Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hello, Tahsin. On Mon, Feb 27, 2017 at 12:37:59PM -0800, Tahsin Erdogan wrote: > > Ah, absolutely, that's a stupid failure but we should be able to fix > > that by making the blkg functions take gfp mask and allocate > > accordingly, right? It'll probably take preallocation tricks because > > of

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hello, Tahsin. On Mon, Feb 27, 2017 at 12:37:59PM -0800, Tahsin Erdogan wrote: > > Ah, absolutely, that's a stupid failure but we should be able to fix > > that by making the blkg functions take gfp mask and allocate > > accordingly, right? It'll probably take preallocation tricks because > > of

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hello, On Mon, Feb 27, 2017 at 05:00:31AM -0800, Tahsin Erdogan wrote: > On Mon, Feb 27, 2017 at 1:52 AM, Michal Hocko wrote: > > On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: > >> When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of > >> a failure is higher

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hello, On Mon, Feb 27, 2017 at 05:00:31AM -0800, Tahsin Erdogan wrote: > On Mon, Feb 27, 2017 at 1:52 AM, Michal Hocko wrote: > > On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: > >> When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of > >> a failure is higher than GFP_KERNEL

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 01:12:11PM -0800, Tahsin Erdogan wrote: > That makes sense. I will work a patch that does that (unless you are > interested in implementing it yourself). I'd really appreciate if you can work on it. Thanks a lot! -- tejun

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 01:12:11PM -0800, Tahsin Erdogan wrote: > That makes sense. I will work a patch that does that (unless you are > interested in implementing it yourself). I'd really appreciate if you can work on it. Thanks a lot! -- tejun

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

>> Doing preallocations would probably work but not sure if that can be >> done without >> complicating code too much. Could you describe what you have in mind? > > So, blkg_create() already takes @new_blkg argument which is the > preallocated blkg and used during q init. Wouldn't it work to make

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

>> Doing preallocations would probably work but not sure if that can be >> done without >> complicating code too much. Could you describe what you have in mind? > > So, blkg_create() already takes @new_blkg argument which is the > preallocated blkg and used during q init. Wouldn't it work to make

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 12:29 PM, Tejun Heo wrote: > Hello, > > On Mon, Feb 27, 2017 at 12:27:08PM -0800, Tahsin Erdogan wrote: >> A better example is the call path below: >> >> pcpu_alloc+0x68f/0x710 >> __alloc_percpu_gfp+0xd/0x10 >> __percpu_counter_init+0x55/0xc0 >>

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 12:29 PM, Tejun Heo wrote: > Hello, > > On Mon, Feb 27, 2017 at 12:27:08PM -0800, Tahsin Erdogan wrote: >> A better example is the call path below: >> >> pcpu_alloc+0x68f/0x710 >> __alloc_percpu_gfp+0xd/0x10 >> __percpu_counter_init+0x55/0xc0 >> cfq_pd_alloc+0x3b2/0x4e0 >>

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hello, On Mon, Feb 27, 2017 at 12:27:08PM -0800, Tahsin Erdogan wrote: > A better example is the call path below: > > pcpu_alloc+0x68f/0x710 > __alloc_percpu_gfp+0xd/0x10 > __percpu_counter_init+0x55/0xc0 > cfq_pd_alloc+0x3b2/0x4e0 > blkg_alloc+0x187/0x230 > blkg_create+0x489/0x670 >

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hello, On Mon, Feb 27, 2017 at 12:27:08PM -0800, Tahsin Erdogan wrote: > A better example is the call path below: > > pcpu_alloc+0x68f/0x710 > __alloc_percpu_gfp+0xd/0x10 > __percpu_counter_init+0x55/0xc0 > cfq_pd_alloc+0x3b2/0x4e0 > blkg_alloc+0x187/0x230 > blkg_create+0x489/0x670 >

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hi Tejun, On Mon, Feb 27, 2017 at 11:51 AM, Tejun Heo wrote: >> __vmalloc+0x45/0x50 >> pcpu_mem_zalloc+0x50/0x80 >> pcpu_populate_chunk+0x3b/0x380 >> pcpu_alloc+0x588/0x6e0 >> __alloc_percpu_gfp+0xd/0x10 >> __percpu_counter_init+0x55/0xc0 >> blkg_alloc+0x76/0x230 >>

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

Hi Tejun, On Mon, Feb 27, 2017 at 11:51 AM, Tejun Heo wrote: >> __vmalloc+0x45/0x50 >> pcpu_mem_zalloc+0x50/0x80 >> pcpu_populate_chunk+0x3b/0x380 >> pcpu_alloc+0x588/0x6e0 >> __alloc_percpu_gfp+0xd/0x10 >> __percpu_counter_init+0x55/0xc0 >> blkg_alloc+0x76/0x230 >> blkg_create+0x489/0x670 >>

