From: Jerome Marchand <jmarchan@redhat.com>
To: Vivek Goyal <vgoyal@redhat.com>
Cc: linux-kernel@vger.kernel.org,
containers@lists.linux-foundation.org, dm-devel@redhat.com,
jens.axboe@oracle.com, ryov@valinux.co.jp,
balbir@linux.vnet.ibm.com, righi.andrea@gmail.com,
nauman@google.com, dpshah@google.com, lizf@cn.fujitsu.com,
mikew@google.com, fchecconi@gmail.com, paolo.valente@unimore.it,
fernando@oss.ntt.co.jp, s-uchida@ap.jp.nec.com,
taka@valinux.co.jp, guijianfeng@cn.fujitsu.com,
jmoyer@redhat.com, dhaval@linux.vnet.ibm.com,
m-ikeda@ds.jp.nec.com, agk@redhat.com, akpm@linux-foundation.org,
peterz@infradead.org
Subject: Re: [PATCH 02/24] io-controller: Core of the elevator fair queuing
Date: Thu, 20 Aug 2009 16:51:46 +0200 [thread overview]
Message-ID: <4A8D6302.3080301@redhat.com> (raw)
In-Reply-To: <20090819184142.GD4391@redhat.com>
Vivek Goyal wrote:
> On Wed, Aug 19, 2009 at 06:01:34PM +0200, Jerome Marchand wrote:
>> Hi Vivek,
>>
>> Vivek Goyal wrote:
>>> o This is core of the io scheduler implemented at elevator layer. This is a mix
>>> of cpu CFS scheduler and CFQ IO scheduler. Some of the bits from CFS have
>>> to be derived so that we can support hierarchical scheduling. Without
>>> cgroups or with-in group, we should essentially get same behavior as CFQ.
>>>
>>> o This patch only shows non-hierarchical bits. Hierarhical code comes in later
>>> patches.
>>>
>>> o This code is the building base of introducing fair queuing logic in common
>>> elevator layer so that it can be used by all the four IO schedulers.
>>> +static void enqueue_io_entity(struct io_entity *entity)
>>> +{
>>> + struct io_service_tree *st = entity->st;
>>> + struct io_sched_data *sd = io_entity_sched_data(entity);
>>> +
>>> + /* In case task ioprio class changed while entity was off tree */
>>> + io_entity_update_prio(entity);
>>> + st->nr_active++;
>>> + sd->nr_active++;
>>> + entity->on_st = 1;
>>> + place_entity(st, entity, 0);
>>> + __enqueue_io_entity(st, entity);
>>> +}
One more thing. io_entity_update_prio(entity) can change entity->st, so we'd
better set st after io_entity_update_prio() call. I think it fixes a bug I've
seen when changing elevator and ioprio_class at the same time:
------------[ cut here ]------------
kernel BUG at block/elevator-fq.c:1921!
invalid opcode: 0000 [#1] SMP
last sysfs file: /sys/block/sda/queue/scheduler
Modules linked in: ipt_REJECT xt_tcpudp iptable_filter ip_tables x_tables bridge stp ipv6 autofs4 hidp rfcomm l2cap bluetooth rfkill sunrpc dm_mirror dm_region_hash dm_log dm_multipath scsi_dh dm_mod video output sbs sbshc battery ac netconsole lp snd_hda_codec_analog snd_hda_intel snd_hda_codec snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device sg snd_pcm_oss snd_mixer_oss rtc_cmos sr_mod snd_pcm cdrom rtc_core tg3 snd_timer libphy snd serio_raw dcdbas i2c_i801 soundcore button pcspkr rtc_lib snd_page_alloc i2c_core parport_pc parport ata_piix libatasd_mod scsi_mod ext3 jbd uhci_hcd ohci_hcd ehci_hcd [last unloaded: microcode]
Pid: 9127, comm: change_elevator Tainted: G M (2.6.31-rc6-io-controller-v8 #69) OptiPlex 745
EIP: 0060:[<c0537867>] EFLAGS: 00010286 CPU: 1
