linux-toradex.git/kernel/bpf/task_iter.c, branch v5.12-rc5

bpf: Introduce task_vma bpf_iter

2021-02-12T20:56:53+00:00

Introduce task_vma bpf_iter to print memory information of a process. It
can be used to print customized information similar to /proc//maps.

Current /proc//maps and /proc//smaps provide information of
vma's of a process. However, these information are not flexible enough to
cover all use cases. For example, if a vma cover mixed 2MB pages and 4kB
pages (x86_64), there is no easy way to tell which address ranges are
backed by 2MB pages. task_vma solves the problem by enabling the user to
generate customize information based on the vma (and vma->vm_mm,
vma->vm_file, etc.).

To access the vma safely in the BPF program, task_vma iterator holds
target mmap_lock while calling the BPF program. If the mmap_lock is
contended, task_vma unlocks mmap_lock between iterations to unblock the
writer(s). This lock contention avoidance mechanism is similar to the one
used in show_smaps_rollup().

Signed-off-by: Song Liu 
Signed-off-by: Alexei Starovoitov 
Acked-by: Yonghong Song 
Link: https://lore.kernel.org/bpf/20210212183107.50963-2-songliubraving@fb.com

bpf: Fix a task_iter bug caused by a merge conflict resolution

2021-01-03T00:41:32+00:00

Latest bpf tree has a bug for bpf_iter selftest:

  $ ./test_progs -n 4/25
  test_bpf_sk_storage_get:PASS:bpf_iter_bpf_sk_storage_helpers__open_and_load 0 nsec
  test_bpf_sk_storage_get:PASS:socket 0 nsec
  ...
  do_dummy_read:PASS:read 0 nsec
  test_bpf_sk_storage_get:FAIL:bpf_map_lookup_elem map value wasn't set correctly
                          (expected 1792, got -1, err=0)
  #4/25 bpf_sk_storage_get:FAIL
  #4 bpf_iter:FAIL
  Summary: 0/0 PASSED, 0 SKIPPED, 2 FAILED

When doing merge conflict resolution, Commit 4bfc4714849d missed to
save curr_task to seq_file private data. The task pointer in seq_file
private data is passed to bpf program. This caused NULL-pointer task
passed to bpf program which will immediately return upon checking
whether task pointer is NULL.

This patch added back the assignment of curr_task to seq_file private
data and fixed the issue.

Fixes: 4bfc4714849d ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf")
Signed-off-by: Yonghong Song 
Signed-off-by: Daniel Borkmann 
Acked-by: John Fastabend 
Link: https://lore.kernel.org/bpf/20201231052418.577024-1-yhs@fb.com

Merge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf

2020-12-28T23:26:11+00:00

Daniel Borkmann says:

====================
pull-request: bpf 2020-12-28

The following pull-request contains BPF updates for your *net* tree.

There is a small merge conflict between bpf tree commit 69ca310f3416
("bpf: Save correct stopping point in file seq iteration") and net tree
commit 66ed594409a1 ("bpf/task_iter: In task_file_seq_get_next use
task_lookup_next_fd_rcu"). The get_files_struct() does not exist anymore
in net, so take the hunk in HEAD and add the `info->tid = curr_tid` to
the error path:

  [...]
                curr_task = task_seq_get_next(ns, &curr_tid, true);
                if (!curr_task) {
                        info->task = NULL;
                        info->tid = curr_tid;
                        return NULL;
                }

                /* set info->task and info->tid */
  [...]

We've added 10 non-merge commits during the last 9 day(s) which contain
a total of 11 files changed, 75 insertions(+), 20 deletions(-).

The main changes are:

1) Various AF_XDP fixes such as fill/completion ring leak on failed bind and
   fixing a race in skb mode's backpressure mechanism, from Magnus Karlsson.

2) Fix latency spikes on lockdep enabled kernels by adding a rescheduling
   point to BPF hashtab initialization, from Eric Dumazet.

3) Fix a splat in task iterator by saving the correct stopping point in the
   seq file iteration, from Jonathan Lemon.

4) Fix BPF maps selftest by adding retries in case hashtab returns EBUSY
   errors on update/deletes, from Andrii Nakryiko.

5) Fix BPF selftest error reporting to something more user friendly if the
   vmlinux BTF cannot be found, from Kamal Mostafa.
====================

Signed-off-by: David S. Miller

bpf: Use thread_group_leader()

2020-12-24T01:04:53+00:00

Instead of directly comparing task->tgid and task->pid, use the
thread_group_leader() helper.  This helps with readability, and
there should be no functional change.

