From owner-linux-mm-bugs@nl.linux.org Thu Feb 1 17:37:23 2001 Received: by humbolt.nl.linux.org id ; Thu, 1 Feb 2001 17:37:12 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:47624 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 1 Feb 2001 17:36:58 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id NAA22596; Thu, 1 Feb 2001 13:32:16 -0200 Date: Thu, 1 Feb 2001 13:32:16 -0200 Message-Id: <200102011532.NAA22596@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1186] New - 2.4.1: OOM killer doesn't trigger Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1186 --- shadow/1186 Thu Feb 1 13:32:16 2001 +++ shadow/1186.tmp.22592 Thu Feb 1 13:32:16 2001 @@ -0,0 +1,18 @@ ++============================================================================+ +| 2.4.1: OOM killer doesn't trigger | ++----------------------------------------------------------------------------+ +| Bug #: 1186 Product: Linux kernel MM | +| Status: NEW Version: 2.4 (stable) | +| Resolution: Platform: All | +| Severity: minor OS/Version: Linux | +| Priority: P2 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +for some reason, the OOM killer doesn't trigger under kernel versions 2.4.0 and +2.4.1 ... obviously, this needs to be fixed. \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Thu Feb 1 18:47:33 2001 Received: by humbolt.nl.linux.org id ; Thu, 1 Feb 2001 18:47:24 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:41993 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 1 Feb 2001 18:46:56 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id OAA24109; Thu, 1 Feb 2001 14:46:49 -0200 Date: Thu, 1 Feb 2001 14:46:49 -0200 Message-Id: <200102011646.OAA24109@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org, marcelo@conectiva.com.br Subject: [Bug 1175] Changed - swap readahead sets the page referenced bit Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1175 --- shadow/1175 Wed Jan 31 16:54:08 2001 +++ shadow/1175.tmp.24105 Thu Feb 1 14:46:48 2001 @@ -3,8 +3,8 @@ +----------------------------------------------------------------------------+ | Bug #: 1175 Product: Linux kernel MM | | Status: NEW Version: 2.4 (stable) | -| Resolution: Platform: | -| Severity: normal OS/Version: | +| Resolution: Platform: All | +| Severity: normal OS/Version: Linux | | Priority: P2 Component: linux mm | +----------------------------------------------------------------------------+ | Assigned To: linux-mm-bugs@nl.linux.org | @@ -25,3 +25,19 @@ used by any process, they get freed before other (probably more useful) pages. This can be easily fixed using Rik's __find_page_simple(). \ No newline at end of file + +------- Additional Comments From riel@conectiva.com.br 2001-02-01 14:46 ------- +swapin readahead also has the potential to evict lots of useful pages from +memory, this could/should be most evident when running big processes with very +random access patterns + +the solution in this case would be to place (most of) the pages we do readaround +on on the inactive_dirty list, so they won't put too much pressure on the +working set but still have a decent chance of being used + +for linear access patterns, we can do things differently ... in this case we can +put the pages we read in on the active list and do drop_behind() on an equal +number of pages we've already used + +FreeBSD's vm/vm_page.c::vm_page_dontneed() could be a useful model for doing +this ... anybody have opinions on this idea ? \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Thu Feb 1 23:47:29 2001 Received: by humbolt.nl.linux.org id ; Thu, 1 Feb 2001 23:47:15 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:58386 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 1 Feb 2001 23:46:53 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id TAA29344; Thu, 1 Feb 2001 19:43:12 -0200 Date: Thu, 1 Feb 2001 19:43:12 -0200 Message-Id: <200102012143.TAA29344@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org, riel@conectiva.com.br Subject: [Bug 1186] Changed - 2.4.1: OOM killer doesn't trigger Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1186 --- shadow/1186 Thu Feb 1 13:32:16 2001 +++ shadow/1186.tmp.29340 Thu Feb 1 19:43:12 2001 @@ -16,3 +16,79 @@ | DESCRIPTION | for some reason, the OOM killer doesn't trigger under kernel versions 2.4.0 and 2.4.1 ... obviously, this needs to be fixed. \ No newline at end of file + +------- Additional Comments From roger.larsson@norran.net 2001-02-01 19:43 ------- +Not shure that this is the problem - it might be the execution of the OOM +killer causes the kernel to freeze. +I am not running anything special. + KDE 2.1 beta 1, starting kmail quickly (not giving the ageing a chance...) + move around a little reading some messages, freeze. +Today I succeeded in starting kmail and then switching virtual terminals doing +some cat /proc/meminfo | tee -a meminfo.tee only seconds before freeze, +results like this (last one on file): + + total: used: free: shared: buffers: cached: +Mem: 96931840 95182848 1748992 0 20283392 32808960 +Swap: 679100416 61440 679038976 +MemTotal: 94660 kB +MemFree: 1708 kB +MemShared: 0 kB +Buffers: 19808 kB +Cached: 32040 kB +Active: 48632 kB +Inact_dirty: 1116 kB +Inact_clean: 2100 kB +Inact_target: 1020 kB +HighTotal: 0 kB +HighFree: 0 kB +LowTotal: 94660 kB +LowFree: 1708 kB +SwapTotal: 663184 kB +SwapFree: 663124 kB + +This one captured on paper. + + total: used: free: shared: buffers: cached: +Mem: 96931840 95203328 1728512 0 19750912 31633408 +Swap: 679100416 61440 679038976 +MemTotal: 94660 kB +MemFree: 1688 kB +MemShared: 0 kB +Buffers: (missed) kB +Cached: 30892 kB +Active: 45776 kB +Inact_dirty: 2312 kB +Inact_clean: 2092 kB +Inact_target: 1064 kB +HighTotal: 0 kB +HighFree: 0 kB +LowTotal: 94660 kB +LowFree: 1688 kB +SwapTotal: 663184 kB +SwapFree: 663124 kB + +doing one right now gives + + total: used: free: shared: buffers: cached: +Mem: 96931840 95686656 1245184 0 1630208 31776768 +Swap: 679100416 25690112 653410304 +MemTotal: 94660 kB +MemFree: 1216 kB +MemShared: 0 kB +Buffers: 1592 kB +Cached: 31032 kB +Active: 23064 kB +Inact_dirty: 7312 kB +Inact_clean: 2248 kB +Inact_target: 48 kB +HighTotal: 0 kB +HighFree: 0 kB +LowTotal: 94660 kB +LowFree: 1216 kB +SwapTotal: 663184 kB +SwapFree: 638096 kB + +Running with more on swap and fewer buffers. + +Note: system is completely frozen neither SysRQ nor Alt-Fx works... +(interrupts disabled?) - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 00:17:37 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 00:17:15 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:28435 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 00:16:48 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id UAA29831; Thu, 1 Feb 2001 20:10:10 -0200 Date: Thu, 1 Feb 2001 20:10:10 -0200 Message-Id: <200102012210.UAA29831@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org, riel@conectiva.com.br Subject: [Bug 1186] Changed - 2.4.1: OOM killer doesn't trigger Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1186 --- shadow/1186 Thu Feb 1 19:43:12 2001 +++ shadow/1186.tmp.29827 Thu Feb 1 20:10:10 2001 @@ -92,3 +92,7 @@ Note: system is completely frozen neither SysRQ nor Alt-Fx works... (interrupts disabled?) + + +------- Additional Comments From roger.larsson@norran.net 2001-02-01 20:10 ------- +The previous comment was UNRELATED and too long... /RogerL \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 17:22:51 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 17:22:41 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:12562 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 17:22:16 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id NAA10566; Fri, 2 Feb 2001 13:22:07 -0200 Date: Fri, 2 Feb 2001 13:22:07 -0200 Message-Id: <200102021522.NAA10566@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1194] New - test Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1194 --- shadow/1194 Fri Feb 2 13:22:07 2001 +++ shadow/1194.tmp.10560 Fri Feb 2 13:22:07 2001 @@ -0,0 +1,17 @@ ++============================================================================+ +| test | ++----------------------------------------------------------------------------+ +| Bug #: 1194 Product: Linux kernel MM | +| Status: NEW Version: 2.2 (stable) | +| Resolution: Platform: All | +| Severity: trivial OS/Version: Linux | +| Priority: P2 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: rodrigob@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +test, junk me \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 17:37:42 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 17:37:31 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:40210 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 17:37:12 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id NAA10713; Fri, 2 Feb 2001 13:29:17 -0200 Date: Fri, 2 Feb 2001 13:29:17 -0200 Message-Id: <200102021529.NAA10713@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1194] Changed - test Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1194 --- shadow/1194 Fri Feb 2 13:22:07 2001 +++ shadow/1194.tmp.10708 Fri Feb 2 13:29:17 2001 @@ -2,8 +2,8 @@ | test | +----------------------------------------------------------------------------+ | Bug #: 1194 Product: Linux kernel MM | -| Status: NEW Version: 2.2 (stable) | -| Resolution: Platform: All | +| Status: RESOLVED Version: 2.2 (stable) | +| Resolution: FIXED Platform: All | | Severity: trivial OS/Version: Linux | | Priority: P2 Component: linux mm | +----------------------------------------------------------------------------+ @@ -15,3 +15,8 @@ +============================================================================+ | DESCRIPTION | test, junk me \ No newline at end of file + +------- Additional Comments From riel@conectiva.com.br 2001-02-02 13:29 ------- +ok, now the status update _is_ sent to the mailing list + +... I wonder why the other message got lost this morning \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 19:07:41 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 19:07:22 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:23301 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 19:06:54 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id PAA12531; Fri, 2 Feb 2001 15:05:50 -0200 Date: Fri, 2 Feb 2001 15:05:50 -0200 Message-Id: <200102021705.PAA12531@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1199 (linux mm)] New - 2.4.1: prune_icache horribly inefficient Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1199 --- shadow/1199 Fri Feb 2 15:05:50 2001 +++ shadow/1199.tmp.12527 Fri Feb 2 15:05:50 2001 @@ -0,0 +1,33 @@ ++============================================================================+ +| 2.4.1: prune_icache horribly inefficient | ++----------------------------------------------------------------------------+ +| Bug #: 1199 Product: Linux kernel MM | +| Status: NEW Version: 2.4 (stable) | +| Resolution: Platform: All | +| Severity: minor OS/Version: Linux | +| Priority: P2 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +fs/inode.c::prune_icache() is horribly inefficient and has the potential to make +kswapd (or applications) block on IO for a very long time. + +Near the top of prune_icache(): + /* go simple and safe syncing everything before starting */ + sync_all_inodes(); + +The sync_all_inodes() ends up doing an fsync on all inodes in the system - in +sync_one() - and waiting for the data to be flushed synchronously. + +What we would want instead is to get rid of the call to sync_all_inodes() and +add code to sync out those dirty inodes - with no data pages attached - that we +want to free .. asynchronously. + +This is mostly a performance issue, but it looks like an easy to fix problem +that doesn't involve much new code ... just a sync_one() variant that leaves the +data pages alone and syncs dirty inodes with no data pages attached. \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 19:48:13 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 19:47:53 +0100 Received: from brutus.conectiva.com.br ([200.250.58.146]:52474 "EHLO brutus.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 19:47:37 +0100 Received: from localhost (riel@localhost) by brutus.conectiva.com.br (8.11.2/8.11.2) with ESMTP id f12Il6J07916; Fri, 2 Feb 2001 16:47:07 -0200 X-Authentication-Warning: duckman.distro.conectiva: riel owned process doing -bs Date: Fri, 2 Feb 2001 16:47:06 -0200 (BRDT) From: Rik van Riel X-Sender: riel@duckman.distro.conectiva To: Linus Torvalds cc: linux-mm-bugs@nl.linux.org Subject: MM bug tracking Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list Hi Linus, we (the MM folks) would like to know if you'd mind getting the updates from the Linux MM bugzilla by email. Changed bug status, seeing who's working on which solution, etc. is all nicely in unidiff and you can see the bug URL for the full report, etc. We're NOT asking you to use a bug tracking system, we'd just like you to have the opportunity to stay informed of what we're doing on the bug tracking system. If needed, I'd be more than happy to be your webform-secretary for feeding your stuff back into the bugzilla... ;) regards, Rik -- Linux MM bugzilla: http://linux-mm.org/bugzilla.shtml Virtual memory is like a game you can't win; However, without VM there's truly nothing to lose... http://www.surriel.com/ http://www.conectiva.com/ http://distro.conectiva.com/ - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 21:07:14 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 21:07:05 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:15369 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 21:06:50 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id RAA15239; Fri, 2 Feb 2001 17:07:20 -0200 Date: Fri, 2 Feb 2001 17:07:20 -0200 Message-Id: <200102021907.RAA15239@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1203 (linux mm)] New - 2.5: load control Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1203 --- shadow/1203 Fri Feb 2 17:07:20 2001 +++ shadow/1203.tmp.15235 Fri Feb 2 17:07:20 2001 @@ -0,0 +1,22 @@ ++============================================================================+ +| 2.5: load control | ++----------------------------------------------------------------------------+ +| Bug #: 1203 Product: Linux kernel MM | +| Status: NEW Version: 2.5 (beta) | +| Resolution: Platform: | +| Severity: enhancement OS/Version: | +| Priority: P3 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +when the system reaches the thrashing point, we'll need to suspend some +processes or do other magic tricks to reduce the VM load and make it possible +for the tasks in the system to make progress + +we could look at something like the BSD swapper as an example ... but maybe +we'll want to do something else instead ? \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 21:07:44 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 21:07:33 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:17417 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 21:07:00 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id RAA15179; Fri, 2 Feb 2001 17:03:22 -0200 Date: Fri, 2 Feb 2001 17:03:22 -0200 Message-Id: <200102021903.RAA15179@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1202 (linux mm)] New - RSS ulimit enforcement Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1202 --- shadow/1202 Fri Feb 2 17:03:22 2001 +++ shadow/1202.tmp.15175 Fri Feb 2 17:03:22 