EIP is at elv_put_iog+0x61/0x83
EAX: 00000000 EBX: 00000000 ECX: f607c000 EDX: f607c001
ESI: f725d300 EDI: f3b59ee0 EBP: f60e5ec0 ESP: f60e5ebc
DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
Process change_elevator (pid: 9127, ti=f60e5000 task=f38083f0 task.ti=f60e5000)
Stack:
f607c000 f60e5ee8 c05382f5 f70ab480 f6abb368 00000086 f70ab480 00000000
<0> f725d300 f725d334 f38d3cc0 f60e5ef8 c05248f7 f725d30c f6abb368 f60e5f28
<0> c0524a67 00000005 00000000 f725d300 706f6f6e c0820000 f60e5f28 c069e51b
Call Trace:
[<c05382f5>] ? elv_exit_fq_data+0x18d/0x195
[<c05248f7>] ? elevator_exit+0x28/0x4e
[<c0524a67>] ? elv_iosched_store+0x14a/0x1db
[<c069e51b>] ? mutex_lock_nested+0x2e/0x36
[<c052d1f8>] ? queue_attr_store+0x50/0x61
[<c04e9f56>] ? sysfs_write_file+0xb9/0xe4
[<c04e9e9d>] ? sysfs_write_file+0x0/0xe4
[<c04a9e57>] ? vfs_write+0x84/0xdf
[<c04a9f4b>] ? sys_write+0x3b/0x60
[<c04029b4>] ? sysenter_do_call+0x12/0x32
Code: 0b eb fe 83 79 48 00 75 18 83 79 4c 00 75 1c 83 79 60 00 75 0c 83 79 64 00 75 10 83 79 78 00 74 04 0f 0b eb fe 83 79 7c 00 74 04 <0f> 0b eb fe 8d 81 ec 00 00 00 ba f6 65 53 c0 e8 e9 41 f3 ff 85
EIP: [<c0537867>] elv_put_iog+0x61/0x83 SS:ESP 0068:f60e5ebc
---[ end trace f56eaa89c47eb665 ]---
With the following patch, I can't reproduce it anymore.
Regards,
Jerome
Signed-off-by: Jerome Marchand <jmarchan@redhat.com>
---
elevator-fq.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
--- block/elevator-fq.c.orig 2009-08-20 16:39:13.000000000 +0200
+++ block/elevator-fq.c 2009-08-20 14:29:04.000000000 +0200
@@ -519,11 +519,12 @@ __enqueue_io_entity(struct io_service_tr
static void enqueue_io_entity(struct io_entity *entity)
{
- struct io_service_tree *st = entity->st;
+ struct io_service_tree *st;
struct io_sched_data *sd = io_entity_sched_data(entity);
/* In case task ioprio class changed while entity was off tree */
io_entity_update_prio(entity);
+ st = entity->st;
st->nr_active++;
sd->nr_active++;
entity->on_st = 1;
next prev parent reply other threads:[~2009-08-20 14:54 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-08-16 19:30 [RFC] IO scheduler based IO controller V8 Vivek Goyal
2009-08-16 19:30 ` [PATCH 01/24] io-controller: Documentation Vivek Goyal
2009-08-25 3:36 ` Rik van Riel
2009-08-16 19:30 ` [PATCH 02/24] io-controller: Core of the elevator fair queuing Vivek Goyal
2009-08-17 5:29 ` Gui Jianfeng
2009-08-17 20:37 ` Vivek Goyal
2009-08-19 16:01 ` Jerome Marchand
2009-08-19 18:41 ` Vivek Goyal
2009-08-20 14:51 ` Jerome Marchand [this message]
2009-08-20 15:04 ` Vivek Goyal
2009-08-19 18:30 ` Vivek Goyal
2009-08-21 1:54 ` Gui Jianfeng
2009-08-21 2:00 ` Vivek Goyal
2009-08-27 2:49 ` Gui Jianfeng
2009-08-27 21:08 ` Vivek Goyal
2009-08-16 19:30 ` [PATCH 03/24] io-controller: Common flat fair queuing code in elevaotor layer Vivek Goyal
2009-08-19 3:36 ` Gui Jianfeng
2009-08-19 18:39 ` Vivek Goyal
2009-08-16 19:30 ` [PATCH 04/24] io-controller: Modify cfq to make use of flat elevator fair queuing Vivek Goyal
2009-08-16 19:30 ` [PATCH 05/24] io-controller: Core scheduler changes to support hierarhical scheduling Vivek Goyal
