Re: [Qemu-devel] [PATCH] qemu-iotests: Fix core dump suppression in test 039

[Kevin Wolf <kwolf@redhat.com>]

Am 14.05.2014 um 13:16 hat Markus Armbruster geschrieben:
> Markus Armbruster <armbru@redhat.com> writes:
> 
> > Kevin Wolf <kwolf@redhat.com> writes:
> >
> >> Am 13.05.2014 um 19:44 hat Markus Armbruster geschrieben:
> >>> Fam Zheng <famz@redhat.com> writes:
> >>> 
> >>> > On Tue, 05/13 10:46, Markus Armbruster wrote:
> >>> >> The shell script attempts to suppress core dumps like this:
> >>> >> 
> >>> >>     old_ulimit=$(ulimit -c)
> >>> >>     ulimit -c 0
> >>> >>     $QEMU_IO arg...
> >>> >>     ulimit -c "$old_ulimit"
> >>> >> 
> >>> >> This breaks the test hard unless the limit was zero to begin with!
> >>> >> ulimit sets both hard and soft limit by default, and (re-)raising the
> >>> >> hard limit requires privileges.  Broken since it was added in commit
> >>> >> dc68afe.
> >>> >> 
> >>> >> Could be fixed by adding -S to set only the soft limit, but I'm not
> >>> >> sure how portable that is in practice.  Simply do it in a subshell
> >>> >> instead, like this:
> >>> >> 
> >>> >>     (ulimit -c 0; exec $QEMU_IO arg...)
> >>> >> 
> >>> >> Signed-off-by: Markus Armbruster <armbru@redhat.com>
> >>> >> ---
> >>> >>  tests/qemu-iotests/039 | 18 ++++++------------
> >>> >>  1 file changed, 6 insertions(+), 12 deletions(-)
> >>> >> 
> >>> >> diff --git a/tests/qemu-iotests/039 b/tests/qemu-iotests/039
> >>> >> index b9cbe99..182b0f0 100755
> >>> >> --- a/tests/qemu-iotests/039
> >>> >> +++ b/tests/qemu-iotests/039
> >>> >> @@ -67,10 +67,8 @@ echo "== Creating a dirty image file =="
> >>> >>  IMGOPTS="compat=1.1,lazy_refcounts=on"
> >>> >>  _make_test_img $size
> >>> >>  
> >>> >> -old_ulimit=$(ulimit -c)
> >>> >> -ulimit -c 0 # do not produce a core dump on abort(3)
> >>> >> -$QEMU_IO -c "write -P 0x5a 0 512" -c "abort" "$TEST_IMG" | _filter_qemu_io
> >>> >> -ulimit -c "$old_ulimit"
> >>> >> +(ulimit -c 0 # do not produce a core dump on abort(3)
> >>> >> +exec $QEMU_IO -c "write -P 0x5a 0 512" -c "abort" "$TEST_IMG") | _filter_qemu_io
> >>> >
> >>> > This works well.
> >>> >
> >>> > But when I try to put this in a function to avoid repeating:
> >>> >
> >>> >     function _no_dump_exec()
> >>> >     {
> >>> >         (ulimit -c 0; exec "$@")
> >>> >     }
> >>> >
> >>> >     _no_dump_exec $QEMU_IO -c "write -P 0x5a 0 512" -c "abort" "$TEST_IMG") | _filter_qemu_io
> >>> >
> >>> > it doesn't work:
> >>> >
> >>> >     039 1s ... - output mismatch (see 039.out.bad)
> >>> >     --- 039.out     2014-05-13 12:10:39.248866480 +0800
> >>> >     +++ 039.out.bad 2014-05-13 17:19:46.161986618 +0800
> >>> >     @@ -9,6 +9,7 @@
> >>> >
> >>> >      == Creating a dirty image file ==
> >>> >      Formatting 'TEST_DIR/t.IMGFMT', fmt=IMGFMT size=134217728
> >>> >     +./039: line 51: 10517 Aborted                 "$@"
> >>> >      wrote 512/512 bytes at offset 0
> >>> >      512 bytes, X ops; XX:XX:XX.X (XXX YYY/sec and XXX ops/sec)
> >>> >      incompatible_features     0x1
> >>> >
> >>> > Any idea what the difference is here?
> >>> 
> >>> This is qemu-io aborting, as instructed.  The command is
> >>> 
> >>>     qemu-io --cache writeback --cache writethrough -c 'write -P 0x5a
> >>> 0 512' -c abort scratch/t.qcow2
> >>> 
> >>> [...]
> >>> 
> >>> The additional "Aborted" line appears as soon as I put pass the qemu-io
> >>> command to a function that runs it using "$@".  I don't need a subshell,
> >>> exec or anything:
> >>
> >> So that looks fine, I'd even consider it a feature to have the abort
> >> recorded explicitly.  Let's just update the reference output. Another
> >> reason why qemu-iotests is bash-only, but we already have the same kind
> >> of output in other test cases, so this is not setting a precedence.
> >
> > Okay, I'll respin it that way.
> 
> The message printed by the shell looks like:
> 
> ./039: line <LINENR>:  <PID> Aborted <LINETEXT>
> 
> Need to filter out the PID.  Okay to add that to the sed script in
> _filter_qemu_io, or would you like to have it elsewhere?

Fine with me in _filter_qemu_io. We may later need it elsewhere too,
but we can still move it then.

Kevin