rpms/blktrace
5
0
Fork 0
Code Issues Pull Requests Releases Activity

fix hang when BLKTRACESETUP fails and "-o -" is used

Browse Source 
Resolves: RHEL-54662

Signed-off-by: Pavel Reichl <preichl@redhat.com>
This commit is contained in:
Pavel Reichl 2024-08-20 20:29:25 +02:00
parent a43c639d87
commit e45df32741
2 changed files with 63 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
blktrace.spec
View File

@ -1,7 +1,7 @@
Summary: Utilities for performing block layer IO tracing in the Linux kernel
Name: blktrace
Version: 1.3.0
Release: 11%{?dist}
Release: 12%{?dist}
License: GPL-2.0-or-later
Source0: http://brick.kernel.dk/snaps/blktrace-%{version}.tar.bz2
Source1: https://brick.kernel.dk/snaps/blktrace-%{version}.tar.bz2.asc
@ -16,6 +16,9 @@ BuildRequires: gcc, libaio-devel, librsvg2-devel
BuildRequires: make
BuildRequires: gnupg2
# Upstream: blktrace-1.3.0-5-g1836be5
Patch0: for-next-fix-hang-when-BLKTRACESETUP-fails-and-o-is-used.patch
%description
blktrace is a block layer IO tracing mechanism which provides detailed
information about request queue operations to user space. This package
@ -85,6 +88,10 @@ information about IO patterns.
%{_mandir}/man1/iowatcher.*
%changelog
* Tue Aug 20 2024 Pavel Reichl <preichl@redhat.com> - 1.3.0-12
- fix hang when BLKTRACESETUP fails and "-o -" is used
- Related: RHEL-54662
* Mon Jun 24 2024 Troy Dawson <tdawson@redhat.com> - 1.3.0-11
- Bump release for June 2024 mass rebuild

55
for-next-fix-hang-when-BLKTRACESETUP-fails-and-o-is-used.patch Normal file
View File

@ -0,0 +1,55 @@
From 1836be5d99c9362f1e2b39206c95270f19cb7faa Mon Sep 17 00:00:00 2001
From: Eric Sandeen <sandeen@redhat.com>
Date: Wed, 5 Jun 2024 23:07:27 -0500
Subject: [PATCH] fix hang when BLKTRACESETUP fails and "-o -" is used
# blktrace -o - /dev/sda /dev/sdb /dev/sdc
has to be SIGKILLed if BLKTRACESETUP fails for any or all of the devices
listed. (I simulated this by just catching one of the devices in
setup_buts(), skipping the ioctl, and doing ret++).
This seems to be because with "-o -" on the command line, use_tracer_devpaths()
sees piped_output set, so we call process_trace_bufs which ends up waiting on
(!done) and "done" is never set. i.e.
atexit(exit_tracing)
wait_tracers
if (use_tracer_devpaths()) // true because "-o -"
process_trace_bufs
while (wait_empty_entries())
wait_empty_entries
while (!done ... )
<loop forever>
I think this can be avoided by just setting "done = 1" before returning
when setup_buts() fails in run_tracers().
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Link: https://lore.kernel.org/r/f3204c9d-1384-40b5-a5fb-3bb967ca2bec@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Pavel Reichl <preichl@redhat.com>
---
blktrace.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/blktrace.c b/blktrace.c
index 3444fbb..038b2cb 100644
--- a/blktrace.c
+++ b/blktrace.c
@@ -2684,8 +2684,10 @@ static int run_tracers(void)
if (net_mode == Net_client)
printf("blktrace: connecting to %s\n", hostname);
- if (setup_buts())
+ if (setup_buts()) {
+ done = 1;
return 1;
+ }
if (use_tracer_devpaths()) {
if (setup_tracer_devpaths())
--
2.46.0