Signed-off-by: Jonathan Lemon 
Signed-off-by: Daniel Borkmann 
Acked-by: Andrii Nakryiko 
Link: https://lore.kernel.org/bpf/20201218185032.2464558-3-jonathan.lemon@gmail.com

bpf: Save correct stopping point in file seq iteration

2020-12-24T01:04:47+00:00

On some systems, some variant of the following splat is
repeatedly seen.  The common factor in all traces seems
to be the entry point to task_file_seq_next().  With the
patch, all warnings go away.

    rcu: INFO: rcu_sched self-detected stall on CPU
    rcu: \x0926-....: (20992 ticks this GP) idle=d7e/1/0x4000000000000002 softirq=81556231/81556231 fqs=4876
    \x09(t=21033 jiffies g=159148529 q=223125)
    NMI backtrace for cpu 26
    CPU: 26 PID: 2015853 Comm: bpftool Kdump: loaded Not tainted 5.6.13-0_fbk4_3876_gd8d1f9bf80bb #1
    Hardware name: Quanta Twin Lakes MP/Twin Lakes Passive MP, BIOS F09_3A12 10/08/2018
    Call Trace:
     
     dump_stack+0x50/0x70
     nmi_cpu_backtrace.cold.6+0x13/0x50
     ? lapic_can_unplug_cpu.cold.30+0x40/0x40
     nmi_trigger_cpumask_backtrace+0xba/0xca
     rcu_dump_cpu_stacks+0x99/0xc7
     rcu_sched_clock_irq.cold.90+0x1b4/0x3aa
     ? tick_sched_do_timer+0x60/0x60
     update_process_times+0x24/0x50
     tick_sched_timer+0x37/0x70
     __hrtimer_run_queues+0xfe/0x270
     hrtimer_interrupt+0xf4/0x210
     smp_apic_timer_interrupt+0x5e/0x120
     apic_timer_interrupt+0xf/0x20
     
    RIP: 0010:get_pid_task+0x38/0x80
    Code: 89 f6 48 8d 44 f7 08 48 8b 00 48 85 c0 74 2b 48 83 c6 55 48 c1 e6 04 48 29 f0 74 19 48 8d 78 20 ba 01 00 00 00 f0 0f c1 50 20 <85> d2 74 27 78 11 83 c2 01 78 0c 48 83 c4 08 c3 31 c0 48 83 c4 08
    RSP: 0018:ffffc9000d293dc8 EFLAGS: 00000202 ORIG_RAX: ffffffffffffff13
    RAX: ffff888637c05600 RBX: ffffc9000d293e0c RCX: 0000000000000000
    RDX: 0000000000000001 RSI: 0000000000000550 RDI: ffff888637c05620
    RBP: ffffffff8284eb80 R08: ffff88831341d300 R09: ffff88822ffd8248
    R10: ffff88822ffd82d0 R11: 00000000003a93c0 R12: 0000000000000001
    R13: 00000000ffffffff R14: ffff88831341d300 R15: 0000000000000000
     ? find_ge_pid+0x1b/0x20
     task_seq_get_next+0x52/0xc0
     task_file_seq_get_next+0x159/0x220
     task_file_seq_next+0x4f/0xa0
     bpf_seq_read+0x159/0x390
     vfs_read+0x8a/0x140
     ksys_read+0x59/0xd0
     do_syscall_64+0x42/0x110
     entry_SYSCALL_64_after_hwframe+0x44/0xa9
    RIP: 0033:0x7f95ae73e76e
    Code: Bad RIP value.
    RSP: 002b:00007ffc02c1dbf8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
    RAX: ffffffffffffffda RBX: 000000000170faa0 RCX: 00007f95ae73e76e
    RDX: 0000000000001000 RSI: 00007ffc02c1dc30 RDI: 0000000000000007
    RBP: 00007ffc02c1ec70 R08: 0000000000000005 R09: 0000000000000006
    R10: fffffffffffff20b R11: 0000000000000246 R12: 00000000019112a0
    R13: 0000000000000000 R14: 0000000000000007 R15: 00000000004283c0

If unable to obtain the file structure for the current task,
proceed to the next task number after the one returned from
task_seq_get_next(), instead of the next task number from the
original iterator.

Also, save the stopping task number from task_seq_get_next()
on failure in case of restarts.

