Re: [syzbot] Monthly trace report (May 2024)

On Sun, Jun 02, 2024 at 12:09:50PM +0900, Masami Hiramatsu wrote:

On Thu, 30 May 2024 23:50:32 -0700
syzbot [off-list ref] wrote:

quoted

Hello trace maintainers/developers,

This is a 31-day syzbot report for the trace subsystem.
All related reports/information can be found at:
https://syzkaller.appspot.com/upstream/s/trace

During the period, 1 new issues were detected and 0 were fixed.
In total, 10 issues are still open and 35 have been fixed so far.

Some of the still happening issues:

Ref Crashes Repro Title
<1> 705     Yes   WARNING in format_decode (3)
                  https://syzkaller.appspot.com/bug?extid=e2c932aec5c8a6e1d31c

Could you send this to bpf folks? It seems bpf_trace_printk caused this errror.
(Maybe skipping fmt string check?)

quoted

<2> 26      Yes   INFO: task hung in blk_trace_ioctl (4)
                  https://syzkaller.appspot.com/bug?extid=ed812ed461471ab17a0c

This looks like debugfs_mutex lock leakage. Need to rerun with lockdep.

quoted

<3> 7       Yes   WARNING in get_probe_ref
                  https://syzkaller.appspot.com/bug?extid=8672dcb9d10011c0a160

Hm, fail on register_trace_block_rq_insert(). blktrace issue.

quoted

<4> 6       Yes   INFO: task hung in blk_trace_remove (2)
                  https://syzkaller.appspot.com/bug?extid=2373f6be3e6de4f92562

This looks like debugfs_mutex lock leakage too.

quoted

<5> 5       Yes   general protection fault in bpf_get_attach_cookie_tracing
                  https://syzkaller.appspot.com/bug?extid=3ab78ff125b7979e45f9

This is also BPF problem.

this one seems to be easy to fix, can't reproduce with either the change
below or with instrumenting __bpf_prog_test_run_raw_tp to set current->bpf_ctx
as in __bpf_trace_run

will send a patch

thanks,
jirka


---

diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c
index 593efccc2030..fc303c20f402 100644
--- a/kernel/trace/bpf_trace.c
+++ b/kernel/trace/bpf_trace.c

@@ -1148,6 +1148,8 @@ BPF_CALL_1(bpf_get_attach_cookie_trace, void *, ctx)
 {
 	struct bpf_trace_run_ctx *run_ctx;
 
+	if (!current->bpf_ctx)
+		return 0;
 	run_ctx = container_of(current->bpf_ctx, struct bpf_trace_run_ctx, run_ctx);
 	return run_ctx->bpf_cookie;
 }