2009-08-16 19:30 ` [PATCH 06/24] io-controller: cgroup related changes for hierarchical group support Vivek Goyal
2009-08-16 19:30 ` [PATCH 07/24] io-controller: Common hierarchical fair queuing code in elevaotor layer Vivek Goyal
2009-08-16 19:30 ` [PATCH 08/24] io-controller: cfq changes to use " Vivek Goyal
2009-08-16 19:30 ` [PATCH 09/24] io-controller: Export disk time used and nr sectors dipatched through cgroups Vivek Goyal
2009-08-16 19:30 ` [PATCH 10/24] io-controller: Debug hierarchical IO scheduling Vivek Goyal
2009-08-16 19:30 ` [PATCH 11/24] io-controller: Introduce group idling Vivek Goyal
2009-08-20 1:46 ` [PATCH] IO-Controller: clear ioq wait flag if a request goes into that ioq Gui Jianfeng
2009-08-20 13:42 ` Vivek Goyal
2009-08-21 0:57 ` Gui Jianfeng
2009-08-28 1:12 ` [PATCH 11/24] io-controller: Introduce group idling Gui Jianfeng
2009-08-16 19:30 ` [PATCH 12/24] io-controller: Wait for requests to complete from last queue before new queue is scheduled Vivek Goyal
2009-08-24 3:30 ` Gui Jianfeng
2009-08-16 19:30 ` [PATCH 13/24] io-controller: Separate out queue and data Vivek Goyal
2009-08-16 19:30 ` [PATCH 14/24] io-conroller: Prepare elevator layer for single queue schedulers Vivek Goyal
2009-08-16 19:30 ` [PATCH 15/24] io-controller: noop changes for hierarchical fair queuing Vivek Goyal
2009-08-16 19:30 ` [PATCH 16/24] io-controller: deadline " Vivek Goyal
2009-08-16 19:30 ` [PATCH 17/24] io-controller: anticipatory " Vivek Goyal
2009-08-16 19:30 ` [PATCH 18/24] blkio_cgroup patches from Ryo to track async bios Vivek Goyal
2009-08-18 11:42 ` Ryo Tsuruta
2009-08-18 14:26 ` Vivek Goyal
2009-08-19 1:43 ` Ryo Tsuruta
2009-08-16 19:30 ` [PATCH 19/24] io-controller: map async requests to appropriate cgroup Vivek Goyal
2009-08-16 19:30 ` [PATCH 20/24] io-controller: Per cgroup request descriptor support Vivek Goyal
2009-08-16 19:30 ` [PATCH 21/24] io-controller: Per io group bdi congestion interface Vivek Goyal
2009-08-16 19:30 ` [PATCH 22/24] io-controller: Support per cgroup per device weights and io class Vivek Goyal
2009-08-16 19:30 ` [PATCH 23/24] io-controller: map sync requests to group using bio tracking info Vivek Goyal
2009-08-16 19:30 ` [PATCH 24/24] io-controller: debug elevator fair queuing support Vivek Goyal
2009-08-16 19:53 ` [RFC] IO scheduler based IO controller V8 Vivek Goyal
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4A8D6302.3080301@redhat.com \
--to=jmarchan@redhat.com \
--cc=agk@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=balbir@linux.vnet.ibm.com \
--cc=containers@lists.linux-foundation.org \
--cc=dhaval@linux.vnet.ibm.com \
--cc=dm-devel@redhat.com \
--cc=dpshah@google.com \
--cc=fchecconi@gmail.com \
--cc=fernando@oss.ntt.co.jp \
--cc=guijianfeng@cn.fujitsu.com \
--cc=jens.axboe@oracle.com \
--cc=jmoyer@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lizf@cn.fujitsu.com \
--cc=m-ikeda@ds.jp.nec.com \
--cc=mikew@google.com \
--cc=nauman@google.com \
--cc=paolo.valente@unimore.it \
--cc=peterz@infradead.org \
--cc=righi.andrea@gmail.com \
--cc=ryov@valinux.co.jp \
--cc=s-uchida@ap.jp.nec.com \
--cc=taka@valinux.co.jp \
--cc=vgoyal@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).