Re: [Bug #11308] tbench regression on each kernel releasefrom 2.6.22 -> 2.6.28
From: Mike Galbraith
Date: Thu Sep 18 2008 - 03:13:20 EST
On Wed, 2008-09-17 at 15:57 +0200, Mike Galbraith wrote:
> On Wed, 2008-09-17 at 16:36 +0300, Ilpo JÃrvinen wrote:
> > On Wed, 17 Sep 2008, Mike Galbraith wrote:
> >
> > > On Wed, 2008-09-17 at 14:49 +0200, Ingo Molnar wrote:
> > >
> > > > git log --pretty=format:"%h: %s" 2069f45..847106f | grep -viE \
> > > > 'block|alsa|pcmcia|sound|Merge|iosched|blk|DAC960|scsi|s390|paride|pktcdvd|filter|cdrom|drm'
> > > >
> > > > gives us:
> > > >
> > > > 7daf705: Start using the new '%pS' infrastructure to print symbols
> > > > 6f0f0fd: security: remove register_security hook
> > > > 93cbace: security: remove dummy module fix
> > > > 5915eb5: security: remove dummy module
> > > > b478a9f: security: remove unused sb_get_mnt_opts hook
> > > > 32502b8: splice: fix generic_file_splice_read() race with page invalidation
> > > > 8b3d356: ramfs: enable splice write
> > > > a144ff0: xen: Avoid allocations causing swap activity on the resume path
> > > >
> > > > which really only leaves that security commit your bisection fingered.
> > > > Which _slightly_ raises its likelyhood of being implicated. Structure
> > > > size changes can move two formerly far-apart netperf-relevant symbols on
> > > > the same cacheline, which can start cache ping-pong-ing badly.
> > >
> > > I sure hope it's something like ping-pong, it's driving me NUTS.
> >
> > How about dividing the problem to smaller blocks then by restoring
> > parts of the change...
>
> Well, what I've done is check out the "bad" tree, reverted every darn
> commit between there and the "good" tree, and then reverted the reverts
> so I have a nice merge-free line and don't have to remember to think
> backward. (probably sounds silly to git-foo masters) I'll try
> bisecting that in the a.m. and see what happens.
It bisected to 1c9ce52. Reverting that in 27.git had the expected
result, nada. Full bisection/test log below - you can jump straight to
post run sanity checks. I'm torn between building a straight line tree
from v2.6.26 through git.today and bisecting that sucker, or exorcising
netperf from my box and swearing a sacred oath to never download the
damned thing again. Nuking netperf is most attractive option.
1e65e841bb5584136ed6047c55cf77532afbbb55 is first bad commit
commit 1e65e841bb5584136ed6047c55cf77532afbbb55
Author: Mike Galbraith <efault@xxxxxx>
Date: Wed Sep 17 14:55:50 2008 +0200
Revert "Revert "block: export "ro" attribute""
This reverts commit 2c8803af5c1bf41200167f29349f7f1396683a51.
:040000 040000 08ca8ba7ff3f9506a5462b4122256356cae28ceb bef679485bc924ad1dc867858ebda1b68196b5a8 M block
git bisect start
# good: [7804ad865f7d0cd9bdc51da601772ce4d2e252ca] Revert "[ALSA] soc - tlv320aic3x - revisit clock setup"
git bisect good 7804ad865f7d0cd9bdc51da601772ce4d2e252ca
# bad: [2846693a63a34ac6d582dd55a7e00605b49b1cec] Revert "Revert "[S390] sclp_tty: Fix scheduling while atomic bug.""
git bisect bad 2846693a63a34ac6d582dd55a7e00605b49b1cec
# bad: [70477f86f63640be2dd1d8968aeb47870a5c21c6] Revert "Revert "xen/blkfront: Make sure we don't use bounce buffers, we don't need them.""
git bisect bad 70477f86f63640be2dd1d8968aeb47870a5c21c6
# good: [7c6ccb520424939deff0a50f3fae621c6477dbbe] Revert "Revert "ALSA: hda - Add bdl_pos_adj option""
git bisect good 7c6ccb520424939deff0a50f3fae621c6477dbbe
# good: [66036beae94f043f99044e285935486126a9c4bd] Revert "Revert "pcmcia: fix Alchemy warnings""
git bisect good 66036beae94f043f99044e285935486126a9c4bd
# good: [6b7b5ef18871c8f7c15eedc6eb53270eaf8bc613] Revert "Revert "ALSA: hda - Added SSID for 'Fujitsu Siemens Amilo M1451G' laptop""
git bisect good 6b7b5ef18871c8f7c15eedc6eb53270eaf8bc613
# good: [024905ea4b2d1ab5d6b845ba84ddfc0857fb2d2a] Revert "Revert "ALSA: hda - Add MacBook 3.1 support""
git bisect good 024905ea4b2d1ab5d6b845ba84ddfc0857fb2d2a
# good: [4bbe3501e06eddaa4900894efa519f1167cb8624] Revert "Revert "as-iosched: properly protect ioc_gone and ioc count""
git bisect good 4bbe3501e06eddaa4900894efa519f1167cb8624
# good: [e124683c1cd5c73c494f9ea6cee8a71e68bb70ca] Revert "Revert "Added in user-injected messages into blk traces""
git bisect good e124683c1cd5c73c494f9ea6cee8a71e68bb70ca