2001 @@ -0,0 +1,22 @@ ++============================================================================+ +| RSS ulimit enforcement | ++----------------------------------------------------------------------------+ +| Bug #: 1202 Product: Linux kernel MM | +| Status: NEW Version: 2.5 (beta) | +| Resolution: Platform: All | +| Severity: enhancement OS/Version: Linux | +| Priority: P3 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +feature request for RSS ulimit enforcement ... initial patches already available +from http://www.surriel.com/ and various other sources + +we should have a light-weight RSS ulimit implementation, one that doesn't add +much code overhead for people who don't use it and has simple semantics +everybody agrees on (something my patch doesn't seem to do ;)) \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 21:18:06 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 21:17:46 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:55049 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 21:17:19 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id RAA15349; Fri, 2 Feb 2001 17:13:26 -0200 Date: Fri, 2 Feb 2001 17:13:26 -0200 Message-Id: <200102021913.RAA15349@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1204 (linux mm)] New - 2.5?: read-ahead + drop-behind for everything Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1204 --- shadow/1204 Fri Feb 2 17:13:26 2001 +++ shadow/1204.tmp.15345 Fri Feb 2 17:13:26 2001 @@ -0,0 +1,36 @@ ++============================================================================+ +| 2.5?: read-ahead + drop-behind for everything | ++----------------------------------------------------------------------------+ +| Bug #: 1204 Product: Linux kernel MM | +| Status: NEW Version: 2.5 (beta) | +| Resolution: Platform: All | +| Severity: enhancement OS/Version: Linux | +| Priority: P3 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +Linux currently has pretty good read-ahead code and pretty good IO clustering +code; unfortunately the read-ahead code can only be used for read() and the +clustering code can only be used for mmap() (and, to a certain extent, swap). + +It would be better if both the read-ahead code and the clustering code could be +used for both, as required by the situation. When we have this, it's trivial to +also honour the madvise flags for every type of memory. + +Ben LaHaise has some code available to make read-ahead possible for mmap()ed +files, AFAIK no code exists yet to do read-ahead for anonymous memory. + +For drop-behind for mmap() and anonymous memory no code has been written yet, +but it should be relatively simple to abuse try_to_swap_out() for this. + + +Benefits of these improvements would be mostly people with predictable access +patterns. This should include both the scientific calculation folks and the +people who regularly switch desktops on low-memory machines (don't we all?). +Of course the database people will also be happy when Linux is ready to honour +MADV_RANDOM on any type of memory.. \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 21:27:50 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 21:27:31 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:17674 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 21:27:14 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id RAA15534; Fri, 2 Feb 2001 17:25:34 -0200 Date: Fri, 2 Feb 2001 17:25:34 -0200 Message-Id: <200102021925.RAA15534@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1205 (linux mm)] New - 2.5: fix balancing for good Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1205 --- shadow/1205 Fri Feb 2 17:25:33 2001 +++ shadow/1205.tmp.15530 Fri Feb 2 17:25:33 2001 @@ -0,0 +1,31 @@ ++============================================================================+ +| 2.5: fix balancing for good | ++----------------------------------------------------------------------------+ +| Bug #: 1205 Product: Linux kernel MM | +| Status: NEW Version: 2.5 (beta) | +| Resolution: Platform: All | +| Severity: major OS/Version: Linux | +| Priority: P3 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +VM balancing seems to have been a problem since before Linux 1.0, continuing +with various degrees of (un)balance, having some especially fragile VM balancing +in some of the 1.1, 1.2, 2.2 and 2.3 kernels. + +With balancing, in this report, is meant the balance between stealing pages from +processes, from SHM segments and from the cache. Since we use separate functions +for all of these activities, we have to balance them somehow. + +Now that we're close to having a VM subsystem of which we can understand what's +going on and why things are happening, we might as well try and find a solution +to get rid of the balancing issue forever. + +We can also include the balancing of inode and dentry caches in this problem, +but since these items are not page sized it may be considerably more difficult +to balance them well with normal VM operations. \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 2 21:47:45 2001 Received: by humbolt.nl.linux.org id ; Fri, 2 Feb 2001 21:47:26 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:19723 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 2 Feb 2001 21:46:52 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id RAA15855; Fri, 2 Feb 2001 17:41:08 -0200 Date: Fri, 2 Feb 2001 17:41:08 -0200 Message-Id: <200102021941.RAA15855@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1206 (linux mm)] New - 2.4?: make page aging tunable Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1206 --- shadow/1206 Fri Feb 2 17:41:08 2001 +++ shadow/1206.tmp.15851 Fri Feb 2 17:41:08 2001 @@ -0,0 +1,28 @@ ++============================================================================+ +| 2.4?: make page aging tunable | ++----------------------------------------------------------------------------+ +| Bug #: 1206 Product: Linux kernel MM | +| Status: NEW Version: 2.4 (stable) | +| Resolution: Platform: All | +| Severity: trivial OS/Version: Linux | +| Priority: P3 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: riel@conectiva.com.br | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +since nobody seems to know exactly which page aging tactic works best for +general purpose workloads (or, what a general purpose workload means, for that +matter) it may be an idea to put something in the 2.4 kernel which allows people +to choose between a few page aging tactics. + +Having a test base of several tens of thousands people who actually play with +the stuff in /proc might give us a better idea which page aging tactic works +best, without any noticable cost in code complexity or system stability... + +I already have a trivial piece of code available to tune the page aging tactic. + +Comments ? \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 6 19:48:06 2001 Received: by humbolt.nl.linux.org id ; Tue, 6 Feb 2001 19:47:45 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:11026 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Tue, 6 Feb 2001 19:47:21 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id PAA30321; Tue, 6 Feb 2001 15:46:34 -0200 Date: Tue, 6 Feb 2001 15:46:34 -0200 Message-Id: <200102061746.PAA30321@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br Cc: linux-mm-bugs@nl.linux.org Subject: [Bug 1174 (linux mm)] Changed - vm_enough_memory() doesn't see icache/dcache To: unlisted-recipients:; (no To-header on input) Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1174 --- shadow/1174 Wed Jan 31 16:06:46 2001 +++ shadow/1174.tmp.30315 Tue Feb 6 15:46:33 2001 @@ -5,9 +5,9 @@ | Status: NEW Version: 2.4 (stable) | | Resolution: Platform: All | | Severity: normal OS/Version: