Hello,
I have summarized the topics for the IO controller mini-summit and
written the ideas seen in the mailing list.
- The place where IO controller should be implemented
- Block layer in conjunction with the IO scheduler
- Common layer right above the IO scheduler
- CFQ enhancement.
- Both block and common layer, users can select whichever controller
they want.
- VFS layer
- What kind of bandwidth control policies are needed?
- Proportional weight
- Enforcing upper limit
- Minimum bandwidth guarantee
- How to handle buffered writes?
- Add dirt-ratio in the memory controller
- Add bufferred-write-cgroup to track buffered writebacks
- A per group per bdi pdflush threads
- Who should be charged for swap activity?
- who requests a page.
- who has a page.
- All swap activities are charged to the root group.
And I would also like to discuss about the followings.
- Extensions of struct bio
- Make a bio point to the io_context of a process which creates the
I/O request. This allows to pass the IO scheduling class and
priority information to IO controller even if the IO is submitted
by another process which does not create the request, such as a
worker thread.
- Add a new flag to struct bio to identify the bio as urgent. This
gives IO controller a chance to handle the bio as high
priority. This flag should be set if the bio is created for the
page-out operation.
- Common test methods to verify the functionality of IO controller.
Please give me comments and suggestions. I may be missing or
misunderstanding something.
Thanks,
Ryo Tsuruta
On Wed, Oct 14, 2009 at 09:24:44PM +0900, Ryo Tsuruta wrote:> Hello, >Hi Ryo, CCing people who are planning to attend the mini summit either in person or phone. (As per your list on io mini summit wiki page). Not sure if everybody is scanning mailing list for update on mini summit. I am checking out wiki page for more information like Venue. It says the venue is linux foundation office Japan. Hopefully there is no change in that information. What are the conference call details for the people who might want to join in over phone? Could not find those. Are you yet to post these? Thanks Vivek> I have summarized the topics for the IO controller mini-summit and > written the ideas seen in the mailing list. > > - The place where IO controller should be implemented > - Block layer in conjunction with the IO scheduler > - Common layer right above the IO scheduler > - CFQ enhancement. > - Both block and common layer, users can select whichever controller > they want. > - VFS layer > > - What kind of bandwidth control policies are needed? > - Proportional weight > - Enforcing upper limit > - Minimum bandwidth guarantee > > - How to handle buffered writes? > - Add dirt-ratio in the memory controller > - Add bufferred-write-cgroup to track buffered writebacks > - A per group per bdi pdflush threads > > - Who should be charged for swap activity? > - who requests a page. > - who has a page. > - All swap activities are charged to the root group. > > And I would also like to discuss about the followings. > > - Extensions of struct bio > - Make a bio point to the io_context of a process which creates the > I/O request. This allows to pass the IO scheduling class and > priority information to IO controller even if the IO is submitted > by another process which does not create the request, such as a > worker thread. > - Add a new flag to struct bio to identify the bio as urgent. This > gives IO controller a chance to handle the bio as high > priority. This flag should be set if the bio is created for the > page-out operation. > > - Common test methods to verify the functionality of IO controller. > > Please give me comments and suggestions. I may be missing or > misunderstanding something. > > Thanks, > Ryo Tsuruta > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo at vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/
Hi Vivek, Thank you for CCing me. I just wanted to let you know that Ted Tso was kind enough to let us give a readout for the mini-summit at the kernel summit. There will be other mini-summits sharing 50 minutes so we need to keep each mini-summit readout to 5-6 minutes of presentation and 4-5 minutes of questions/discussions. Yoshikawa-san and myself were planning to take notes of the mini-summit, but it would be great if you could share yours. The idea is to use those to prepare a brief mini-summit report (4-5 slides) that we can show at the kernel summit. Later in the day, after the mini-summit, we would be sending a draft version of the slides to the relevant mailing lists for you to review. Since the mini-summit readouts will take place on Monday I would appreciate it if you could comment on them before Sunday night the latest. Thanks, Fernando Vivek Goyal wrote:> On Wed, Oct 14, 2009 at 09:24:44PM +0900, Ryo Tsuruta wrote: >> Hello, >> > > Hi Ryo, > > CCing people who are planning to attend the mini summit either in person > or phone. (As per your list on io mini summit wiki page). Not sure if > everybody is scanning mailing list for update on mini summit. > > I am checking out wiki page for more information like Venue. It says the > venue is linux foundation office Japan. Hopefully there is no change in > that information. > > What are the conference call details for the people who might want to join > in over phone? Could not find those. Are you yet to post these? > > Thanks > Vivek > >> I have summarized the topics for the IO controller mini-summit and >> written the ideas seen in the mailing list. >> >> - The place where IO controller should be implemented >> - Block layer in conjunction with the IO scheduler >> - Common layer right above the IO scheduler >> - CFQ enhancement. >> - Both block and common layer, users can select whichever controller >> they want. >> - VFS layer >> >> - What kind of bandwidth control