# bad: [f148fae0bc009aa23122c5762368a1a16bb55b86] Revert "Revert "block: kill request_queue_t""
git bisect bad f148fae0bc009aa23122c5762368a1a16bb55b86
# bad: [1e65e841bb5584136ed6047c55cf77532afbbb55] Revert "Revert "block: export "ro" attribute""
git bisect bad 1e65e841bb5584136ed6047c55cf77532afbbb55
v2.6.26-974-g2846693 (847106f)
16384 87380 1 1 60.00 94350.45
16384 87380 1 1 60.01 95857.25
16384 87380 1 1 60.00 95334.84
16384 87380 1 1 60.00 95052.11
v2.6.26-659-g7804ad8 (2069f45)
16384 87380 1 1 60.00 98630.64
16384 87380 1 1 60.00 98653.14
16384 87380 1 1 60.00 99162.65
16384 87380 1 1 60.00 98652.38
v2.6.26-816-g70477f8 (call it bad)
16384 87380 1 1 60.00 95532.19
16384 87380 1 1 60.01 96211.39
16384 87380 1 1 60.01 96246.73
16384 87380 1 1 60.01 96286.40
v2.6.26-737-g7c6ccb5
16384 87380 1 1 60.00 98478.00
16384 87380 1 1 60.00 99221.33
16384 87380 1 1 60.00 98930.70
16384 87380 1 1 60.00 98958.73
v2.6.26-776-g66036be
16384 87380 1 1 60.00 97958.10
16384 87380 1 1 60.01 98683.80
16384 87380 1 1 60.01 98515.34
16384 87380 1 1 60.00 98396.11
v2.6.26-796-g6b7b5ef
16384 87380 1 1 60.00 99047.21
16384 87380 1 1 60.00 98095.23
16384 87380 1 1 60.00 99811.18
16384 87380 1 1 60.00 98651.32
v2.6.26-806-g024905e
16384 87380 1 1 60.00 98823.46
16384 87380 1 1 60.00 98959.11
16384 87380 1 1 60.00 98709.95
16384 87380 1 1 60.00 99042.13
v2.6.26-811-g4bbe350
16384 87380 1 1 60.00 98144.99
16384 87380 1 1 60.01 99023.30
16384 87380 1 1 60.01 98685.45
16384 87380 1 1 60.01 98606.18
v2.6.26-813-ge124683
16384 87380 1 1 60.00 98458.18
16384 87380 1 1 60.00 98163.92
16384 87380 1 1 60.00 98115.62
16384 87380 1 1 60.00 98633.62
v2.6.26-815-gf148fae
16384 87380 1 1 60.00 95649.91
16384 87380 1 1 60.00 96292.34
16384 87380 1 1 60.00 96043.82
16384 87380 1 1 60.00 96093.81
v2.6.26-814-g1e65e84
16384 87380 1 1 60.00 94906.81
16384 87380 1 1 60.00 95445.05
16384 87380 1 1 60.01 94698.68
16384 87380 1 1 60.00 94938.65
Post bisection checkouts
------------------------------------------------
v2.6.26-rc8-208-g02c6230
16384 87380 1 1 60.00 98392.94
16384 87380 1 1 60.00 98199.96
16384 87380 1 1 60.00 98534.27
16384 87380 1 1 60.00 98501.02
v2.6.26-rc8-209-g1c9ce52
16384 87380 1 1 60.00 97583.88
16384 87380 1 1 60.00 97326.23
16384 87380 1 1 60.00 97582.80
16384 87380 1 1 60.00 97568.63
v2.6.26-814-g1e65e84
16384 87380 1 1 60.00 94856.33
16384 87380 1 1 60.00 94594.03
16384 87380 1 1 60.00 94751.74
16384 87380 1 1 60.01 96825.28
v2.6.26-813-ge124683
16384 87380 1 1 60.00 97550.64
16384 87380 1 1 60.00 98024.28
16384 87380 1 1 60.00 98486.85
16384 87380 1 1 60.00 98493.41
marge:..git/linux-2.6 # git rev-list v2.6.26-813-ge124683..v2.6.26-814-g1e65e84
1e65e841bb5584136ed6047c55cf77532afbbb55
marge:..git/linux-2.6 # git show 1e65e841bb5584136ed6047c55cf77532afbbb55
commit 1e65e841bb5584136ed6047c55cf77532afbbb55
Author: Mike Galbraith <efault@xxxxxx>
Date: Wed Sep 17 14:55:50 2008 +0200
Revert "Revert "block: export "ro" attribute""
This reverts commit 2c8803af5c1bf41200167f29349f7f1396683a51.
diff --git a/block/genhd.c b/block/genhd.c
index b922d48..43e468e 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -400,6 +400,14 @@ static ssize_t disk_removable_show(struct device *dev,
(disk->flags & GENHD_FL_REMOVABLE ? 1 : 0));
}
+static ssize_t disk_ro_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct gendisk *disk = dev_to_disk(dev);
+
+ return sprintf(buf, "%d\n", disk->policy ? 1 : 0);
+}
+
static ssize_t disk_size_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
@@ -472,6 +480,7 @@ static ssize_t disk_fail_store(struct device *dev,
static DEVICE_ATTR(range, S_IRUGO, disk_range_show, NULL);
static DEVICE_ATTR(removable, S_IRUGO, disk_removable_show, NULL);
+static DEVICE_ATTR(ro, S_IRUGO, disk_ro_show, NULL);
static DEVICE_ATTR(size, S_IRUGO, disk_size_show, NULL);
static DEVICE_ATTR(capability, S_IRUGO, disk_capability_show, NULL);
static DEVICE_ATTR(stat, S_IRUGO, disk_stat_show, NULL);
@@ -483,6 +492,7 @@ static struct device_attribute dev_attr_fail =
static struct attribute *disk_attrs[] = {
&dev_attr_range.attr,
&dev_attr_removable.attr,
+ &dev_attr_ro.attr,
&dev_attr_size.attr,
&dev_attr_capability.attr,
&dev_attr_stat.attr,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/