Fixes: eaaacd23910f ("bpf: Add task and task/file iterator targets")
Signed-off-by: Jonathan Lemon 
Signed-off-by: Daniel Borkmann 
Acked-by: Andrii Nakryiko 
Link: https://lore.kernel.org/bpf/20201218185032.2464558-2-jonathan.lemon@gmail.com

Merge branch 'exec-for-v5.11' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace

2020-12-16T03:29:43+00:00

Pull execve updates from Eric Biederman:
 "This set of changes ultimately fixes the interaction of posix file
  lock and exec. Fundamentally most of the change is just moving where
  unshare_files is called during exec, and tweaking the users of
  files_struct so that the count of files_struct is not unnecessarily
  played with.

  Along the way fcheck and related helpers were renamed to more
  accurately reflect what they do.

  There were also many other small changes that fell out, as this is the
  first time in a long time much of this code has been touched.

  Benchmarks haven't turned up any practical issues but Al Viro has
  observed a possibility for a lot of pounding on task_lock. So I have
  some changes in progress to convert put_files_struct to always rcu
  free files_struct. That wasn't ready for the merge window so that will
  have to wait until next time"

* 'exec-for-v5.11' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace: (27 commits)
  exec: Move io_uring_task_cancel after the point of no return
  coredump: Document coredump code exclusively used by cell spufs
  file: Remove get_files_struct
  file: Rename __close_fd_get_file close_fd_get_file
  file: Replace ksys_close with close_fd
  file: Rename __close_fd to close_fd and remove the files parameter
  file: Merge __alloc_fd into alloc_fd
  file: In f_dupfd read RLIMIT_NOFILE once.
  file: Merge __fd_install into fd_install
  proc/fd: In fdinfo seq_show don't use get_files_struct
  bpf/task_iter: In task_file_seq_get_next use task_lookup_next_fd_rcu
  proc/fd: In proc_readfd_common use task_lookup_next_fd_rcu
  file: Implement task_lookup_next_fd_rcu
  kcmp: In get_file_raw_ptr use task_lookup_fd_rcu
  proc/fd: In tid_fd_mode use task_lookup_fd_rcu
  file: Implement task_lookup_fd_rcu
  file: Rename fcheck lookup_fd_rcu
  file: Replace fcheck_files with files_lookup_fd_rcu
  file: Factor files_lookup_fd_locked out of fcheck_files
  file: Rename __fcheck_files to files_lookup_fd_raw
  ...

bpf/task_iter: In task_file_seq_get_next use task_lookup_next_fd_rcu

2020-12-10T18:42:58+00:00

When discussing[1] exec and posix file locks it was realized that none
of the callers of get_files_struct fundamentally needed to call
get_files_struct, and that by switching them to helper functions
instead it will both simplify their code and remove unnecessary
increments of files_struct.count.  Those unnecessary increments can
result in exec unnecessarily unsharing files_struct which breaking
posix locks, and it can result in fget_light having to fallback to
fget reducing system performance.

Using task_lookup_next_fd_rcu simplifies task_file_seq_get_next, by
moving the checking for the maximum file descritor into the generic
code, and by remvoing the need for capturing and releasing a reference
on files_struct.  As the reference count of files_struct no longer
needs to be maintained bpf_iter_seq_task_file_info can have it's files
member removed and task_file_seq_get_next no longer needs it's fstruct
argument.

The curr_fd local variable does need to become unsigned to be used
with fnext_task.  As curr_fd is assigned from and assigned a u32
making curr_fd an unsigned int won't cause problems and might prevent
them.

[1] https://lkml.kernel.org/r/20180915160423.GA31461@redhat.com
Suggested-by: Oleg Nesterov 
v1: https://lkml.kernel.org/r/20200817220425.9389-11-ebiederm@xmission.com
Link: https://lkml.kernel.org/r/20201120231441.29911-16-ebiederm@xmission.com
Signed-off-by: Eric W. Biederman

file: Replace fcheck_files with files_lookup_fd_rcu

2020-12-10T18:40:03+00:00

This change renames fcheck_files to files_lookup_fd_rcu.  All of the
remaining callers take the rcu_read_lock before calling this function
so the _rcu suffix is appropriate.  This change also tightens up the
debug check to verify that all callers hold the rcu_read_lock.

All callers that used to call files_check with the files->file_lock
held have now been changed to call files_lookup_fd_locked.

This change of name has helped remind me of which locks and which
guarantees are in place helping me to catch bugs later in the
patchset.

The need for better names became apparent in the last round of
discussion of this set of changes[1].

