Messages in this thread | | | Date | Tue, 15 Aug 2000 03:19:23 -0300 (BRST) | From | Rik van Riel <> | Subject | Re: linux-2.4-test & innd |
| |
On Tue, 15 Aug 2000, Henrik Størner wrote: > On Sat, Aug 12, 2000 at 10:31:21PM -0700, Linus Torvalds wrote: > > Can you please double-check exactly which kernel started doing this? I > > have some mixed reports about the "test1-ac" series kernels doing this > > too. > > A pure 2.4.0-test1 kernel also has the bug. It is actually quite easy > to test for: > > * fire up the kernel and innd > * inject a couple of articles using rnews > * note the numbers listed for the newsgroup in the active file > * restart innd and check the active file again > > If it has the bug, the active file will forget all updates after inn > is restarted. This happens with 2.4.0-test1 also.
I don't know if it is related, but I've noticed a serious bug in filemap.c while playing with my new VM patch.
It seems that something in filemap.c (add_to_page_cache_unique?) is calling lru_cache_add for pages which are _already_ on a page list...
In my new VM patch, this is caught by the DEBUG_ADD_PAGE in the lru_cache_add function, while in the vanilla Linux 2.4 tree, this could/should lead to a corruption of the page lru list, with all bad consequences it brings with it (list corruption, maybe stray pointers, maybe a memory leak and maybe even data corruption or data loss).
regards,
Rik -- "What you're running that piece of shit Gnome?!?!" -- Miguel de Icaza, UKUUG 2000
http://www.conectiva.com/ http://www.surriel.com/
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |