been setup correctly. It can cause the callback
function of a ftrace_ops has FTRACE_OPS_FL_SAVE_REGS to be called
with pt_regs set to NULL.
So setup correct FTRACE_FL_REGS flags for rec when we call
referenced_filters to find ftrace_ops references it.

Link: https://lkml.kernel.org/r/20200728180554.65203-1-zhouchengming@bytedance.com

Cc: stable@vger.kernel.org
Fixes: 8c4f3c3fa9681 ("ftrace: Check module functions being traced on reload")
Signed-off-by: Chengming Zhou 
Signed-off-by: Muchun Song 
Signed-off-by: Steven Rostedt (VMware) 
Signed-off-by: Greg Kroah-Hartman

pidfd: Add missing sock updates for pidfd_getfd()

2020-08-21T11:07:25+00:00

commit 4969f8a073977123504609d7310b42a588297aa4 upstream.

The sock counting (sock_update_netprioidx() and sock_update_classid())
was missing from pidfd's implementation of received fd installation. Add
a call to the new __receive_sock() helper.

Cc: Christian Brauner 
Cc: Christoph Hellwig 
Cc: Sargun Dhillon 
Cc: Jakub Kicinski 
Cc: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: stable@vger.kernel.org
Fixes: 8649c322f75c ("pid: Implement pidfd_getfd syscall")
Signed-off-by: Kees Cook 
Signed-off-by: Greg Kroah-Hartman

genirq/PM: Always unlock IRQ descriptor in rearm_wake_irq()

2020-08-21T11:07:17+00:00

commit e27b1636e9337d1a1d174b191e53d0f86421a822 upstream.

rearm_wake_irq() does not unlock the irq descriptor if the interrupt
is not suspended or if wakeup is not enabled on it.

Restucture the exit conditions so the unlock is always ensured.

Fixes: 3a79bc63d9075 ("PCI: irq: Introduce rearm_wake_irq()")
Signed-off-by: Guenter Roeck 
Signed-off-by: Thomas Gleixner 
Acked-by: Rafael J. Wysocki 
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20200811180001.80203-1-linux@roeck-us.net
Signed-off-by: Greg Kroah-Hartman

genirq/affinity: Make affinity setting if activated opt-in

2020-08-21T11:07:17+00:00

commit f0c7baca180046824e07fc5f1326e83a8fd150c7 upstream.

John reported that on a RK3288 system the perf per CPU interrupts are all
affine to CPU0 and provided the analysis:

 "It looks like what happens is that because the interrupts are not per-CPU
  in the hardware, armpmu_request_irq() calls irq_force_affinity() while
  the interrupt is deactivated and then request_irq() with IRQF_PERCPU |
  IRQF_NOBALANCING.

  Now when irq_startup() runs with IRQ_STARTUP_NORMAL, it calls
  irq_setup_affinity() which returns early because IRQF_PERCPU and
  IRQF_NOBALANCING are set, leaving the interrupt on its original CPU."

This was broken by the recent commit which blocked interrupt affinity
setting in hardware before activation of the interrupt. While this works in
general, it does not work for this particular case. As contrary to the
initial analysis not all interrupt chip drivers implement an activate
callback, the safe cure is to make the deferred interrupt affinity setting
at activation time opt-in.

Implement the necessary core logic and make the two irqchip implementations
for which this is required opt-in. In hindsight this would have been the
right thing to do, but ...

Fixes: baedb87d1b53 ("genirq/affinity: Handle affinity setting on inactive interrupts correctly")
Reported-by: John Keeping 
Signed-off-by: Thomas Gleixner 
Tested-by: Marc Zyngier 
Acked-by: Marc Zyngier 
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/87blk4tzgm.fsf@nanos.tec.linutronix.de
Signed-off-by: Greg Kroah-Hartman