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 11:32:50, Tahsin Erdogan wrote: > >> > > >> > Yes, this prevents adding more pcpu chunks and so cause "atomic" > >> > allocations > >> > to fail more easily. > >> > >> Then I fail to see what is the problem you are trying to fix. > > > > To be more specific. Could you describe

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 11:32:50, Tahsin Erdogan wrote: > >> > > >> > Yes, this prevents adding more pcpu chunks and so cause "atomic" > >> > allocations > >> > to fail more easily. > >> > >> Then I fail to see what is the problem you are trying to fix. > > > > To be more specific. Could you describe

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

>> > >> > Yes, this prevents adding more pcpu chunks and so cause "atomic" >> > allocations >> > to fail more easily. >> >> Then I fail to see what is the problem you are trying to fix. > > To be more specific. Could you describe what more can we do in the > vmalloc layer for GFP_NOWAIT

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

>> > >> > Yes, this prevents adding more pcpu chunks and so cause "atomic" >> > allocations >> > to fail more easily. >> >> Then I fail to see what is the problem you are trying to fix. > > To be more specific. Could you describe what more can we do in the > vmalloc layer for GFP_NOWAIT

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 18:07:53, Michal Hocko wrote: > On Mon 27-02-17 09:01:09, Tahsin Erdogan wrote: > > On Mon, Feb 27, 2017 at 7:25 AM, Michal Hocko wrote: > > > /* > > > * No space left. Create a new chunk. We don't want multiple > > > * tasks to

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 18:07:53, Michal Hocko wrote: > On Mon 27-02-17 09:01:09, Tahsin Erdogan wrote: > > On Mon, Feb 27, 2017 at 7:25 AM, Michal Hocko wrote: > > > /* > > > * No space left. Create a new chunk. We don't want multiple > > > * tasks to create chunks

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 09:01:09, Tahsin Erdogan wrote: > On Mon, Feb 27, 2017 at 7:25 AM, Michal Hocko wrote: > > /* > > * No space left. Create a new chunk. We don't want multiple > > * tasks to create chunks simultaneously. Serialize and create iff > >

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 09:01:09, Tahsin Erdogan wrote: > On Mon, Feb 27, 2017 at 7:25 AM, Michal Hocko wrote: > > /* > > * No space left. Create a new chunk. We don't want multiple > > * tasks to create chunks simultaneously. Serialize and create iff > > * there's

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 7:25 AM, Michal Hocko wrote: > /* > * No space left. Create a new chunk. We don't want multiple > * tasks to create chunks simultaneously. Serialize and create iff > * there's still no empty chunk after grabbing the

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 7:25 AM, Michal Hocko wrote: > /* > * No space left. Create a new chunk. We don't want multiple > * tasks to create chunks simultaneously. Serialize and create iff > * there's still no empty chunk after grabbing the mutex. >

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 05:00:31, Tahsin Erdogan wrote: > On Mon, Feb 27, 2017 at 1:52 AM, Michal Hocko wrote: > > On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: > >> When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of > >> a failure is higher than GFP_KERNEL case.

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon 27-02-17 05:00:31, Tahsin Erdogan wrote: > On Mon, Feb 27, 2017 at 1:52 AM, Michal Hocko wrote: > > On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: > >> When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of > >> a failure is higher than GFP_KERNEL case. This is mainly

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 1:52 AM, Michal Hocko wrote: > On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: >> When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of >> a failure is higher than GFP_KERNEL case. This is mainly because >> pcpu_alloc() relies on

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Mon, Feb 27, 2017 at 1:52 AM, Michal Hocko wrote: > On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: >> When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of >> a failure is higher than GFP_KERNEL case. This is mainly because >> pcpu_alloc() relies on previously allocated

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: > When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of > a failure is higher than GFP_KERNEL case. This is mainly because > pcpu_alloc() relies on previously allocated reserves and does not make > an effort to add memory to its pools

Re: [PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

On Sat 25-02-17 20:38:29, Tahsin Erdogan wrote: > When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of > a failure is higher than GFP_KERNEL case. This is mainly because > pcpu_alloc() relies on previously allocated reserves and does not make > an effort to add memory to its pools

[PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of a failure is higher than GFP_KERNEL case. This is mainly because pcpu_alloc() relies on previously allocated reserves and does not make an effort to add memory to its pools for non-GFP_KERNEL case. This issue is somewhat

[PATCH v2 3/3] percpu: improve allocation success rate for non-GFP_KERNEL callers

When pcpu_alloc() is called with gfp != GFP_KERNEL, the likelihood of a failure is higher than GFP_KERNEL case. This is mainly because pcpu_alloc() relies on previously allocated reserves and does not make an effort to add memory to its pools for non-GFP_KERNEL case. This issue is somewhat