login
Search
==
From: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Documentation update for memory cgroup
Some informations are old, and I think current document doesn't work
as "a guide for users".
We need summary of all of our controls, at least.
This patch updates information for current implementations and add a
summary of interfaces. etc...
Changelog:
- fixed tons of typos.
- replaced "memcg" with "memory cgroup" AMAP.
- replaced "mem+swap" with "memory+swap"
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
Documentation/cgroups/memory.txt | 210 ++++++++++++++++++++++++++++-----------
1 file changed, 152 insertions(+), 58 deletions(-)
Index: mmotm-temp/Documentation/cgroups/memory.txt
===================================================================
--- mmotm-temp.orig/Documentation/cgroups/memory.txt
+++ mmotm-temp/Documentation/cgroups/memory.txt
@@ -4,16 +4,6 @@ NOTE: The Memory Resource Controller has
to as the memory controller in this document. Do not confuse memory controller
used here with the memory controller that is used in hardware.
-Salient features
-
-a. Enable control of Anonymous, Page Cache (mapped and unmapped) and
- Swap Cache memory pages.
-b. The infrastructure allows easy addition of other types of memory to control
-c. Provides *zero overhead* for non memory controller users
-d. Provides a double LRU: global memory pressure causes reclaim from the
- global LRU; a cgroup on hitting a limit, reclaims from the per
- cgroup LRU
-
Benefits and Purpose of the memory controller
The memory controller isolates the memory behaviour of a group of tasks
@@ -33,6 +23,45 @@ d. A CD/DVD burner could control the amo
e. There are several other use cases, find one or use the controller just
for fun (to learn and hack on the VM subsystem).
+Current Status: linux-2.6.34-mmotm(development version of 2010/April)
+
+Features:
+ - accounting anonymous pages, file caches, swap caches usage and limit them.
+ ...It would be better to use LRU throughout the file instead of using LRU in some --- ~Randy --
On Thu, 8 Apr 2010 10:32:09 -0700 Ragards. --
It seems nice overall. Thank you for your work. I have a few minor comments. IIUC, writing to cgroup.procs isn't supported yet. So, I think we don't have to Is this line necessary ? We say the similar thing above("A PageCache page is hmm, chapter 2 is used as "2. Memory Control" already :( Otherwise, it looks good to me. Thanks, Daisuke Nishimura. --
On Fri, 9 Apr 2010 10:26:26 +0900 It's supported. See Documetaion/cgroup/cgroup.txt Thanks, --
I can't write to it now.
# echo $$ >/cgroup/memory/01/cgroup.procs
-bash: echo: write error: Invalid argument
# ls -l /cgroup/memory/01/cgroup.procs
-r--r--r-- 1 root root 0 2010-04-09 10:41 /cgroup/memory/01/cgroup.procs
# uname -a
Linux GibsonE 2.6.34-rc3-mm1-00432-g37c11f5 #1 SMP Thu Apr 8 11:03:39 JST 2010 x86_64 x86_64 x86_64 GNU/Linux
And kernel/cgroup.c says:
3161 {
3162 .name = CGROUP_FILE_GENERIC_PREFIX "procs",
3163 .open = cgroup_procs_open,
3164 /* .write_u64 = cgroup_procs_write, TODO */
3165 .release = cgroup_pidlist_release,
3166 .mode = S_IRUGO,
3167 },
IIRC, it's supported once, but the patch was dropped.
Thanks,
Daisuke Nishimura.
--
On Fri, 9 Apr 2010 11:11:45 +0900 Ouch....I have to fix (our internal) user manual, too..... Thanks, -Kame --
Thank you all.
==
Documentation update.
Some information are old, and I think current documentation doesn't work
as "a guide for users".
We need summary of all of our controls, at least.
Changelog: 2010/04/09
* replace 'lru' with 'LRU' and 'oom' with 'OOM'
* fixed double-space breakage
* applied all comments and fixed wrong parts pointed out.
* fixed cgroup.procs
Changelog: 2009/04/07
* fixed tons of typos.
* replaced "memcg" with "memory cgroup" AMAP.
* replaced "mem+swap" with "memory+swap"
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
Documentation/cgroups/memory.txt | 277 ++++++++++++++++++++++++++-------------
1 file changed, 188 insertions(+), 89 deletions(-)
Index: mmotm-temp/Documentation/cgroups/memory.txt
===================================================================
--- mmotm-temp.orig/Documentation/cgroups/memory.txt
+++ mmotm-temp/Documentation/cgroups/memory.txt
@@ -4,16 +4,6 @@ NOTE: The Memory Resource Controller has
to as the memory controller in this document. Do not confuse memory controller
used here with the memory controller that is used in hardware.
-Salient features
-
-a. Enable control of Anonymous, Page Cache (mapped and unmapped) and
- Swap Cache memory pages.
