on 2020/02/21 7:30, Darrick J. Wong wrote:
On Thu, Feb 20, 2020 at 03:58:47PM +0800, Yang Xu wrote:
When I test this case(default lsunit 256k), this case will fail,
as below:
cycle: 1 version: 2 lsn: 1,0 tail_lsn: 1,0
length of Log Record: 258048 prev offset: -1 num ops: 1
...
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
extended-header: cycle: 1
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
...
It reports this info because xfs_logprint only read 32k header every time, so it
needs to read more times. We can filter this useless info.
common/log also has _filter_logprint function. only library function is
prefixed with "_", remove '_'.
Signed-off-by: Yang Xu <xuyang2018.jy@xxxxxxxxxxxxxx>
---
tests/xfs/029 | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/tests/xfs/029 b/tests/xfs/029
index f532a21b..dbe700ab 100755
--- a/tests/xfs/029
+++ b/tests/xfs/029
@@ -19,7 +19,7 @@ trap "rm -f $tmp.*; exit \$status" 0 1 2 3 15
. ./common/rc
. ./common/filter
-_filter_logprint()
+filter_logprint()
{
perl -ne '
s/data device: ([\w|\/.-]+)/data device: DDEV/;
@@ -31,6 +31,8 @@ _filter_logprint()
s/(length of Log Record:) \d+/$1 <LEN>/;
s/version: \d/version: <VERN>/;
s/h_size: \d+/h_size: <H_SIZE>/;
+ s/^~+[\r|\n]+$//;
+ s/extended-header: cycle: 1[\r|\n]+$//;
I guess this works, but do you want to delete this line for line with
"extended-header: cycle: <some number>"?
Yes, because this line happens depend on log size(every 32k print a
line, exclude the first, so 256k prints 7 line). This case is design to
test mkfs log(internal/external) zeroing. IMO, this line doesn't matter
unless we want to test 32k read header(If so, specify -lsunit=128, but I
want to keep origal design).
Best Regards
Yang Xu
print;
'
}
@@ -45,7 +47,7 @@ echo
_scratch_mkfs_xfs | _filter_mkfs 2>/dev/null
echo
-_scratch_xfs_logprint | _filter_logprint
+_scratch_xfs_logprint | filter_logprint
Thanks for cleaning out the ^ leading underscore.
--D
status=0
exit
--
2.18.0