Skip to contentSkip to conversation

read_all: scale down how many times we read by default

PATCH New Jan Stancek 3 participants 4 messages Jun 9, 2020

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 read_all: scale down how many times we read by default

read_all is running into timeouts on high cpu systems, where access to some per-cpu files is protected by a lock. Latest example is /sys/kernel/tracing/per_cpu/*.

At the moment we read each file 10 times, and we have been excluding files that take too long. Rather than expanding blacklist, scale the default down to 3.

---
 runtest/fs | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
runtest/fs +3 -3
index 464ba8fb9686..5892e9fdaee5 100644
@@ -69,9 +69,9 @@ fs_di fs_di -d $TMPDIR
 # Was not sure why it should reside in runtest/crashme and won't get tested ever
 proc01 proc01 -m 128
 
-read_all_dev read_all -d /dev -p -q -r 10
-read_all_proc read_all -d /proc -q -r 10
-read_all_sys read_all -d /sys -q -r 10
+read_all_dev read_all -d /dev -p -q -r 3
+read_all_proc read_all -d /proc -q -r 3
+read_all_sys read_all -d /sys -q -r 3
 
 #Run the File System Race Condition Check tests as well
 fs_racer fs_racer.sh -t 5
Signature · 1 line
2.18.1
#1 lore raw

PATCH 1/1 read_all: scale down how many times we read by default

[Cc Richard] in case he has extra suggest.

Quoted 10 lines · Jan Stancek

read_all is running into timeouts on high cpu systems, where access to some per-cpu files is protected by a lock. Latest example is /sys/kernel/tracing/per_cpu/*.

At the moment we read each file 10 times, and we have been excluding files that take too long. Rather than expanding blacklist, scale the default down to 3.

Signature · 5 lines
Regards,
Li Wang
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linux.it/pipermail/ltp/attachments/20200609/b70c248a/attachment.htm>
#2 lore raw

PATCH 1/1 read_all: scale down how many times we read by default

In reply to Jan Stancek

Hello,

Quoted 32 lines · Jan Stancek

read_all is running into timeouts on high cpu systems, where access to some per-cpu files is protected by a lock. Latest example is /sys/kernel/tracing/per_cpu/*.

At the moment we read each file 10 times, and we have been excluding files that take too long. Rather than expanding blacklist, scale the default down to 3.

---
 runtest/fs | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
runtest/fs +3 -3
index 464ba8fb9686..5892e9fdaee5 100644
@@ -69,9 +69,9 @@ fs_di fs_di -d $TMPDIR
 # Was not sure why it should reside in runtest/crashme and won't get tested ever
 proc01 proc01 -m 128
 
-read_all_dev read_all -d /dev -p -q -r 10
-read_all_proc read_all -d /proc -q -r 10
-read_all_sys read_all -d /sys -q -r 10
+read_all_dev read_all -d /dev -p -q -r 3
+read_all_proc read_all -d /proc -q -r 3
+read_all_sys read_all -d /sys -q -r 3
 
 #Run the File System Race Condition Check tests as well
 fs_racer fs_racer.sh -t 5
Signature · 1 line
2.18.1

OK this makes sense. We shouldn't be stress testing the system in this runtest file.

Signature · 2 lines
Thank you,
Richard.
#3 lore raw

PATCH 1/1 read_all: scale down how many times we read by default

----- Original Message -----

Quoted 2 lines

OK this makes sense. We shouldn't be stress testing the system in this runtest file.

Pushed.