Linux | -| Priority: P2 Component: linux mm | +| Priority: P3 Component: linux mm | +----------------------------------------------------------------------------+ -| Assigned To: linux-mm-bugs@nl.linux.org | +| Assigned To: riel@conectiva.com.br | | Reported By: riel@conectiva.com.br | | CC list: Cc: | +----------------------------------------------------------------------------+ @@ -24,3 +24,9 @@ the solution is to fix mm/mmap.c::vm_enough_memory() to know about the memory use of the inode and dentry cache and take this into account when doing the calculation \ No newline at end of file + +------- Additional Comments From riel@conectiva.com.br 2001-02-06 15:46 ------- +OK, I'm taking this bug ... + +(can easily fix+merge it now since it's independent of the other code changes I +did) \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 7 08:47:10 2001 Received: by humbolt.nl.linux.org id ; Wed, 7 Feb 2001 08:47:00 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:11532 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 7 Feb 2001 08:46:36 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id EAA10573; Wed, 7 Feb 2001 04:45:54 -0200 Date: Wed, 7 Feb 2001 04:45:54 -0200 Message-Id: <200102070645.EAA10573@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1186 (linux mm)] Changed - 2.4.1: OOM killer doesn't trigger Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1186 --- shadow/1186 Thu Feb 1 20:10:10 2001 +++ shadow/1186.tmp.10567 Wed Feb 7 04:45:54 2001 @@ -4,8 +4,8 @@ | Bug #: 1186 Product: Linux kernel MM | | Status: NEW Version: 2.4 (stable) | | Resolution: Platform: All | -| Severity: minor OS/Version: Linux | -| Priority: P2 Component: linux mm | +| Severity: normal OS/Version: Linux | +| Priority: P3 Component: linux mm | +----------------------------------------------------------------------------+ | Assigned To: linux-mm-bugs@nl.linux.org | | Reported By: riel@conectiva.com.br | - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 7 08:47:37 2001 Received: by humbolt.nl.linux.org id ; Wed, 7 Feb 2001 08:47:10 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:12812 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 7 Feb 2001 08:46:45 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id EAA10554; Wed, 7 Feb 2001 04:45:05 -0200 Date: Wed, 7 Feb 2001 04:45:05 -0200 Message-Id: <200102070645.EAA10554@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org, marcelo@conectiva.com.br Subject: [Bug 1175 (linux mm)] Changed - swap readahead sets the page referenced bit Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1175 --- shadow/1175 Thu Feb 1 14:46:48 2001 +++ shadow/1175.tmp.10550 Wed Feb 7 04:45:05 2001 @@ -4,8 +4,8 @@ | Bug #: 1175 Product: Linux kernel MM | | Status: NEW Version: 2.4 (stable) | | Resolution: Platform: All | -| Severity: normal OS/Version: Linux | -| Priority: P2 Component: linux mm | +| Severity: enhancement OS/Version: Linux | +| Priority: P3 Component: linux mm | +----------------------------------------------------------------------------+ | Assigned To: linux-mm-bugs@nl.linux.org | | Reported By: marcelo@conectiva.com.br | - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 7 08:47:54 2001 Received: by humbolt.nl.linux.org id ; Wed, 7 Feb 2001 08:47:28 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:11788 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 7 Feb 2001 08:46:41 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id EAA10591; Wed, 7 Feb 2001 04:46:07 -0200 Date: Wed, 7 Feb 2001 04:46:07 -0200 Message-Id: <200102070646.EAA10591@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1177 (linux mm)] Changed - VM: undead swap entry xxx Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1177 --- shadow/1177 Wed Jan 31 18:15:22 2001 +++ shadow/1177.tmp.10587 Wed Feb 7 04:46:07 2001 @@ -3,9 +3,9 @@ +----------------------------------------------------------------------------+ | Bug #: 1177 Product: Linux kernel MM | | Status: NEW Version: 2.4 (stable) | -| Resolution: Platform: | -| Severity: minor OS/Version: | -| Priority: P2 Component: linux mm | +| Resolution: Platform: All | +| Severity: normal OS/Version: Linux | +| Priority: P3 Component: linux mm | +----------------------------------------------------------------------------+ | Assigned To: linux-mm-bugs@nl.linux.org | | Reported By: riel@conectiva.com.br | - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 7 09:07:16 2001 Received: by humbolt.nl.linux.org id ; Wed, 7 Feb 2001 09:06:56 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:10253 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 7 Feb 2001 09:06:35 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id FAA10857; Wed, 7 Feb 2001 05:01:49 -0200 Date: Wed, 7 Feb 2001 05:01:49 -0200 Message-Id: <200102070701.FAA10857@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1186 (linux mm)] Changed - 2.4.1: OOM killer doesn't trigger Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1186 --- shadow/1186 Wed Feb 7 04:45:54 2001 +++ shadow/1186.tmp.10853 Wed Feb 7 05:01:49 2001 @@ -2,7 +2,7 @@ | 2.4.1: OOM killer doesn't trigger | +----------------------------------------------------------------------------+ | Bug #: 1186 Product: Linux kernel MM | -| Status: NEW Version: 2.4 (stable) | +| Status: ASSIGNED Version: 2.4 (stable) | | Resolution: Platform: All | | Severity: normal OS/Version: Linux | | Priority: P3 Component: linux mm | @@ -96,3 +96,8 @@ ------- Additional Comments From roger.larsson@norran.net 2001-02-01 20:10 ------- The previous comment was UNRELATED and too long... /RogerL \ No newline at end of file + +------- Additional Comments From riel@conectiva.com.br 2001-02-07 05:01 ------- +this one is mine, I'll take a good look at this in the morning and will try if I +can create the condition where the OOM killer doesn't get triggered and make the +OOM killing more reliable ... \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Thu Feb 8 07:27:18 2001 Received: by humbolt.nl.linux.org id ; Thu, 8 Feb 2001 07:27:01 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:780 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 8 Feb 2001 07:26:35 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id DAA01718; Thu, 8 Feb 2001 03:22:04 -0200 Date: Thu, 8 Feb 2001 03:22:04 -0200 Message-Id: <200102080522.DAA01718@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1336 (linux mm)] New - Cache flushing bug Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1336 --- shadow/1336 Thu Feb 8 03:22:03 2001 +++ shadow/1336.tmp.1710 Thu Feb 8 03:22:03 2001 @@ -0,0 +1,57 @@ ++============================================================================+ +| Cache flushing bug | ++----------------------------------------------------------------------------+ +| Bug #: 1336 Product: Linux kernel MM | +| Status: UNCONFIRMED Version: 2.4 (stable) | +| Resolution: Platform: All | +| Severity: major OS/Version: Linux | +| Priority: P3 Component: linux mm | ++----------------------------------------------------------------------------+ +| Assigned To: linux-mm-bugs@nl.linux.org | +| Reported By: gniibe@m17n.org | +| CC list: Cc: | ++----------------------------------------------------------------------------+ +| URL: | ++============================================================================+ +| DESCRIPTION | +Platform: non-x86 + +For the arch which cache system is not coherent, we need to +flush the cache (my case is SH-4). I think that current code +has a bug in do_swap_page. + +It only flushs the cache after read_swap_cache. If it hits +lookup_swap_cache, the cache is not flushed. If the page is +read asynchronously (say, readahead), the page is left unflushed. + +Here's a patch to