Re: [PATCH v1 5/6] rcu: rename kfree_call_rcu()/__kfree_rcu()

Re: [PATCH v1 5/6] rcu: rename kfree_call_rcu()/__kfree_rcu()

[Andrew Morton]

On Mon, 16 Mar 2020 20:01:44 +0100 Uladzislau Rezki <urezki@gmail.com> wrote:

> On Mon, Mar 16, 2020 at 11:25:41AM -0400, Joel Fernandes wrote:
> > On Sun, Mar 15, 2020 at 07:18:39PM +0100, Uladzislau Rezki (Sony) wrote:
> > > Rename kfree_call_rcu() to the kvfree_call_rcu().
> > > The reason is, it is capable of freeing vmalloc()
> > > memory now.
> > > 
> > > Do the same with __kfree_rcu() macro, it becomes
> > > __kvfree_rcu(), the reason is the same as pointed
> > > above.
> > 
> > Vlad, this patch does not apply to my branch that I shared with you. Sorry if
> > I was not clear earlier, could we work on the same branch to avoid conflicts?
> > 
> It was clear to me. Basically i knew that you would be able to apply it
> because of slim changes. I based my work on latest Paul's branch simply
> because that my current setup was based on that, it would take more time
> to switch.
> 
> Next changes i will base on your branch.
> 
> > I based the kfree_rcu shrinker patches on an 'rcu/kfree' branch in my git
> > tree: https://github.com/joelagnel/linux-kernel/tree/rcu/kfree
> > 
> > For now I manually applied 5/6. All others applied cleanly.
> > 
> > Updated the tree as I continue to review your patches.
> >

I'm not sure what's happening here - these patches aren't yet in
linux-next.

A couple of thoughts:

- Please cc linux-mm@kvack.org on this patchset and anything else
  which impacts MM.

- It's a bit strange to create new infrastructure which has but a
  single call site.  Please tell us much more about "there was also
  request/interest so there will be new comers" to set minds at ease. 

  Who/where are these possible new callsites and when can we expect to
  see that code?

Re: [PATCH v1 5/6] rcu: rename kfree_call_rcu()/__kfree_rcu()

[Uladzislau Rezki]

Hello, Andrew.

> > > I based the kfree_rcu shrinker patches on an 'rcu/kfree' branch in my git
> > > tree: https://github.com/joelagnel/linux-kernel/tree/rcu/kfree
> > > 
> > > For now I manually applied 5/6. All others applied cleanly.
> > > 
> > > Updated the tree as I continue to review your patches.
> > >
> 
> I'm not sure what's happening here - these patches aren't yet in
> linux-next.
> 
Please have a look here, i explained in detail in my cover latter
what we would like to achieve: https://lkml.org/lkml/2020/3/23/242

> A couple of thoughts:
> 
> - Please cc linux-mm@kvack.org on this patchset and anything else
>   which impacts MM.
> 
OK. I thought that is related to RCU, therefore i did not want to
spam, but i included you :) Next time we will keep linux-mm@kvack.org
in a loop. Sorry for that.

>
> - It's a bit strange to create new infrastructure which has but a
>   single call site.  Please tell us much more about "there was also
>   request/interest so there will be new comers" to set minds at ease. 
> 
>   Who/where are these possible new callsites and when can we expect to
>   see that code?
> 
https://lkml.org/lkml/2020/3/23/242 here i wrote motivation. Please
check it. If you have some questions please ask.

--
Vlad Rezki

Re: [PATCH v1 5/6] rcu: rename kfree_call_rcu()/__kfree_rcu()

[Joel Fernandes]

On Tue, Apr 21, 2020 at 02:03:58PM +0200, Uladzislau Rezki wrote:
> Hello, Andrew.
> 
> > > > I based the kfree_rcu shrinker patches on an 'rcu/kfree' branch in my git
> > > > tree: https://github.com/joelagnel/linux-kernel/tree/rcu/kfree
> > > > 
> > > > For now I manually applied 5/6. All others applied cleanly.
> > > > 
> > > > Updated the tree as I continue to review your patches.
> > > >
> > 
> > I'm not sure what's happening here - these patches aren't yet in
> > linux-next.
> > 
> Please have a look here, i explained in detail in my cover latter
> what we would like to achieve: https://lkml.org/lkml/2020/3/23/242
> 
> > A couple of thoughts:
> > 
> > - Please cc linux-mm@kvack.org on this patchset and anything else
> >   which impacts MM.
> > 
> OK. I thought that is related to RCU, therefore i did not want to
> spam, but i included you :) Next time we will keep linux-mm@kvack.org
> in a loop. Sorry for that.

Sorry from my side as well. Will keep linux-mm in the loop henceforth.

thanks,

 - Joel


> 
> >
> > - It's a bit strange to create new infrastructure which has but a
> >   single call site.  Please tell us much more about "there was also
> >   request/interest so there will be new comers" to set minds at ease. 
> > 
> >   Who/where are these possible new callsites and when can we expect to
> >   see that code?
> > 
> https://lkml.org/lkml/2020/3/23/242 here i wrote motivation. Please
> check it. If you have some questions please ask.
> 
> --
> Vlad Rezki

Re: [PATCH v1 5/6] rcu: rename kfree_call_rcu()/__kfree_rcu()

[Paul E. McKenney]

On Wed, Apr 22, 2020 at 07:03:22PM -0400, Joel Fernandes wrote:
> On Tue, Apr 21, 2020 at 02:03:58PM +0200, Uladzislau Rezki wrote:
> > Hello, Andrew.
> > 
> > > > > I based the kfree_rcu shrinker patches on an 'rcu/kfree' branch in my git
> > > > > tree: https://github.com/joelagnel/linux-kernel/tree/rcu/kfree
> > > > > 
> > > > > For now I manually applied 5/6. All others applied cleanly.
> > > > > 
> > > > > Updated the tree as I continue to review your patches.
> > > > >
> > > 
> > > I'm not sure what's happening here - these patches aren't yet in
> > > linux-next.
> > > 
> > Please have a look here, i explained in detail in my cover latter
> > what we would like to achieve: https://lkml.org/lkml/2020/3/23/242
> > 
> > > A couple of thoughts:
> > > 
> > > - Please cc linux-mm@kvack.org on this patchset and anything else
> > >   which impacts MM.
> > > 
> > OK. I thought that is related to RCU, therefore i did not want to
> > spam, but i included you :) Next time we will keep linux-mm@kvack.org
> > in a loop. Sorry for that.
> 
> Sorry from my side as well. Will keep linux-mm in the loop henceforth.

I must confess that it came as a bit of a surprise to me as well.

But what is life without a surprise or two along the way?  ;-)

							Thanx, Paul

> thanks,
> 
>  - Joel
> 
> 
> > 
> > >
> > > - It's a bit strange to create new infrastructure which has but a
> > >   single call site.  Please tell us much more about "there was also
> > >   request/interest so there will be new comers" to set minds at ease. 
> > > 
> > >   Who/where are these possible new callsites and when can we expect to
> > >   see that code?
> > > 
> > https://lkml.org/lkml/2020/3/23/242 here i wrote motivation. Please
> > check it. If you have some questions please ask.
> > 
> > --
> > Vlad Rezki