[1] https://lkml.kernel.org/r/CAHk-=wj8BQbgJFLa+J0e=iT-1qpmCRTbPAJ8gd6MJQ=kbRPqyQ@mail.gmail.com
Link: https://lkml.kernel.org/r/20201120231441.29911-9-ebiederm@xmission.com
Signed-off-by: Eric W. Biederman

bpf: Simplify task_file_seq_get_next()

2020-11-20T19:36:34+00:00

Simplify task_file_seq_get_next() by removing two in/out arguments: task
and fstruct. Use info->task and info->files instead.

Signed-off-by: Song Liu 
Signed-off-by: Daniel Borkmann 
Acked-by: Yonghong Song 
Link: https://lore.kernel.org/bpf/20201120002833.2481110-1-songliubraving@fb.com

bpf: Permit cond_resched for some iterators

2020-10-28T21:54:31+00:00

Commit e679654a704e ("bpf: Fix a rcu_sched stall issue with
bpf task/task_file iterator") tries to fix rcu stalls warning
which is caused by bpf task_file iterator when running
"bpftool prog".

      rcu: INFO: rcu_sched self-detected stall on CPU
      rcu: \x097-....: (20999 ticks this GP) idle=302/1/0x4000000000000000 softirq=1508852/1508852 fqs=4913
      \x09(t=21031 jiffies g=2534773 q=179750)
      NMI backtrace for cpu 7
      CPU: 7 PID: 184195 Comm: bpftool Kdump: loaded Tainted: G        W         5.8.0-00004-g68bfc7f8c1b4 #6
      Hardware name: Quanta Twin Lakes MP/Twin Lakes Passive MP, BIOS F09_3A17 05/03/2019
      Call Trace:
      
      dump_stack+0x57/0x70
      nmi_cpu_backtrace.cold+0x14/0x53
      ? lapic_can_unplug_cpu.cold+0x39/0x39
      nmi_trigger_cpumask_backtrace+0xb7/0xc7
      rcu_dump_cpu_stacks+0xa2/0xd0
      rcu_sched_clock_irq.cold+0x1ff/0x3d9
      ? tick_nohz_handler+0x100/0x100
      update_process_times+0x5b/0x90
      tick_sched_timer+0x5e/0xf0
      __hrtimer_run_queues+0x12a/0x2a0
      hrtimer_interrupt+0x10e/0x280
      __sysvec_apic_timer_interrupt+0x51/0xe0
      asm_call_on_stack+0xf/0x20
      
      sysvec_apic_timer_interrupt+0x6f/0x80
      ...
      task_file_seq_next+0x52/0xa0
      bpf_seq_read+0xb9/0x320
      vfs_read+0x9d/0x180
      ksys_read+0x5f/0xe0
      do_syscall_64+0x38/0x60
      entry_SYSCALL_64_after_hwframe+0x44/0xa9

The fix is to limit the number of bpf program runs to be
one million. This fixed the program in most cases. But
we also found under heavy load, which can increase the wallclock
time for bpf_seq_read(), the warning may still be possible.

For example, calling bpf_delay() in the "while" loop of
bpf_seq_read(), which will introduce artificial delay,
the warning will show up in my qemu run.

  static unsigned q;
  volatile unsigned *p = &q;
  volatile unsigned long long ll;
  static void bpf_delay(void)
  {
         int i, j;

         for (i = 0; i < 10000; i++)
                 for (j = 0; j < 10000; j++)
                         ll += *p;
  }

There are two ways to fix this issue. One is to reduce the above
one million threshold to say 100,000 and hopefully rcu warning will
not show up any more. Another is to introduce a target feature
which enables bpf_seq_read() calling cond_resched().

This patch took second approach as the first approach may cause
more -EAGAIN failures for read() syscalls. Note that not all bpf_iter
targets can permit cond_resched() in bpf_seq_read() as some, e.g.,
netlink seq iterator, rcu read lock critical section spans through
seq_ops->next() -> seq_ops->show() -> seq_ops->next().

For the kernel code with the above hack, "bpftool p" roughly takes
38 seconds to finish on my VM with 184 bpf program runs.
Using the following command, I am able to collect the number of
context switches:
   perf stat -e context-switches -- ./bpftool p >& log
Without this patch,
   69      context-switches
With this patch,
   75      context-switches
This patch added additional 6 context switches, roughly every 6 seconds
to reschedule, to avoid lengthy no-rescheduling which may cause the
above RCU warnings.

Signed-off-by: Yonghong Song 
Signed-off-by: Alexei Starovoitov 
Acked-by: Andrii Nakryiko 
Link: https://lore.kernel.org/bpf/20201028061054.1411116-1-yhs@fb.com