fix that. + +--- v2.4.1/mm/memory.c Sun Jan 28 12:12:35 2001 ++++ linux/mm/memory.c Thu Feb 8 10:05:07 2001 +@@ -1033,9 +1033,6 @@ static int do_swap_page(struct mm_struct + + unlock_kernel(); + + if (!page) + + return -1; +- +- + flush_page_to_ram(page); +- + flush_icache_page(vma, page); + } + + mm->rss++; +@@ -1053,6 +1050,8 @@ static int do_swap_page(struct mm_struct + + pte = pte_mkwrite(pte_mkdirty(pte)); + UnlockPage(page); + ++ +flush_page_to_ram(page); ++ +flush_icache_page(vma, page); + set_pte(page_table, pte); + /* No need to invalidate - it was non-present before */ + update_mmu_cache(vma, address, pte); \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Thu Feb 8 15:37:04 2001 Received: by humbolt.nl.linux.org id ; Thu, 8 Feb 2001 15:36:56 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:40965 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 8 Feb 2001 15:36:38 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id LAA09025; Thu, 8 Feb 2001 11:30:55 -0200 Date: Thu, 8 Feb 2001 11:30:55 -0200 Message-Id: <200102081330.LAA09025@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1194 (linux mm)] Changed - test Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1194 --- shadow/1194 Fri Feb 2 13:29:17 2001 +++ shadow/1194.tmp.9020 Thu Feb 8 11:30:54 2001 @@ -2,7 +2,7 @@ | test | +----------------------------------------------------------------------------+ | Bug #: 1194 Product: Linux kernel MM | -| Status: RESOLVED Version: 2.2 (stable) | +| Status: CLOSED Version: 2.2 (stable) | | Resolution: FIXED Platform: All | | Severity: trivial OS/Version: Linux | | Priority: P2 Component: linux mm | @@ -20,3 +20,6 @@ ok, now the status update _is_ sent to the mailing list ... I wonder why the other message got lost this morning \ No newline at end of file + +------- Additional Comments From riel@conectiva.com.br 2001-02-08 11:30 ------- +closed test bug ;) \ No newline at end of file - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 9 16:58:58 2001 Received: by humbolt.nl.linux.org id ; Fri, 9 Feb 2001 16:58:47 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:29701 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 9 Feb 2001 16:58:31 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id MAA14659; Fri, 9 Feb 2001 12:50:19 -0200 Date: Fri, 9 Feb 2001 12:50:19 -0200 Message-Id: <200102091450.MAA14659@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1205 (linux mm)] Changed - 2.5: fix balancing for good Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1205 riel@conectiva.com.br changed: What |Old Value |New Value ---------------------------------------------------------------------------- Severity|major |enhancement Priority|P3 |P1 ------- Additional Comments From riel@conectiva.com.br 2001-02-09 12:50 ------- hmmm, this is more of a high-priority enhancement ... ;) - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 9 18:18:54 2001 Received: by humbolt.nl.linux.org id ; Fri, 9 Feb 2001 18:18:41 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:31756 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 9 Feb 2001 18:18:26 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id OAA16881; Fri, 9 Feb 2001 14:11:31 -0200 Date: Fri, 9 Feb 2001 14:11:31 -0200 Message-Id: <200102091611.OAA16881@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1403 (linux mm)] Changed - stalls with tmpfs Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1403 ------- Additional Comments From arjan@fenrus.demon.nl 2001-02-09 14:11 ------- Created an attachment (id=41) vmstat output - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 9 18:19:35 2001 Received: by humbolt.nl.linux.org id ; Fri, 9 Feb 2001 18:19:16 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:33292 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 9 Feb 2001 18:18:31 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id OAA16868; Fri, 9 Feb 2001 14:11:06 -0200 Date: Fri, 9 Feb 2001 14:11:06 -0200 Message-Id: <200102091611.OAA16868@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1403 (linux mm)] New - stalls with tmpfs Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1403 Summary: stalls with tmpfs Product: Linux kernel MM Version: 2.4 (stable) Platform: All URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: normal Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: arjan@fenrus.demon.nl While untarring a kerneltree to a tmpfs mount, I get 1/2 to 1 second stalls. vmstat output is attached - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 9 18:38:45 2001 Received: by humbolt.nl.linux.org id ; Fri, 9 Feb 2001 18:38:42 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:41997 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 9 Feb 2001 18:38:27 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id OAA17453; Fri, 9 Feb 2001 14:32:56 -0200 Date: Fri, 9 Feb 2001 14:32:56 -0200 Message-Id: <200102091632.OAA17453@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1403 (linux mm)] Changed - stalls with tmpfs Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1403 ------- Additional Comments From arjan@fenrus.demon.nl 2001-02-09 14:32 ------- (this is with the vm patch from Rik's site) - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 9 21:09:35 2001 Received: by humbolt.nl.linux.org id ; Fri, 9 Feb 2001 21:09:25 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:9223 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 9 Feb 2001 21:08:58 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id RAA21123; Fri, 9 Feb 2001 17:05:12 -0200 Date: Fri, 9 Feb 2001 17:05:12 -0200 Message-Id: <200102091905.RAA21123@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1403 (linux mm)] Changed - stalls with tmpfs Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1403 riel@conectiva.com.br changed: What |Old Value |New Value ---------------------------------------------------------------------------- Status|UNCONFIRMED |NEW everconfirmed|0 |1 ------- Additional Comments From riel@conectiva.com.br 2001-02-09 17:05 ------- Verified the bug (on my dual-pentium 120MHz, so I got to see even longer stalls) My initial suspicion was that big IO spikes were responsible for the stalls, but while I saw the IO spikes they didn't seem to be causing the stalls. in the periods when the stall happens, both kswapd and kflushd (bdflush) are eating CPU like mad. I haven't quite figured out why but I am pretty sure that kswapd and bdflush behaviour together is responsible for the bug. Rik. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 13 02:09:04 2001 Received: by humbolt.nl.linux.org id ; Tue, 13 Feb 2001 02:08:42 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:33797 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Tue, 13 Feb 2001 02:08:19 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id WAA24193; Mon, 12 Feb 2001 22:57:19 -0300 Date: Mon, 12 Feb 2001 22:57:19 -0300 Message-Id: <200102130157.WAA24193@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1336 (linux mm)] Changed - Cache flushing bug Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1336 marcelo@conectiva.com.br changed: What |Old Value |New Value ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |FIXED ------- Additional Comments From marcelo@conectiva.com.br 2001-02-12 22:57 ------- Patch sent to Linus and Alan. Thanks Niibe! - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 13 11:18:52 2001 Received: by humbolt.nl.linux.org id ; Tue, 13 Feb 2001 11:18:41 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:6160 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Tue, 13 Feb 2001 11:18:17 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id IAA31323; Tue, 13 Feb 2001 08:11:23 -0300 Date: Tue, 13 Feb 2001 08:11:23 -0300 Message-Id: <200102131111.IAA31323@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1336 (linux mm)] Changed - Cache flushing bug Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1336 ------- Additional Comments From sct@redhat.com 2001-02-13 08:11 ------- This is a hot path: if you have a parent process continually forking children, then the parent may have a swapped pte which is physically in memory and the children will be looking that page up in the swap cache on each fork. Slowing that down with unnecessary cache flushes is a bad thing. Sendmail and apache show this behaviour. If we need to flush the cache because of IO, then surely we should be moving it to the IO path, not adding cache flushes to the in-memory swap-cache lookup path? - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 13 16:39:33 2001 Received: by humbolt.nl.linux.org id ; Tue, 13 Feb 2001 16:39:25 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:16402 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Tue, 13 Feb 2001 16:38:31 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id NAA03894; Tue, 13 Feb 2001 13:26:49 -0300 Date: Tue, 13 Feb 2001 13:26:49 -0300 Message-Id: <200102131626.NAA03894@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1336 (linux mm)] Changed - Cache flushing bug Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1336 sct@redhat.com changed: What |Old Value |New Value ---------------------------------------------------------------------------- Status|RESOLVED |UNCONFIRMED Resolution|FIXED | ------- Additional Comments From sct@redhat.com 2001-02-13 13:26 ------- Needs a better fix. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 14 17:02:18 2001 Received: by humbolt.nl.linux.org id ; Wed, 14 Feb 2001 17:01:59 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:48136 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 14 Feb 2001 17:01:24 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id NAA25776; Wed, 14 Feb 2001 13:54:39 -0300 Date: Wed, 14 Feb 2001 13:54:39 -0300 Message-Id: <200102141654.NAA25776@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1501 (linux mm)] New - Swapoff freezes machine for 60+ seconds? Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1501 Summary: Swapoff freezes machine for 60+ seconds? Product: Linux kernel MM Version: 2.4 (stable) Platform: i386 URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: trivial Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: burton@fint.org Hello. Is this "normal"? toy:/home/bwindle# uname -a Linux toy 2.4.1-ac9 #1 Fri Feb 9 04:55:09 EST 2001 i586 unknown toy:/home/bwindle# free ; date ; swapoff -a ; date ; free total used free shared buffers cached Mem: 320280 260232 60048 0 9004 196888 -/+ buffers/cache: 54340 265940 Swap: 191512 61628 129884 Tue Feb 13 21:11:09 EST 2001 Tue Feb 13 21:12:22 EST 2001 total used free shared buffers cached Mem: 320280 235328 84952 0 9004 145612 -/+ buffers/cache: 80712 239568 Swap: 0 0 0 toy:# As you can see, the 'swapoff -a' command took well over 60 seconds. During this time, X was running, and the machine pretty much became non-responsive (no mouse movement, xmms stopped, seconds stop in clock, etc, except to pings) for the last 45 seconds before swapoff returned. Once swapoff was done, machine was normal again. This is just a fairly typical x86 box (Debian Sid, K6-2 400, Asus P5A-B mobo, 320mb ram, one ide hdd, two ide cdroms, one ne2000 nic, ensonic sound card). Any ideas? Or is this normal? - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Mon Feb 19 03:21:36 2001 Received: by humbolt.nl.linux.org id ; Mon, 19 Feb 2001 03:21:27 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:18187 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Mon, 19 Feb 2001 03:21:09 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id AAA21625; Mon, 19 Feb 2001 00:14:20 -0300 Date: Mon, 19 Feb 2001 00:14:20 -0300 Message-Id: <200102190314.AAA21625@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1551 (linux mm)] New - grep causes strange memory behaviour Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1551 Summary: grep causes strange memory behaviour Product: Linux kernel MM Version: 2.4 (stable) Platform: i386 URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: normal Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: morten@wtf.dk Hi, I'm no developer, or any technical head, but I think this is a bug in the Linux kernel, actually I don't even know if it belongs on the MM buglist - please ignore if it's not. This is running Linux kernel 2.4.1 on Debian unstable. While running X, and being root in a terminal, this causes my system to act funny: $ cd / $ grep -r "0:0.0" * grep: dev/kmem: Bad address Binary file dev/mem matches Binary file dev/core matches grep: memory exhausted This causes my X screen to flicker a lot, and my system generaly runnning bad, when quitting X - I see that my system uses 180 mb ram (I have a total of 256 mb ram) on console only, and in top or ps, I can't see which process which consumes this ram. I tried it on a friends system, and the exact same behaviour happened. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Mon Feb 19 20:48:51 2001 Received: by humbolt.nl.linux.org id ; Mon, 19 Feb 2001 20:48:42 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:38667 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Mon, 19 Feb 2001 20:48:26 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id QAA04170; Mon, 19 Feb 2001 16:40:18 -0300 Date: Mon, 19 Feb 2001 16:40:18 -0300 Message-Id: <200102191940.QAA04170@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1501 (linux mm)] Changed - Swapoff freezes machine for 60+ seconds? Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1501 marcelo@conectiva.com.br changed: What |Old Value |New Value ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |INVALID ------- Additional Comments From marcelo@conectiva.com.br 2001-02-19 16:40 ------- Its normal behaviour. You had 60MB in swap, and the kernel read all this memory from disk when the swapoff happened. Thats why it took so long. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 20 20:54:58 2001 Received: by humbolt.nl.linux.org id ; Tue, 20 Feb 2001 20:54:22 +0100 Received: from imladris.infradead.org ([194.205.184.45]:38156 "EHLO infradead.org") by humbolt.nl.linux.org with ESMTP id ; Tue, 20 Feb 2001 20:51:40 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]) by infradead.org with esmtp (Exim 3.20 #2) id 14VHDE-0005Rn-00 for linux-mm-bugs@nl.linux.org; Tue, 20 Feb 2001 18:08:34 +0000 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id OAA04012; Tue, 20 Feb 2001 14:59:18 -0300 Date: Tue, 20 Feb 2001 14:59:18 -0300 Message-Id: <200102201759.OAA04012@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1566 (linux mm)] New - vmbalance patch in ac7 triggers OOM killer too often Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1566 Summary: vmbalance patch in ac7 triggers OOM killer too often Product: Linux kernel MM Version: 2.4 (stable) Platform: All URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: normal Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: arjan@fenrus.demon.nl On a big box (2 or more Gb), the OOM killer gets triggered MUCH earlier with the ac7 vmrebalance patch applied. The testset in question had to be reduced by 500Mb (!!) to avoid being killed. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 21 23:49:24 2001 Received: by humbolt.nl.linux.org id ; Wed, 21 Feb 2001 23:49:04 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:36370 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 21 Feb 2001 23:48:37 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id TAA04843; Wed, 21 Feb 2001 19:41:23 -0300 Date: Wed, 21 Feb 2001 19:41:23 -0300 Message-Id: <200102212241.TAA04843@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1551 (linux mm)] Changed - grep causes strange memory behaviour Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1551 riel@conectiva.com.br changed: What |Old Value |New Value ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED Resolution| |INVALID ------- Additional Comments From riel@conectiva.com.br 2001-02-21 19:41 ------- sounds like grep is grepping device nodes for the string you are looking for, and since character devices cannot be mmap()ed, it tries to load all of /dev/core and /dev/mem into memory. No surprise your system is acting funny ... however, this is not a VM-related thing. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Wed Feb 21 23:49:40 2001 Received: by humbolt.nl.linux.org id ; Wed, 21 Feb 2001 23:49:13 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:35090 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Wed, 21 Feb 2001 23:48:29 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id TAA04890; Wed, 21 Feb 2001 19:43:41 -0300 Date: Wed, 21 Feb 2001 19:43:41 -0300 Message-Id: <200102212243.TAA04890@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1566 (linux mm)] Changed - vmbalance patch in ac7 triggers OOM killer too often Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1566 ------- Additional Comments From riel@conectiva.com.br 2001-02-21 19:43 ------- I take it you've ran out of swap at that point ? If that is the case, it's not the vm rebalance which is responsible for this bug, but the fact that we don't reclaim swap space any longer when we have a page both in memory and on swap ... - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Thu Feb 22 10:18:56 2001 Received: by humbolt.nl.linux.org id ; Thu, 22 Feb 2001 10:18:43 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:38926 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 22 Feb 2001 10:18:14 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id GAA16156; Thu, 22 Feb 2001 06:07:36 -0300 Date: Thu, 22 Feb 2001 06:07:36 -0300 Message-Id: <200102220907.GAA16156@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1566 (linux mm)] Changed - vmbalance patch in ac7 triggers OOM killer too often Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1566 ------- Additional Comments From arjan@fenrus.demon.nl 2001-02-22 06:07 ------- How do you explain that the 500Mb "leak" is gone once vmrebalance is backed out? (same everything for the rest) - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Thu Feb 22 15:59:30 2001 Received: by humbolt.nl.linux.org id ; Thu, 22 Feb 2001 15:59:07 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:14341 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Thu, 22 Feb 2001 15:58:17 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id LAA22358; Thu, 22 Feb 2001 11:47:15 -0300 Date: Thu, 22 Feb 2001 11:47:15 -0300 Message-Id: <200102221447.LAA22358@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1594 (linux mm)] New - mm bug Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1594 Summary: mm bug Product: Linux kernel MM Version: 2.4 (stable) Platform: All URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: blocker Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: xose@smi-ps.com try petalo.c and petalo_t like a normal user on kernel-2.4.1 plus patch-2.4.1-ac20 ---petalo.c--- #include #include main() { void *buf[2048]; long int mem = 0x3000000; int i = 0; for(;;) { printf("%lX\n", mem); buf[i] = malloc(mem); if (buf[i] == NULL) mem /= 2; else { memset(buf[i], i, mem); i++; } if(mem < 1) { sleep(1); mem = 0x20000; } } getchar(); } ---end--- --petalo_t.c--- /* * compile with gcc -o petalo petalo.c -lpthread */ #include #include #include #define MAXT 4096 void petalo(void *n); main() { int i; pthread_t th_num[MAXT]; void *retval; for(;;) { for(i = 0; i < MAXT; i++) pthread_create(&th_num[i], NULL,(void*)&petalo,(void*)i); for(i = 0; i < MAXT; i++) pthread_join(th_num[i], &retval); } return 0; } void petalo(void *n) { void *buf[2048]; long int mem = 0x3000000 >> ((int)n); int i = 0; for(;;) { buf[i] = malloc(mem); if (buf[i] == NULL) mem /= 2; else { memset(buf[i], i, mem); i++; } if(mem < 1) { usleep(5000); mem = 0x20000; } } } ---end--- linux-mm with petalo.c is ok, you can more than one time. But if you run petalo_t more than one time the system hang. thanks - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 23 03:08:43 2001 Received: by humbolt.nl.linux.org id ; Fri, 23 Feb 2001 03:08:33 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:40970 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 23 Feb 2001 03:08:13 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id XAA03899; Thu, 22 Feb 2001 23:04:05 -0300 Date: Thu, 22 Feb 2001 23:04:05 -0300 Message-Id: <200102230204.XAA03899@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1594 (linux mm)] Changed - mm bug Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1594 ------- Additional Comments From riel@conectiva.com.br 2001-02-22 23:04 ------- I'm wondering if this is a case of the OOM killer not being triggered (and killing the threads as it should in this case) Could you please test if the Linux system actually runs out of memory AND swap or if it just runs out of memory and for some reason never gets around to swapping ? - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Fri Feb 23 03:59:18 2001 Received: by humbolt.nl.linux.org id ; Fri, 23 Feb 2001 03:58:59 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:31243 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Fri, 23 Feb 2001 03:58:14 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id XAA04765; Thu, 22 Feb 2001 23:52:44 -0300 Date: Thu, 22 Feb 2001 23:52:44 -0300 Message-Id: <200102230252.XAA04765@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1566 (linux mm)] Changed - vmbalance patch in ac7 triggers OOM killer too often Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1566 ------- Additional Comments From riel@conectiva.com.br 2001-02-22 23:52 ------- 1. things get swapped out earlier 2. swap space never gets deallocated 3. because of 1 and 2, you run out of space sooner - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Sat Feb 24 03:00:30 2001 Received: by humbolt.nl.linux.org id ; Sat, 24 Feb 2001 03:00:15 +0100 Received: from p108.usnyc3.stsn.com ([199.106.218.108]:38666 "EHLO localhost.localdomain") by humbolt.nl.linux.org with ESMTP id ; Sat, 24 Feb 2001 02:59:44 +0100 Received: from localhost (riel@localhost) by localhost.localdomain (8.11.2/8.11.2) with ESMTP id f1O21OL11392; Fri, 23 Feb 2001 23:01:29 -0300 X-Authentication-Warning: localhost.localdomain: riel owned process doing -bs Date: Fri, 23 Feb 2001 21:01:24 -0500 (EST) From: Rik van Riel X-X-Sender: To: =?iso-8859-1?Q?Xos=C9=20V=E1zquez?= cc: Subject: Re: [Bug 1594 (linux mm)] Changed - mm bug In-Reply-To: <3A96C7AF.AD915DB1@smi-ps.