Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Kirill Tkhai writes: > On 01.06.2018 18:25, Eric W. Biederman wrote: >> Michal Hocko writes: >> >>> On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: Michal Hocko writes: >>> [...] > Group leader exiting early without tearing down the whole thread > group should be quite rare as

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Kirill Tkhai writes: > On 01.06.2018 18:25, Eric W. Biederman wrote: >> Michal Hocko writes: >> >>> On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: Michal Hocko writes: >>> [...] > Group leader exiting early without tearing down the whole thread > group should be quite rare as

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 01.06.2018 18:25, Eric W. Biederman wrote: > Michal Hocko writes: > >> On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: >>> Michal Hocko writes: >> [...] Group leader exiting early without tearing down the whole thread group should be quite rare as well. No question that somebody

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 01.06.2018 18:25, Eric W. Biederman wrote: > Michal Hocko writes: > >> On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: >>> Michal Hocko writes: >> [...] Group leader exiting early without tearing down the whole thread group should be quite rare as well. No question that somebody

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Mon 04-06-18 09:31:46, Eric W. Biederman wrote: [...] > My key point is that it is easy to trigger which makes the current > mm_update_next_owner a fundamentally flawed design, and something that > needs to be fixed. Ohh, absolutely agreed! I was not trying to argue that part of course. --

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Mon 04-06-18 09:31:46, Eric W. Biederman wrote: [...] > My key point is that it is easy to trigger which makes the current > mm_update_next_owner a fundamentally flawed design, and something that > needs to be fixed. Ohh, absolutely agreed! I was not trying to argue that part of course. --

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Fri 01-06-18 10:25:59, Eric W. Biederman wrote: >> Michal Hocko writes: >> >> > On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: >> >> Michal Hocko writes: >> > [...] >> >> > Group leader exiting early without tearing down the whole thread >> >> > group should be

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Fri 01-06-18 10:25:59, Eric W. Biederman wrote: >> Michal Hocko writes: >> >> > On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: >> >> Michal Hocko writes: >> > [...] >> >> > Group leader exiting early without tearing down the whole thread >> >> > group should be

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Fri 01-06-18 10:25:59, Eric W. Biederman wrote: > Michal Hocko writes: > > > On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: > >> Michal Hocko writes: > > [...] > >> > Group leader exiting early without tearing down the whole thread > >> > group should be quite rare as well. No question

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Fri 01-06-18 10:25:59, Eric W. Biederman wrote: > Michal Hocko writes: > > > On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: > >> Michal Hocko writes: > > [...] > >> > Group leader exiting early without tearing down the whole thread > >> > group should be quite rare as well. No question

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: >> Michal Hocko writes: > [...] >> > Group leader exiting early without tearing down the whole thread >> > group should be quite rare as well. No question that somebody might do >> > that on purpose though... >> >> The

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: >> Michal Hocko writes: > [...] >> > Group leader exiting early without tearing down the whole thread >> > group should be quite rare as well. No question that somebody might do >> > that on purpose though... >> >> The

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: > Michal Hocko writes: [...] > > Group leader exiting early without tearing down the whole thread > > group should be quite rare as well. No question that somebody might do > > that on purpose though... > > The group leader exiting early is a

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Fri 01-06-18 09:32:42, Eric W. Biederman wrote: > Michal Hocko writes: [...] > > Group leader exiting early without tearing down the whole thread > > group should be quite rare as well. No question that somebody might do > > that on purpose though... > > The group leader exiting early is a

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Thu 31-05-18 20:07:28, Eric W. Biederman wrote: >> Michal Hocko writes: >> >> > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: >> >> This function searches for a new mm owner in children and siblings, >> >> and then iterates over all processes in the system in unlikely

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Thu 31-05-18 20:07:28, Eric W. Biederman wrote: >> Michal Hocko writes: >> >> > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: >> >> This function searches for a new mm owner in children and siblings, >> >> and then iterates over all processes in the system in unlikely

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 31-05-18 20:07:28, Eric W. Biederman wrote: > Michal Hocko writes: > > > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: > >> This function searches for a new mm owner in children and siblings, > >> and then iterates over all processes in the system in unlikely case. > >> Despite the case

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 31-05-18 20:07:28, Eric W. Biederman wrote: > Michal Hocko writes: > > > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: > >> This function searches for a new mm owner in children and siblings, > >> and then iterates over all processes in the system in unlikely case. > >> Despite the case

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: >> This function searches for a new mm owner in children and siblings, >> and then iterates over all processes in the system in unlikely case. >> Despite the case is unlikely, its probability growths with the number >> of

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: >> This function searches for a new mm owner in children and siblings, >> and then iterates over all processes in the system in unlikely case. >> Despite the case is unlikely, its probability growths with the number >> of

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 27.04.2018 21:05, Eric W. Biederman wrote: > Michal Hocko writes: > >> On Thu 26-04-18 21:28:18, Michal Hocko wrote: >>> On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: Michal Hocko writes: > I've had a patch to remove owner few years