policies are needed? >> - Proportional weight >> - Enforcing upper limit >> - Minimum bandwidth guarantee >> >> - How to handle buffered writes? >> - Add dirt-ratio in the memory controller >> - Add bufferred-write-cgroup to track buffered writebacks >> - A per group per bdi pdflush threads >> >> - Who should be charged for swap activity? >> - who requests a page. >> - who has a page. >> - All swap activities are charged to the root group. >> >> And I would also like to discuss about the followings. >> >> - Extensions of struct bio >> - Make a bio point to the io_context of a process which creates the >> I/O request. This allows to pass the IO scheduling class and >> priority information to IO controller even if the IO is submitted >> by another process which does not create the request, such as a >> worker thread. >> - Add a new flag to struct bio to identify the bio as urgent. This >> gives IO controller a chance to handle the bio as high >> priority. This flag should be set if the bio is created for the >> page-out operation. >> >> - Common test methods to verify the functionality of IO controller. >> >> Please give me comments and suggestions. I may be missing or >> misunderstanding something. >> >> Thanks, >> Ryo Tsuruta >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in >> the body of a message to majordomo at vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/
Hi all, As Fernando said below, I will be glad if I can help you by taking notes of the mini summit. I would also be glad if someone helps me take notes. Tsuruta-san, thank you for summarizing the topics. It looks good! The only thing I am worrying about is whether we can summarize our discussions for 5-6 minutes of presentation. I would be happy if we all have it in mind that we can send our message to maintainers, including Ted Tso, by sending our summary. So before starting the discussion, how about taking some time to talk about which topic we want to include in the summary. IMHO, >>> - The place where IO controller should be implemented is the most important, basic, topic, and other things should be summarized so as to explain why we think the place, e.g. block layer, VFS layer, common layer, is better for implementing io controllers. It would also be better if we can answer to the comments we have already received from maintainers, e.g. Andrew, Jens. Thanks, Takuya Yoshikawa Fernando Luis V?zquez Cao wrote:> Hi Vivek, > > Thank you for CCing me. > > I just wanted to let you know that Ted Tso was kind enough to let > us give a readout for the mini-summit at the kernel summit. > > There will be other mini-summits sharing 50 minutes so we need to > keep each mini-summit readout to 5-6 minutes of presentation and > 4-5 minutes of questions/discussions. > > Yoshikawa-san and myself were planning to take notes of the > mini-summit, but it would be great if you could share yours. The > idea is to use those to prepare a brief mini-summit report (4-5 > slides) that we can show at the kernel summit. > > Later in the day, after the mini-summit, we would be sending a > draft version of the slides to the relevant mailing lists for you > to review. Since the mini-summit readouts will take place on > Monday I would appreciate it if you could comment on them before > Sunday night the latest. > > Thanks, > > Fernando > > Vivek Goyal wrote: >> On Wed, Oct 14, 2009 at 09:24:44PM +0900, Ryo Tsuruta wrote: >>> Hello, >>> >> >> Hi Ryo, >> >> CCing people who are planning to attend the mini summit either in person >> or phone. (As per your list on io mini summit wiki page). Not sure if >> everybody is scanning mailing list for update on mini summit. >> >> I am checking out wiki page for more information like Venue. It says the >> venue is linux foundation office Japan. Hopefully there is no change in >> that information. >> >> What are the conference call details for the people who might want to >> join >> in over phone? Could not find those. Are you yet to post these? >> >> Thanks >> Vivek >> >>> I have summarized the topics for the IO controller mini-summit and >>> written the ideas seen in the mailing list. >>> >>> - The place where IO controller should be implemented >>> - Block layer in conjunction with the IO scheduler >>> - Common layer right above the IO scheduler >>> - CFQ enhancement. >>> - Both block and common layer, users can select whichever controller >>> they want. >>> - VFS layer >>> >>> - What kind of bandwidth control policies are needed? >>> - Proportional weight >>> - Enforcing upper limit >>> - Minimum bandwidth guarantee >>> >>> - How to handle buffered writes? >>> - Add dirt-ratio in the memory controller >>> - Add bufferred-write-cgroup to track buffered writebacks >>> - A per group per bdi pdflush threads >>> >>> - Who should be charged for swap activity? >>> - who requests a page. >>> - who has a page. >>> - All swap activities are charged to the root group. >>> >>> And I would also like to discuss about the followings. >>> >>> - Extensions of struct bio >>> - Make a bio point to the io_context of a process which creates the >>> I/O request. This allows to pass the IO scheduling class and >>> priority information to IO controller even if the IO is submitted >>> by another process which does not create the request, such as a >>> worker thread. >>> - Add a new flag to struct bio to identify the bio as urgent. This >>> gives IO controller a chance to handle the bio as high >>> priority. This flag should be set if the bio is created for the >>> page-out operation. >>> - Common test methods to verify the functionality of IO controller. >>> >>> Please give me comments and suggestions. I may be missing or >>> misunderstanding something. >>> >>> Thanks, >>> Ryo Tsuruta >>> -- >>> To unsubscribe from this list: send the line "unsubscribe >>> linux-kernel" in >>> the body of a message to majordomo at vger.kernel.org >>> More majordomo info at http://vger.kernel.org/majordomo-info.html >>> Please read the FAQ at http://www.tux.org/lkml/ > >
