2022-11-22 04:15:29

by Yang Jihong

Subject: [PATCH] blktrace: Fix output non-blktrace event when blk_classic option enabled

When the blk_classic option is enabled, non-blktrace events must be
filtered out. Otherwise, events of other types are output in the blktrace
classic format, which is unexpected.

The problem can be triggered in the following ways:

# echo 1 > /sys/kernel/debug/tracing/options/blk_classic
# echo 1 > /sys/kernel/debug/tracing/events/enable
# echo blk > /sys/kernel/debug/tracing/current_tracer
# cat /sys/kernel/debug/tracing/trace_pipe

Fixes: c71a89615411 ("blktrace: add ftrace plugin")
Signed-off-by: Yang Jihong <[email protected]>
---
kernel/trace/blktrace.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c
index a995ea1ef849..a66cff5a1857 100644
--- a/kernel/trace/blktrace.c
+++ b/kernel/trace/blktrace.c
@@ -1548,7 +1548,8 @@ blk_trace_event_print_binary(struct trace_iterator *iter, int flags,

static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter)
{
- if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC))
+ if ((iter->ent->type != TRACE_BLK) ||
+ !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC))
return TRACE_TYPE_UNHANDLED;

return print_one_line(iter, true);
--
2.30.GIT

2022-12-07 02:01:13

by Yang Jihong

Subject: Re: [PATCH] blktrace: Fix output non-blktrace event when blk_classic option enabled

Hello,

PING.

Thanks,
Yang.

On 2022/11/22 12:04, Yang Jihong wrote:
> When the blk_classic option is enabled, non-blktrace events must be
> filtered out. Otherwise, events of other types are output in the blktrace
> classic format, which is unexpected.
>
> The problem can be triggered in the following ways:
>
> # echo 1 > /sys/kernel/debug/tracing/options/blk_classic
> # echo 1 > /sys/kernel/debug/tracing/events/enable
> # echo blk > /sys/kernel/debug/tracing/current_tracer
> # cat /sys/kernel/debug/tracing/trace_pipe
>
> Fixes: c71a89615411 ("blktrace: add ftrace plugin")
> Signed-off-by: Yang Jihong <[email protected]>
> ---
> kernel/trace/blktrace.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c
> index a995ea1ef849..a66cff5a1857 100644
> --- a/kernel/trace/blktrace.c
> +++ b/kernel/trace/blktrace.c
> @@ -1548,7 +1548,8 @@ blk_trace_event_print_binary(struct trace_iterator *iter, int flags,
>
> static enum print_line_t blk_tracer_print_line(struct trace_iterator *iter)
> {
> - if (!(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC))
> + if ((iter->ent->type != TRACE_BLK) ||
> + !(blk_tracer_flags.val & TRACE_BLK_OPT_CLASSIC))
> return TRACE_TYPE_UNHANDLED;
>
> return print_one_line(iter, true);
>

2022-12-08 16:41:04

by Jens Axboe

Subject: Re: [PATCH] blktrace: Fix output non-blktrace event when blk_classic option enabled

On 12/6/22 6:34 PM, Yang Jihong wrote:
> Hello,
>
> PING.

Get your company email fixed so that messages don't get marked as spam,
that would help ensure your patches are more visible.

--
Jens Axboe

2022-12-08 18:04:31

by Jens Axboe

Subject: Re: [PATCH] blktrace: Fix output non-blktrace event when blk_classic option enabled

On Tue, 22 Nov 2022 12:04:10 +0800, Yang Jihong wrote:
> When the blk_classic option is enabled, non-blktrace events must be
> filtered out. Otherwise, events of other types are output in the blktrace
> classic format, which is unexpected.
>
> The problem can be triggered in the following ways:
>
> # echo 1 > /sys/kernel/debug/tracing/options/blk_classic
> # echo 1 > /sys/kernel/debug/tracing/events/enable
> # echo blk > /sys/kernel/debug/tracing/current_tracer
> # cat /sys/kernel/debug/tracing/trace_pipe
>
> [...]

Applied, thanks!

[1/1] blktrace: Fix output non-blktrace event when blk_classic option enabled
commit: f596da3efaf4130ff61cd029558845808df9bf99

Best regards,
--
Jens Axboe

2022-12-09 02:09:13

by Yang Jihong

Subject: Re: [PATCH] blktrace: Fix output non-blktrace event when blk_classic option enabled

Hello,

On 2022/12/9 0:26, Jens Axboe wrote:
> Get your company email fixed so that messages don't get marked as spam,
> that would help ensure your patches are more visible.

OK, I'll consult with the relevant department of our company about this
issue.

Thanks,
Yang