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 27.04.2018 21:05, Eric W. Biederman wrote: > Michal Hocko writes: > >> On Thu 26-04-18 21:28:18, Michal Hocko wrote: >>> On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: Michal Hocko writes: > I've had a patch to remove owner few years back. It needed some work > to

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

ebied...@xmission.com (Eric W. Biederman) writes: > Kirill Tkhai do you think you would be able adapt Michal Hoko's old > patch at https://marc.info/?l=linux-kernel=143635857131756=2 > that replaces mm->owner with mm->memcg? Ouch. At least compared to the current kernel your patch is wide of

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

ebied...@xmission.com (Eric W. Biederman) writes: > Kirill Tkhai do you think you would be able adapt Michal Hoko's old > patch at https://marc.info/?l=linux-kernel=143635857131756=2 > that replaces mm->owner with mm->memcg? Ouch. At least compared to the current kernel your patch is wide of

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Thu 26-04-18 21:28:18, Michal Hocko wrote: >> On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: >> > Michal Hocko writes: >> > >> > > I've had a patch to remove owner few years back. It needed some work >> > > to finish but maybe

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > On Thu 26-04-18 21:28:18, Michal Hocko wrote: >> On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: >> > Michal Hocko writes: >> > >> > > I've had a patch to remove owner few years back. It needed some work >> > > to finish but maybe that would be a better than try to

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 26-04-18 21:28:18, Michal Hocko wrote: > On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: > > Michal Hocko writes: > > > > > I've had a patch to remove owner few years back. It needed some work > > > to finish but maybe that would be a better than try to make > > >

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 26-04-18 21:28:18, Michal Hocko wrote: > On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: > > Michal Hocko writes: > > > > > I've had a patch to remove owner few years back. It needed some work > > > to finish but maybe that would be a better than try to make > > > non-scalable thing

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: > Michal Hocko writes: > > > I've had a patch to remove owner few years back. It needed some work > > to finish but maybe that would be a better than try to make > > non-scalable thing suck less. > > I have a question. Would

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 26-04-18 11:19:33, Eric W. Biederman wrote: > Michal Hocko writes: > > > I've had a patch to remove owner few years back. It needed some work > > to finish but maybe that would be a better than try to make > > non-scalable thing suck less. > > I have a question. Would it be reasonable

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > I've had a patch to remove owner few years back. It needed some work > to finish but maybe that would be a better than try to make > non-scalable thing suck less. I have a question. Would it be reasonable to just have a mm->memcg? That would appear to

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

Michal Hocko writes: > I've had a patch to remove owner few years back. It needed some work > to finish but maybe that would be a better than try to make > non-scalable thing suck less. I have a question. Would it be reasonable to just have a mm->memcg? That would appear to be the simplest

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 04/26, Kirill Tkhai wrote: > > We can rework this simply by adding a list of tasks to mm. Perhaps, but then I think this list should not depend on mm->owner. I mean, mm->list_of_group_leaders_which_use_this_mm can be used by coredump and oom-killer at least. But this is not that simple...

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 04/26, Kirill Tkhai wrote: > > We can rework this simply by adding a list of tasks to mm. Perhaps, but then I think this list should not depend on mm->owner. I mean, mm->list_of_group_leaders_which_use_this_mm can be used by coredump and oom-killer at least. But this is not that simple...

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 26.04.2018 16:07, Michal Hocko wrote: > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: >> This function searches for a new mm owner in children and siblings, >> and then iterates over all processes in the system in unlikely case. >> Despite the case is unlikely, its probability growths with the

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 26.04.2018 16:07, Michal Hocko wrote: > On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: >> This function searches for a new mm owner in children and siblings, >> and then iterates over all processes in the system in unlikely case. >> Despite the case is unlikely, its probability growths with the

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 04/26, Michal Hocko wrote: > > then I suspect that we should be better off reconsidering > mm->owner and try to come up with something more clever. I've had a > patch to remove owner few years back. Yeees, I do remember we discussed this but I forgot everything. I agree, it would nice to turn

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On 04/26, Michal Hocko wrote: > > then I suspect that we should be better off reconsidering > mm->owner and try to come up with something more clever. I've had a > patch to remove owner few years back. Yeees, I do remember we discussed this but I forgot everything. I agree, it would nice to turn

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: > This function searches for a new mm owner in children and siblings, > and then iterates over all processes in the system in unlikely case. > Despite the case is unlikely, its probability growths with the number > of processes in the system. The time,

Re: [PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

On Thu 26-04-18 14:00:19, Kirill Tkhai wrote: > This function searches for a new mm owner in children and siblings, > and then iterates over all processes in the system in unlikely case. > Despite the case is unlikely, its probability growths with the number > of processes in the system. The time,

[PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

This function searches for a new mm owner in children and siblings, and then iterates over all processes in the system in unlikely case. Despite the case is unlikely, its probability growths with the number of processes in the system. The time, spent on iterations, also growths. I regulary observe

[PATCH 0/4] exit: Make unlikely case in mm_update_next_owner() more scalable

This function searches for a new mm owner in children and siblings, and then iterates over all processes in the system in unlikely case. Despite the case is unlikely, its probability growths with the number of processes in the system. The time, spent on iterations, also growths. I regulary observe