-b. The infrastructure allows easy addition of other types of memory to control
-c. Provides *zero overhead* for non memory controller users
-d. Provides a double LRU: global memory pressure causes reclaim from the
- global LRU; a cgroup on hitting a limit, reclaims from the per
- cgroup LRU
-
Benefits and Purpose of the memory controller
The memory controller isolates the memory behaviour of a group of tasks
@@ -33,6 +23,45 @@ d. A CD/DVD burner could control the amo
e. There are several other use cases, find one or use the controller just
for fun (to learn and hack on the VM subsystem).
+Current Status: linux-2.6.34-mmotm(development version of 2010/April)
+
+Features:
+ - accounting anonymous pages, file ...Looks good to me. Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Thanks, --
I would have said: Why remove those leading spaces (indent/text alignment)? Almost there. :) thanks, --- ~Randy --
Thank you very much for your kindly helps!.
-Kame
=
Documentation update.
Some information are old, and I think current document doesn't work
as "a guide for users".
We need summary of all of our controls, at least.
Changelog: 2010/04/12
* applied feedback
Changelog: 2010/04/09
* replace 'lru' with 'LRU' and 'oom' with 'OOM'
* fixed double-space breakage
* applied all comments and fixed wrong parts pointed out.
* fixed cgroup.procs
Changelog: 2009/04/07
* fixed tons of typos.
* replaced "memcg" with "memory cgroup" AMAP.
* replaced "mem+swap" with "memory+swap"
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
Documentation/cgroups/memory.txt | 276 ++++++++++++++++++++++++++-------------
1 file changed, 188 insertions(+), 88 deletions(-)
Index: mmotm-temp/Documentation/cgroups/memory.txt
===================================================================
--- mmotm-temp.orig/Documentation/cgroups/memory.txt
+++ mmotm-temp/Documentation/cgroups/memory.txt
@@ -4,16 +4,6 @@ NOTE: The Memory Resource Controller has
to as the memory controller in this document. Do not confuse memory controller
used here with the memory controller that is used in hardware.
-Salient features
-
-a. Enable control of Anonymous, Page Cache (mapped and unmapped) and
- Swap Cache memory pages.
-b. The infrastructure allows easy addition of other types of memory to control
-c. Provides *zero overhead* for non memory controller users
-d. Provides a double LRU: global memory pressure causes reclaim from the
- global LRU; a cgroup on hitting a limit, reclaims from the per
- cgroup LRU
-
Benefits and Purpose of the memory controller
The memory controller isolates the memory behaviour of a group of tasks
@@ -33,6 +23,45 @@ d. A CD/DVD burner could control the amo
e. There are several other use cases, find one or use the controller just
for fun (to learn and hack on the VM subsystem).
+Current Status: linux-2.6.34-mmotm(development version ...^^ period might not be appropriate, if it is, n (not) should -- Three Cheers, Balbir --
On Tue, 13 Apr 2010 11:34:06 +0530 Hmm ? I'm not sure. Could you explain why you think removing <> is better ? Thanks, -Kame --
* KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> [2010-04-13 15:20:48]: [snip] Because it is an actual name as compared to using <> for to be replaced with the right thing. -- Three Cheers, Balbir --
On Tue, 13 Apr 2010 12:18:55 +0530 Should I replace TABs with SPACEs ? I think my mailer doesn't Did you read cgroup eventer listener ? "write string like "<event_fd> <memory.usage_in_bytes> <threshold>" to" means <fd of event fd> <fd of memory.usage_in_bytes> <threshold> Then, it's not actual name. It's can be changed. Hmm, I'll replace them with <fd of memory.usage_in_bytes> Here. == Documentation update. Some information are old, and I think current document doesn't work as "a guide for users". We need summary of all of our controls, at least. Changelog: 2010/04/13 * applied feedback * fixed <memory.usage_in_bytes> with <fd of memory.usage_in_bytes> at explaining how-to-use eventfd. Changelog: 2010/04/12 * applied feedback Changelog: 2010/04/09 * replace 'lru' with 'LRU' and 'oom' with 'OOM' * fixed double-space breakage * applied all comments and fixed wrong parts pointed out. * fixed cgroup.procs Changelog: 2009/04/07 * fixed tons of typos. * replaced "memcg" with "memory cgroup" AMAP. * replaced "mem+swap" with "memory+swap" Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> --- Documentation/cgroups/memory.txt | 277 ++++++++++++++++++++++++++------------- 1 file changed, 188 insertions(+), 89 deletions(-) Index: mmotm-temp/Documentation/cgroups/memory.txt =================================================================== --- mmotm-temp.orig/Documentation/cgroups/memory.txt +++ mmotm-temp/Documentation/cgroups/memory.txt @@ -4,16 +4,6 @@ NOTE: The Memory Resource Controller has to as the memory controller in this document. Do not confuse memory controller used here with the memory controller that is used in hardware. -Salient features - -a. Enable control of Anonymous, Page Cache (mapped and unmapped) and - Swap Cache memory pages. -b. The infrastructure allows easy addition of other types of memory to control -c. Provides *zero overhead* for non memory controller users -d. Provides a double LRU: ...
using a ; there would be even better. we -- ~Randy --
On Tue, 13 Apr 2010 08:54:28 -0700 Ah, I thought I changed that ....seems my mistake, sorry. ==here== Documentation update. Some information are old, and I think current document doesn't work as "a guide for users". We need summary of all of our controls, at least. Changelog: 2010/04/14 * applied feedback * Add a text about "memory cgroup" at the place explaining "memory controller" * adjusted onto Nishimura's memcg-move-charge-of-file-pages.patch (add an fix mmaped -> mmapped...Hmm? mapped is better ?) Changelog: 2010/04/13 * applied feedback * fixed <memory.usage_in_bytes> with <fd of memory.usage_in_bytes> at explaining how-to-use eventfd. Changelog: 2010/04/12 * applied feedback Changelog: 2010/04/09 * replace 'lru' with 'LRU' and 'oom' with 'OOM' * fixed double-space breakage * applied all comments and fixed wrong parts pointed out. * fixed cgroup.procs Changelog: 2009/04/07 * fixed tons of typos. * replaced "memcg" with "memory cgroup" AMAP. * replaced "mem+swap" with "memory+swap" Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> --- Documentation/cgroups/memory.txt | 289 ++++++++++++++++++++++++++------------- 1 file changed, 197 insertions(+), 92 deletions(-) Index: mmotm-temp/Documentation/cgroups/memory.txt =================================================================== --- mmotm-temp.orig/Documentation/cgroups/memory.txt +++ mmotm-temp/Documentation/cgroups/memory.txt @@ -1,18 +1,15 @@ Memory Resource Controller NOTE: The Memory Resource Controller has been generically been referred -to as the memory controller in this document. Do not confuse memory controller -used here with the memory controller that is used in hardware. + to as the memory controller in this document. Do not confuse memory + controller used here with the memory controller that is used in hardware. -Salient features - -a. Enable control of Anonymous, Page Cache (mapped and unmapped) and - Swap Cache memory pages. -b. ...
BTW: it would be a lot easier [for reviewing] if you could freeze (or merge) this version and then apply fixes on top of it with a different (and shorter) patch. Reviewed-by: Randy Dunlap <randy.dunlap@oracle.com> --- ~Randy --
On Wed, 14 Apr 2010 11:11:46 -0700 Thank you very much for your patient review. It has been very helpful. Here is fixed one. == Documentation update. Some information are old, and I think current document doesn't work as "a guide for users". We need summary of all of our controls, at least. Changelog: 2010/04/15 * applied feedback Changelog: 2010/04/14 * applied feedback * Add a text about "memory cgroup" at the place explaining "memory controller" * adjusted onto Nishimura's memcg-move-charge-of-file-pages.patch (add an fix mmaped -> mmapped...Hmm? mapped is better ?) Changelog: 2010/04/13 * applied feedback * fixed <memory.usage_in_bytes> with <fd of memory.usage_in_bytes> at explaining how-to-use eventfd. Changelog: 2010/04/12 * applied feedback Changelog: 2010/04/09 * replace 'lru' with 'LRU' and 'oom' with 'OOM' * fixed double-space breakage * applied all comments and fixed wrong parts pointed out. * fixed cgroup.procs Changelog: 2009/04/07 * fixed tons of typos. * replaced "memcg" with "memory cgroup" AMAP. * replaced "mem+swap" with "memory+swap" Reviewed-by: Randy Dunlap <randy.dunlap@oracle.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> --- Documentation/cgroups/memory.txt | 291 ++++++++++++++++++++++++++------------- 1 file changed, 198 insertions(+), 93 deletions(-) Index: mmotm-temp/Documentation/cgroups/memory.txt =================================================================== --- mmotm-temp.orig/Documentation/cgroups/memory.txt +++ mmotm-temp/Documentation/cgroups/memory.txt @@ -1,18 +1,15 @@ Memory Resource Controller NOTE: The Memory Resource Controller has been generically been referred -to as the memory controller in this document. Do not confuse memory controller -used here with the memory controller that is used in hardware. + to as the memory controller in this document. Do not confuse memory + controller used here with the memory controller that is used in hardware. ...
On Tue, Apr 13, 2010 at 01:45:53PM +0900, KAMEZAWA Hiroyuki wrote: Because I never understood very well, I will ask. Why lock_page_cgroup() should not be called under mapping->tree_lock? Thanks Vivek --
Typed wrong email id last time and mail bounced. So here is another --
The closest reference I can find to a conversation regarding this is http://linux.derkeiler.com/Mailing-Lists/Kernel/2009-05/msg05158.html -- Three Cheers, Balbir --
Thanks Balbir. So basically idea is that page_cgroup_lock() does not disable interrupts hence can be interrupted. So don't do lock_page_cgroup() in interrupt context at all otherwise it can lead to various kind of deadlock scenarios. One of those scenarios is lock_page_cgroup() under mapping->tree_lock. That helps. Thanks --
