Re: [bug] splat in perf event

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Dec 19, 2023 at 05:34:13PM -0800, Alexei Starovoitov wrote:
> Hi,
> 
> after rebasing bpf-next to the latest net-next.
> I consistently see the following while running
> test_progs -t attach_probe/manual-default
> 
> [   28.638654] WARNING: CPU: 1 PID: 2135 at kernel/events/core.c:1950
> __do_sys_perf_event_open+0x14e0/0x15b0
> [   28.639329] Modules linked in: bpf_testmod(O)
> [   28.639632] CPU: 1 PID: 2135 Comm: test_progs Tainted: G
> O       6.7.0-rc5-01520-gc337f237291b #5281
> [   28.640299] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
> BIOS rel-1.12.0-59-gc9ba5276e321-prebuilt.qemu.org 04/01/2014
> [   28.641062] RIP: 0010:__do_sys_perf_event_open+0x14e0/0x15b0
> [   28.647751] Call Trace:
> [   28.647919]  <TASK>
> [   28.648082]  ? __warn+0xa1/0x1f0
> [   28.648311]  ? __do_sys_perf_event_open+0x14e0/0x15b0
> [   28.648641]  ? report_bug+0x1fa/0x230
> [   28.648902]  ? handle_bug+0x3c/0x70
> [   28.649164]  ? exc_invalid_op+0x17/0x40
> [   28.649416]  ? asm_exc_invalid_op+0x1a/0x20
> [   28.649699]  ? entry_SYSCALL_64_after_hwframe+0x46/0x4e
> [   28.650062]  ? __do_sys_perf_event_open+0x14e0/0x15b0
> [   28.650406]  ? perf_event_set_output+0x2a0/0x2a0
> [   28.650727]  ? __audit_syscall_entry+0x4f/0x200
> [   28.651063]  do_syscall_64+0x2f/0xa0
> [   28.651306]  entry_SYSCALL_64_after_hwframe+0x46/0x4e
> [   28.651635] RIP: 0033:0x7fc0846f752d
> [   28.656060]  </TASK>
> [   28.656219] irq event stamp: 413681
> [   28.656461] hardirqs last  enabled at (413689):
> [<ffffffff81193e67>] console_unlock+0x137/0x140
> [   28.657083] hardirqs last disabled at (413698):
> [<ffffffff81193e4c>] console_unlock+0x11c/0x140
> [   28.657663] softirqs last  enabled at (413368):
> [<ffffffff810c0e89>] irq_exit_rcu+0x99/0xf0
> [   28.658215] softirqs last disabled at (413351):
> [<ffffffff810c0e89>] irq_exit_rcu+0x99/0xf0
> 
> Line 1950 is
>         for_each_sibling_event(sibling, group_leader) {
>                 if (__perf_event_read_size(sibling->attr.read_format,
>                                            group_leader->nr_siblings +
> 1) > 16*1024)
>                         return false;
>         }
> 
> 
> Probably a known issue?

yes, https://lore.kernel.org/linux-perf-users/20231214000620.3081018-1-lucas.demarchi@xxxxxxxxx/

looks like there'll be patch out soon

jirka




[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux