From: Vasily Averin <vvs@openvz.org> To: Steven Rostedt <rostedt@goodmis.org> Cc: YoPOhRctb8wwbmY5@carbon, Shakeel Butt <shakeelb@google.com>, Roman Gushchin <roman.gushchin@linux.dev>, Vlastimil Babka <vbabka@suse.cz>, Matthew Wilcox <willy@infradead.org>, Hyeonggon Yoo <42.hyeyoo@gmail.com>, Muchun Song <songmuchun@bytedance.com>, kernel@openvz.org, linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>, Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org, Joonsoo Kim <iamjoonsoo.kim@lge.com>, David Rientjes <rientjes@google.com>, Pekka Enberg <penberg@kernel.org>, Christoph Lameter <cl@linux.com>, Michal Hocko <mhocko@suse.com> Subject: Re: [PATCH v3] tracing: add 'accounted' entry into output of allocation tracepoints Date: Thu, 19 May 2022 19:29:36 +0300 [thread overview] Message-ID: <e018be81-f4f2-a26f-7c5a-7adddd9c56c4@openvz.org> (raw) In-Reply-To: <20220519100348.101d027d@gandalf.local.home> On 5/19/22 17:03, Steven Rostedt wrote: > On Thu, 19 May 2022 14:35:46 +0300 > Vasily Averin <vvs@openvz.org> wrote: > >>>> @@ -33,42 +35,46 @@ DECLARE_EVENT_CLASS(kmem_alloc, >>>> __entry->bytes_req = bytes_req; >>>> __entry->bytes_alloc = bytes_alloc; >>>> __entry->gfp_flags = (__force unsigned long)gfp_flags; >>>> + __entry->accounted = (gfp_flags & __GFP_ACCOUNT) || >>>> + (s && s->flags & SLAB_ACCOUNT); >>> >>> Now you could make this even faster in the fast path and save just the >>> s->flags. >>> >>> __entry->sflags = s ? s->flags : 0; >>> >>>> ), >>>> >>>> - TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s", >>>> + TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s accounted=%s", >>>> (void *)__entry->call_site, >>>> __entry->ptr, >>>> __entry->bytes_req, >>>> __entry->bytes_alloc, >>>> - show_gfp_flags(__entry->gfp_flags)) >>>> + show_gfp_flags(__entry->gfp_flags), >>>> + __entry->accounted ? "true" : "false") >>> >>> And then have: "accounted=%s": >>> >>> (__entry->gfp_flags & __GFP_ACCOUNT) || >>> (__entry->sflags & SLAB_ACCOUNT) ? "true" : "false" >> >> Unfortunately this returns back sparse warnings about bitwise gfp_t and slab_flags_t casts. >> Could you please explain why your variant is faster? > > Micro-optimization, grant you, but it is faster because it moves some of > the logic into the slow path (the read side), and takes it out of the fast > path (the write side). > > The idea of tracing is to squeeze out every cycle we can to keep the > tracing overhead down. > > But it's really up to you if you need that. I'm not going to let this be a > blocker. This is more of an FYI than anything else. Frankly speaking I vote for performance with both hands. However I'm still would like to avoid new sparse warnings. Christoph Hellwig just recently taught me, "never add '__force' before thinking hard about them", but in this case I would need to use it three times. I found that bitwise typecasts can be avoided by using translation unions. What do you think about following trick? diff --git a/mm/slab.h b/mm/slab.h index 95eb34174c1b..f676612ca40f 100644 --- a/mm/slab.h +++ b/mm/slab.h @@ -882,4 +882,14 @@ void __check_heap_object(const void *ptr, unsigned long n, } #endif +union gfp_flags_u { + unsigned long ulong; + gfp_t flags; +}; + +union slab_flags_u { + unsigned int uint; + slab_flags_t sflags; +}; + #endif /* MM_SLAB_H */ diff --git a/include/trace/events/kmem.h b/include/trace/events/kmem.h index 71c141804222..91632a61e16d 100644 --- a/include/trace/events/kmem.h +++ b/include/trace/events/kmem.h @@ -13,18 +13,20 @@ DECLARE_EVENT_CLASS(kmem_alloc, TP_PROTO(unsigned long call_site, const void *ptr, + struct kmem_cache *s, size_t bytes_req, size_t bytes_alloc, gfp_t gfp_flags), - TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags), + TP_ARGS(call_site, ptr, s, bytes_req, bytes_alloc, gfp_flags), TP_STRUCT__entry( __field( unsigned long, call_site ) __field( const void *, ptr ) __field( size_t, bytes_req ) __field( size_t, bytes_alloc ) - __field( unsigned long, gfp_flags ) + __field_struct( union gfp_flags_u, gfp ) + __field_struct( union slab_flags_u, s ) ), TP_fast_assign( @@ -32,51 +34,57 @@ DECLARE_EVENT_CLASS(kmem_alloc, __entry->ptr = ptr; __entry->bytes_req = bytes_req; __entry->bytes_alloc = bytes_alloc; - __entry->gfp_flags = (__force unsigned long)gfp_flags; + __entry->gfp.flags = gfp_flags; + __entry->s.sflags = s ? s->flags : 0; ), - TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s", + TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s accounted=%s", (void *)__entry->call_site, __entry->ptr, __entry->bytes_req, __entry->bytes_alloc, - show_gfp_flags(__entry->gfp_flags)) + show_gfp_flags(__entry->gfp.ulong), + ((__entry->gfp.flags & __GFP_ACCOUNT) || + (__entry->s.sflags & SLAB_ACCOUNT)) ? "true" : "false") ); Thank you, Vasily Averin
