Re: reiser4 panic in do_readpage_extent

[Edward Shishkin <edward@namesys.com>]

[-- Attachment #1: Type: text/plain, Size: 591 bytes --]

Devils-Hawk wrote:

> Recently tried switching from 2.6.18 + reiser4-for-2.6.18-r3.patch.gz, 
> which works perfectly fine to 2.6.19 + reiser4-for-2.6.19-r3.patch.gz
> I also tried 2.6.20 laurent riffard's reiser4-for-2.6.20. The last 
> both die somewhere during init when one of the 2 following asserts fails:
> extent_file_ops, Line 1160: assert("nikita-2688"),j)
> extent_file_ops, Line 1161: assert("vs-1426"),jnode_page(j) == NULL )
>
> fs was fsck'ed with reiser4progs-1.0.5
>
> regards DevH
>
>

Would you please try the attached patch over reiser4-for-2.6.[19, 20]

Thanks,
Edward.

[-- Attachment #2: reiser4-handle-race-in-readpages-filler.patch --]
[-- Type: text/x-patch, Size: 542 bytes --]

Handle possible race:
do not proceed uf_readpages_filler if page is already uptodate.

Signed-off-by: Edward Shishkin <edward@namesys.com>
---
 linux-2.6.20-mm2/fs/reiser4/plugin/file/file.c |    4 ++++
 1 files changed, 4 insertions(+)

--- linux-2.6.20-mm2/fs/reiser4/plugin/file/file.c.orig
+++ linux-2.6.20-mm2/fs/reiser4/plugin/file/file.c
@@ -1619,6 +1619,10 @@
 		lock_page(page);
 		cbk_done = 1;
 	}
+	if (PageUptodate(page)) {
+		unlock_page(page);
+		return 0;
+	}
 	ret = zload(rc->coord.node);
 	if (ret) {
 		unlock_page(page);