review: xfstests/016 change

PATCH Timothy Shimmin 2 participants 5 messages May 5, 2008 – May 6, 2008

Share & tools

Timothy Shimmin

author

2008-05-05 03:29 UTC #0 lore raw

Hi there,

xfstests/016 has been failing for a while in many cases. It was relying on the log traffic for create/rm to be unaltered in XFS over the life of XFS.
Its initial purpose was to test out the writing of the log around wrapping and checking that as we wrapped we didn't corrupt any data after the end of the log.

The patch is to modify xfstest, 016, to take a sampling of the log traffic to guestimate how many ops are needed to get to the end of the log, instead of hard-coding in how many ops are expected to get there.

--Tim

===========================================================================
Index: xfstests/016
===========================================================================

xfstests/016 +34 ~~-10~~

@@ -50,7 +50,7 @@ _init()
     echo "*** reset partition"
     $here/src/devzero -b 2048 -n 50 -v 198 $SCRATCH_DEV
     echo "*** mkfs"
-    force_opts="-dsize=50m -lsize=2097152"
+    force_opts="-dsize=50m -lsize=$log_size"
     echo mkfs_xfs $force_opts $SCRATCH_DEV >>$seq.full
     _scratch_mkfs_xfs $force_opts >$tmp.mkfs0 2>&1
     [ $? -ne 0 ] && \

@@ -156,6 +156,10 @@ _supported_os Linux
 
 rm -f $seq.full
 
+# mkfs sizes
+log_size=2097152
+log_size_bb=`expr $log_size / 512`
+
 _require_scratch
 _init

@@ -163,27 +167,47 @@ block=`_after_log $SCRATCH_DEV`
 echo "fsblock after log = $block"               >>$seq.full
 _check_corrupt $SCRATCH_DEV $block
 
-size=`_log_size`
-echo "log size = $size BB"                      >>$seq.full
+actual_log_size=`_log_size`
+echo "log size = $actual_log_size BB"                      >>$seq.full
 head=`_log_head`
 echo "log position = $head"                     >>$seq.full
 lsunit=`_log_sunit`
 echo "log sunit = $lsunit"			>>$seq.full
 
-[ $size -eq 4096 ] || \
+# sanity checks
+[ $actual_log_size -eq $log_size_bb ] || \
     _fail "!!! unexpected log size $size"
 [ $head -eq 2 -o $head -eq $((lsunit/512)) ] || \
     _fail "!!! unexpected initial log position $head vs. $((lsunit/512))"
 
-echo "    lots of traffic"                      >>$seq.full
-_log_traffic 850
+# find how how many blocks per op for 100 ops
+# ignore the fact that it will also include an unmount record etc...
+# this should be small overall
+echo "    lots of traffic for sampling" >>$seq.full
+sample_size_ops=100
+_log_traffic $sample_size_ops
+head1=`_log_head`
+num_blocks=`expr $head1 - $head`
+blocks_per_op=`echo "scale=3; $num_blocks / $sample_size_ops" | bc`
+echo "blocks_per_op = $blocks_per_op" >>$seq.full
+num_expected_ops=`echo "$log_size_bb / $blocks_per_op" | bc`
+echo "num_expected_ops = $num_expected_ops" >>$seq.full
+num_expected_to_go=`echo "$num_expected_ops - $sample_size_ops" | bc`
+echo "num_expected_to_go = $num_expected_to_go" >>$seq.full
+
+echo "    lots more traffic" >>$seq.full
+_log_traffic $num_expected_to_go
 head=`_log_head`
 echo "log position = $head"                     >>$seq.full
 
-[ $head -gt 3850 -a $head -lt 4050 ] || \
-    _fail "!!! unexpected log position $head"
+# e.g. 3891
+near_end_min=`echo "0.95 * $log_size_bb" | bc | sed 's/\..*//'`
+echo "near_end_min = $near_end_min" >>$seq.full
+
+[ $head -gt $near_end_min -a $head -lt $log_size_bb ] || \
+    _fail "!!! unexpected near end log position $head"
 
-for c in 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
+for c in `seq 0 20`
 do
     echo "   little traffic"            >>$seq.full
     _log_traffic 2

@@ -193,7 +217,7 @@ do
 done
 
 [ $head -lt 1000 ] || \
-    _fail "!!! unexpected log position $head"
+    _fail "!!! unexpected wrapped log position $head"
 
 # success, all done
 status=0

Christoph Hellwig

2008-05-05 05:54 UTC #1 lore raw

Quoted 9 lines · Timothy Shimmin

Hi there,
xfstests/016 has been failing for a while in many cases. It was relying on the log traffic for create/rm to be unaltered in XFS over the life of XFS.
Its initial purpose was to test out the writing of the log around wrapping and checking that as we wrapped we didn't corrupt any data after the end of the log.

Quoted 3 lines

The patch is to modify xfstest, 016, to take a sampling of the log traffic to guestimate how many ops are needed to get to the end of the log, instead of hard-coding in how many ops are expected to get there.

I get four rejects out of four hunk when trying to apply this patch, looks like whitespace-damage.

Timothy Shimmin

author

2008-05-05 06:19 UTC #2 lore raw

Quoted 10 lines · Christoph Hellwig

Quoted 9 lines · Timothy Shimmin
Hi there,
xfstests/016 has been failing for a while in many cases. It was relying on the log traffic for create/rm to be unaltered in XFS over the life of XFS.
Its initial purpose was to test out the writing of the log around wrapping and checking that as we wrapped we didn't corrupt any data after the end of the log.