Hi Yoshikawa-san and all, Takuya Yoshikawa <yoshikawa.takuya at oss.ntt.co.jp> wrote:> Hi all, > > As Fernando said below, I will be glad if I can help you by taking > notes of the mini summit. I would also be glad if someone helps me > take notes.It is very helpful, thank you very much.> Tsuruta-san, thank you for summarizing the topics. It looks good! > The only thing I am worrying about is whether we can summarize our > discussions for 5-6 minutes of presentation.I would like to do what little I can to help it. And if you permit, I would like to publish the meeting miniutes and the presentation slides on the mini-summit web site.> I would be happy if we all have it in mind that we can send our > message to maintainers, including Ted Tso, by sending our summary. > So before starting the discussion, how about taking some time to > talk about which topic we want to include in the summary.I think it is a good idea to talk about it at the start.> IMHO, > >>> - The place where IO controller should be implemented > is the most important, basic, topic, and other things should be > summarized so as to explain why we think the place, e.g. block layer, > VFS layer, common layer, is better for implementing io controllers. > > It would also be better if we can answer to the comments we have > already received from maintainers, e.g. Andrew, Jens.I would like to do so, too. Thanks, Ryo Tsuruta> Thanks, > Takuya Yoshikawa > > > > > Fernando Luis V?zquez Cao wrote: > > Hi Vivek, > > Thank you for CCing me. > > I just wanted to let you know that Ted Tso was kind enough to let > > us give a readout for the mini-summit at the kernel summit. > > There will be other mini-summits sharing 50 minutes so we need to > > keep each mini-summit readout to 5-6 minutes of presentation and > > 4-5 minutes of questions/discussions. > > Yoshikawa-san and myself were planning to take notes of the > > mini-summit, but it would be great if you could share yours. The > > idea is to use those to prepare a brief mini-summit report (4-5 > > slides) that we can show at the kernel summit. > > Later in the day, after the mini-summit, we would be sending a > > draft version of the slides to the relevant mailing lists for you > > to review. Since the mini-summit readouts will take place on > > Monday I would appreciate it if you could comment on them before > > Sunday night the latest. > > Thanks, > > Fernando > > Vivek Goyal wrote: > >> On Wed, Oct 14, 2009 at 09:24:44PM +0900, Ryo Tsuruta wrote: > >>> Hello, > >>> > >> > >> Hi Ryo, > >> > >> CCing people who are planning to attend the mini summit either in person > >> or phone. (As per your list on io mini summit wiki page). Not sure if > >> everybody is scanning mailing list for update on mini summit. > >> > >> I am checking out wiki page for more information like Venue. It says the > >> venue is linux foundation office Japan. Hopefully there is no change in > >> that information. > >> > >> What are the conference call details for the people who might want to join > >> in over phone? Could not find those. Are you yet to post these? > >> > >> Thanks > >> Vivek > >> > >>> I have summarized the topics for the IO controller mini-summit and > >>> written the ideas seen in the mailing list. > >>> > >>> - The place where IO controller should be implemented > >>> - Block layer in conjunction with the IO scheduler > >>> - Common layer right above the IO scheduler > >>> - CFQ enhancement. > >>> - Both block and common layer, users can select whichever controller > >>> they want. > >>> - VFS layer > >>> > >>> - What kind of bandwidth control policies are needed? > >>> - Proportional weight > >>> - Enforcing upper limit > >>> - Minimum bandwidth guarantee > >>> > >>> - How to handle buffered writes? > >>> - Add dirt-ratio in the memory controller > >>> - Add bufferred-write-cgroup to track buffered writebacks > >>> - A per group per bdi pdflush threads > >>> > >>> - Who should be charged for swap activity? > >>> - who requests a page. > >>> - who has a page. > >>> - All swap activities are charged to the root group. > >>> > >>> And I would also like to discuss about the followings. > >>> > >>> - Extensions of struct bio > >>> - Make a bio point to the io_context of a process which creates the > >>> I/O request. This allows to pass the IO scheduling class and > >>> priority information to IO controller even if the IO is submitted > >>> by another process which does not create the request, such as a > >>> worker thread. > >>> - Add a new flag to struct bio to identify the bio as urgent. This > >>> gives IO controller a chance to handle the bio as high > >>> priority. This flag should be set if the bio is created for the > >>> page-out operation. > >>> - Common test methods to verify the functionality of IO controller. > >>> > >>> Please give me comments and suggestions. I may be missing or > >>> misunderstanding something. > >>> > >>> Thanks, > >>> Ryo Tsuruta > >>> -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > >>> the body of a message to majordomo at vger.kernel.org > >>> More majordomo info at http://vger.kernel.org/majordomo-info.html > >>> Please read the FAQ at http://www.tux.org/lkml/ > > > > -- > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo at vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html > Please read the FAQ at http://www.tux.org/lkml/
Hi all, I've uploaded the conslusions of IO controller Mini-summit 2009 on the web page. Thanks Fernando for creating the slides. http://sourceforge.net/apps/trac/ioband/wiki/iosummit Thank you to all attendees and thank you to the Linux Foundation Japan for providing us with a conference venue and facilities. Thanks, Ryo Tsuruta