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list On Fri, 23 Feb 2001, XosÉ Vázquez wrote: > bugzilla-daemon@distro.conectiva.com.br wrote: > > > > http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1594 > > > > --- shadow/1594 Thu Feb 22 11:47:15 2001 > > +++ shadow/1594.tmp.3897 Thu Feb 22 23:04:05 2001 > > @@ -109,3 +109,11 @@ > > if you run petalo_t more than one time the system hang. > > > > thanks > > \ No newline at end of file > > + > > +------- Additional Comments From riel@conectiva.com.br 2001-02-22 23:04 ------- > > +I'm wondering if this is a case of the OOM killer not being triggered (and > > +killing the threads as it should in this case) > > + > > +Could you please test if the Linux system actually runs out of memory AND swap > > +or if it just runs out of memory and for some reason never gets around to > > +swapping ? > > \ No newline at end of file > > The system has 48M_RAM and 32M_SWAP. > But it is only when I run petal_t more than one time. > The next(second) time the system hang. > > How can i to know if the system run out of memory and > don't make swapping the second time ? Press --M when the system is hanging to see if it has any free memory and/or swap left. If it doesn't, or if the amount is very low, it is likely that the OOM killer doesn't trigger and your programs never get killed (and the system will remain hung). regards, Rik -- Virtual memory is like a game you can't win; However, without VM there's truly nothing to lose... http://www.surriel.com/ http://www.conectiva.com/ http://distro.conectiva.com.br/ - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Sun Feb 25 10:34:31 2001 Received: by humbolt.nl.linux.org id ; Sun, 25 Feb 2001 10:34:19 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:33801 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Sun, 25 Feb 2001 10:33:45 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id GAA18737; Sun, 25 Feb 2001 06:26:05 -0300 Date: Sun, 25 Feb 2001 06:26:05 -0300 Message-Id: <200102250926.GAA18737@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1622 (linux mm)] New - Incoherent Count of Users of Task Substructures and Direct Memory Access Bugs in 2.4.2 Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1622 Summary: Incoherent Count of Users of Task Substructures and Direct Memory Access Bugs in 2.4.2 Product: Linux kernel MM Version: 2.4 (stable) Platform: All URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: critical Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: moshe@moelabs.com Bug 1 ----- Direct access to user-memory from kernel: The routines "proc_write_register" and "proc_write_status" of "fs/binfmt_misc.c" wrongly access the "buffer" argument which is in user-space: this is of course not allowed and could cause a kernel-fault if the supplied address is invalid. I have a patch fixing this problem. Bug 2 ----- Incoherent count of users of task sub-structures: Several per-process sub-structures have an atomic "count" field, indicating the number of references to the structure, so that the last user deallocates the structure. This field is often incremented by casual lookers, especially via the /proc file-system, and therefore does not truly represent the number of tasks actually sharing the structure as clones - however, for at least some of those structures, some kernel code wrongly uses the counts in that manner. "fs_struct" and "file_struct": ------------------------------ The security checks in "compute_creds" of "fs/exec.c", lines 687-688, assume that "count" refers to the number of clone-processes using the structures, whereas it may also include onlookers via "/proc": as a result, the "exec" system-call could fail unnecessarily. In fact, this bug could even be used by a hacker, using /proc to cause processes of other users to fail. The solution is to add another atomic field, say "users", to count only processes that actually use the structure as clones. Solution: an extra counter, "clones" was added to both structures. mm_struct: ---------- the "mm_struct" already has both "mm_count" and "mm_users", but neither of them properly conveys the real number of users as "mm_users" is also incremented by "ptrace", "vmscan" and various /proc functions. One outcome, for example, is that a process may randomly fail to produce a core-dump just, for example, because pages happen to be swapped-out at the time, or because some other user is running "ps" or "top". this is because the code in "do_coredump" of "fs/exec.c" looks at "mm_users". Also, in "write_ldt" of "arch/i386/kernel/ldt.c", a confusing warning message may be unnecessarily displayed on the console, again just because someone happened to run "ps" or "top" at a crucial moment. Solution: a third counter, "mm_clones" was added. ====================== The "count" field of the "signal_struct" appears to be safe - for the time being (unless it became possible to watch a process' signal-disposition via /proc or "ptrace"). The patch for Bug 2, is available, too. Please try out both patches and let me know if they work for you as well as they did for me on SMP i386 2.4.2 The patches are at www.moelabs.com/patch01_23 and www.moelabs.com/patch01_24 Kind regards Moshe Bar www.moelabs.com - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 27 02:44:12 2001 Received: by humbolt.nl.linux.org id ; Tue, 27 Feb 2001 02:44:04 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:5391 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Tue, 27 Feb 2001 02:43:43 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id WAA12626; Mon, 26 Feb 2001 22:41:32 -0300 Date: Mon, 26 Feb 2001 22:41:32 -0300 Message-Id: <200102270141.WAA12626@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1631 (linux mm)] New - VM: Bad swap entry Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1631 Summary: VM: Bad swap entry Product: Linux kernel MM Version: 2.4 (stable) Platform: i386 URL: OS/Version: Linux Status: UNCONFIRMED Status Whiteboard: Keywords: Resolution: Severity: normal Priority: P3 Component: linux mm AssignedTo: linux-mm-bugs@nl.linux.org ReportedBy: cesarb@nitnet.com.br I was stress testing the VM in an old 386SX25 with 8MB of memory and 64MB of swap in hda2. The stress test I was using was running the classic memory hog 'dselect' from Debian, while using NTP (locks 26% of the memory I have in the 386; can't turn it off since the clock is drifting badly), and running three ssh sessions from another computer, one with 'vmstat 1', the other one with 'top -s', and the third one idle. After some time (about half an hour -- dselect is evil) I got the first message: VM: Bad swap entry 00186b00 After some more time, I got: VM: Bad swap entry 001ad100 After even more time, I got: VM: Bad swap entry 001fa200 After more time (I didn't see when it appeared): VM: Bad swap entry 002d1b00 And then, out of nothing, I started getting: Unused swap offset entry in swap_count 002e3400 Unused swap offset entry in swap_dup 002e3400 VM: Bad swap entry 002e3400 lots of times (with the same exact values). Sometime before closing the memory hog, the messages ceased appearing. I'll try to attach the dmesg output. - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org From owner-linux-mm-bugs@nl.linux.org Tue Feb 27 02:54:06 2001 Received: by humbolt.nl.linux.org id ; Tue, 27 Feb 2001 02:53:53 +0100 Received: from distro.conectiva.com.br ([200.250.58.149]:7951 "EHLO distro.conectiva.com.br") by humbolt.nl.linux.org with ESMTP id ; Tue, 27 Feb 2001 02:53:40 +0100 Received: (from nobody@localhost) by distro.conectiva.com.br (8.9.3/8.9.3) id WAA12665; Mon, 26 Feb 2001 22:44:17 -0300 Date: Mon, 26 Feb 2001 22:44:17 -0300 Message-Id: <200102270144.WAA12665@distro.conectiva.com.br> From: bugzilla-daemon@distro.conectiva.com.br To: linux-mm-bugs@nl.linux.org Subject: [Bug 1631 (linux mm)] Changed - VM: Bad swap entry Sender: owner-linux-mm-bugs@nl.linux.org Precedence: bulk Reply-To: linux-mm-bugs@nl.linux.org Return-Path: X-Orcpt: rfc822;linux-mm-bugs-list http://distro.conectiva.com.br/bugzilla/show_bug.cgi?id=1631 ------- Additional Comments From cesarb@nitnet.com.br 2001-02-26 22:44 ------- Created an attachment (id=51) dmesg output after the dust settled - Linux-mm-bugs: bugzilla list for the Linux-MM subsystem Archive: http://mail.nl.linux.org/linux-mm-bugs/ Web site: http://www.linux.eu.org/Linux-MM/ Development: linux-mm@kvack.org