Quoted 6 lines

Quoted 3 lines
The patch is to modify xfstest, 016, to take a sampling of the log traffic to guestimate how many ops are needed to get to the end of the log, instead of hard-coding in how many ops are expected to get there.
I get four rejects out of four hunk when trying to apply this patch, looks like whitespace-damage.

Groan. Sorry.
I'll attach 016.patch for the moment.

Need to see where things went wrong with the whitespace.

--Tim

===========================================================================
Index: xfstests/016
===========================================================================

xfstests/016 +34 ~~-10~~

@@ -50,7 +50,7 @@ _init()
     echo "*** reset partition"
     $here/src/devzero -b 2048 -n 50 -v 198 $SCRATCH_DEV
     echo "*** mkfs"
-    force_opts="-dsize=50m -lsize=2097152"
+    force_opts="-dsize=50m -lsize=$log_size"
     echo mkfs_xfs $force_opts $SCRATCH_DEV >>$seq.full
     _scratch_mkfs_xfs $force_opts >$tmp.mkfs0 2>&1
     [ $? -ne 0 ] && \

@@ -156,6 +156,10 @@ _supported_os Linux
 
 rm -f $seq.full
 
+# mkfs sizes
+log_size=2097152
+log_size_bb=`expr $log_size / 512`
+
 _require_scratch
 _init

@@ -163,27 +167,47 @@ block=`_after_log $SCRATCH_DEV`
 echo "fsblock after log = $block"               >>$seq.full
 _check_corrupt $SCRATCH_DEV $block
 
-size=`_log_size`
-echo "log size = $size BB"                      >>$seq.full
+actual_log_size=`_log_size`
+echo "log size = $actual_log_size BB"                      >>$seq.full
 head=`_log_head`
 echo "log position = $head"                     >>$seq.full
 lsunit=`_log_sunit`
 echo "log sunit = $lsunit"			>>$seq.full
 
-[ $size -eq 4096 ] || \
+# sanity checks
+[ $actual_log_size -eq $log_size_bb ] || \
     _fail "!!! unexpected log size $size"
 [ $head -eq 2 -o $head -eq $((lsunit/512)) ] || \
     _fail "!!! unexpected initial log position $head vs. $((lsunit/512))"
 
-echo "    lots of traffic"                      >>$seq.full
-_log_traffic 850
+# find how how many blocks per op for 100 ops
+# ignore the fact that it will also include an unmount record etc...
+# this should be small overall
+echo "    lots of traffic for sampling" >>$seq.full
+sample_size_ops=100
+_log_traffic $sample_size_ops
+head1=`_log_head`
+num_blocks=`expr $head1 - $head`
+blocks_per_op=`echo "scale=3; $num_blocks / $sample_size_ops" | bc`
+echo "blocks_per_op = $blocks_per_op" >>$seq.full
+num_expected_ops=`echo "$log_size_bb / $blocks_per_op" | bc`
+echo "num_expected_ops = $num_expected_ops" >>$seq.full
+num_expected_to_go=`echo "$num_expected_ops - $sample_size_ops" | bc`
+echo "num_expected_to_go = $num_expected_to_go" >>$seq.full
+
+echo "    lots more traffic" >>$seq.full
+_log_traffic $num_expected_to_go
 head=`_log_head`
 echo "log position = $head"                     >>$seq.full
 
-[ $head -gt 3850 -a $head -lt 4050 ] || \
-    _fail "!!! unexpected log position $head"
+# e.g. 3891
+near_end_min=`echo "0.95 * $log_size_bb" | bc | sed 's/\..*//'`
+echo "near_end_min = $near_end_min" >>$seq.full
+
+[ $head -gt $near_end_min -a $head -lt $log_size_bb ] || \
+    _fail "!!! unexpected near end log position $head"
 
-for c in 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
+for c in `seq 0 20`
 do
     echo "   little traffic"            >>$seq.full
     _log_traffic 2

@@ -193,7 +217,7 @@ do
 done
 
 [ $head -lt 1000 ] || \
-    _fail "!!! unexpected log position $head"
+    _fail "!!! unexpected wrapped log position $head"
 
 # success, all done
 status=0

Christoph Hellwig

2008-05-05 18:30 UTC #3 lore raw

Quoted 4 lines · Timothy Shimmin

Groan. Sorry.
I'll attach 016.patch for the moment.
Need to see where things went wrong with the whitespace.

Still fails for me with that patch applied:

016 43s ...
- output mismatch (see 016.out.bad)
13a14,17

Quoted 4 lines

*** generate log traffic
*** mount
*** fiddle
*** unmount

Failures: 016
Failed 1 of 1 tests

Timothy Shimmin

author

2008-05-06 01:29 UTC #4 lore raw

Quoted 11 lines · Christoph Hellwig

Quoted 4 lines · Timothy Shimmin
Groan. Sorry.
I'll attach 016.patch for the moment.
Need to see where things went wrong with the whitespace.
Still fails for me with that patch applied:
016 43s ...
- output mismatch (see 016.out.bad)
13a14,17

Quoted 6 lines

Quoted 4 lines
*** generate log traffic
*** mount
*** fiddle
*** unmount
Failures: 016
Failed 1 of 1 tests

Forgot adding 016.out change to patch.
The failure diff above means it worked ;-)

--Tim