WARNING: multiple messages have this Message-ID (diff)
From: Vasily Averin <vvs@openvz.org> To: Steven Rostedt <rostedt@goodmis.org> Cc: YoPOhRctb8wwbmY5@carbon.kvack.org, Shakeel Butt <shakeelb@google.com>, Roman Gushchin <roman.gushchin@linux.dev>, Vlastimil Babka <vbabka@suse.cz>, Matthew Wilcox <willy@infradead.org>, Hyeonggon Yoo <42.hyeyoo@gmail.com>, Muchun Song <songmuchun@bytedance.com>, kernel@openvz.org, linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>, Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org, Joonsoo Kim <iamjoonsoo.kim@lge.com>, David Rientjes <rientjes@google.com>, Pekka Enberg <penberg@kernel.org>, Christoph Lameter <cl@linux.com>, Michal Hocko <mhocko@suse.com> Subject: Re: [PATCH v3] tracing: add 'accounted' entry into output of allocation tracepoints Date: Thu, 19 May 2022 19:29:36 +0300 [thread overview] Message-ID: <e018be81-f4f2-a26f-7c5a-7adddd9c56c4@openvz.org> (raw) In-Reply-To: <20220519100348.101d027d@gandalf.local.home> On 5/19/22 17:03, Steven Rostedt wrote: > On Thu, 19 May 2022 14:35:46 +0300 > Vasily Averin <vvs@openvz.org> wrote: > >>>> @@ -33,42 +35,46 @@ DECLARE_EVENT_CLASS(kmem_alloc, >>>> __entry->bytes_req = bytes_req; >>>> __entry->bytes_alloc = bytes_alloc; >>>> __entry->gfp_flags = (__force unsigned long)gfp_flags; >>>> + __entry->accounted = (gfp_flags & __GFP_ACCOUNT) || >>>> + (s && s->flags & SLAB_ACCOUNT); >>> >>> Now you could make this even faster in the fast path and save just the >>> s->flags. >>> >>> __entry->sflags = s ? s->flags : 0; >>> >>>> ), >>>> >>>> - TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s", >>>> + TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s accounted=%s", >>>> (void *)__entry->call_site, >>>> __entry->ptr, >>>> __entry->bytes_req, >>>> __entry->bytes_alloc, >>>> - show_gfp_flags(__entry->gfp_flags)) >>>> + show_gfp_flags(__entry->gfp_flags), >>>> + __entry->accounted ? "true" : "false") >>> >>> And then have: "accounted=%s": >>> >>> (__entry->gfp_flags & __GFP_ACCOUNT) || >>> (__entry->sflags & SLAB_ACCOUNT) ? "true" : "false" >> >> Unfortunately this returns back sparse warnings about bitwise gfp_t and slab_flags_t casts. >> Could you please explain why your variant is faster? > > Micro-optimization, grant you, but it is faster because it moves some of > the logic into the slow path (the read side), and takes it out of the fast > path (the write side). > > The idea of tracing is to squeeze out every cycle we can to keep the > tracing overhead down. > > But it's really up to you if you need that. I'm not going to let this be a > blocker. This is more of an FYI than anything else. Frankly speaking I vote for performance with both hands. However I'm still would like to avoid new sparse warnings. Christoph Hellwig just recently taught me, "never add '__force' before thinking hard about them", but in this case I would need to use it three times. I found that bitwise typecasts can be avoided by using translation unions. What do you think about following trick? diff --git a/mm/slab.h b/mm/slab.h index 95eb34174c1b..f676612ca40f 100644 --- a/mm/slab.h +++ b/mm/slab.h @@ -882,4 +882,14 @@ void __check_heap_object(const void *ptr, unsigned long n, } #endif +union gfp_flags_u { + unsigned long ulong; + gfp_t flags; +}; + +union slab_flags_u { + unsigned int uint; + slab_flags_t sflags; +}; + #endif /* MM_SLAB_H */ diff --git a/include/trace/events/kmem.h b/include/trace/events/kmem.h index 71c141804222..91632a61e16d 100644 --- a/include/trace/events/kmem.h +++ b/include/trace/events/kmem.h @@ -13,18 +13,20 @@ DECLARE_EVENT_CLASS(kmem_alloc, TP_PROTO(unsigned long call_site, const void *ptr, + struct kmem_cache *s, size_t bytes_req, size_t bytes_alloc, gfp_t gfp_flags), - TP_ARGS(call_site, ptr, bytes_req, bytes_alloc, gfp_flags), + TP_ARGS(call_site, ptr, s, bytes_req, bytes_alloc, gfp_flags), TP_STRUCT__entry( __field( unsigned long, call_site ) __field( const void *, ptr ) __field( size_t, bytes_req ) __field( size_t, bytes_alloc ) - __field( unsigned long, gfp_flags ) + __field_struct( union gfp_flags_u, gfp ) + __field_struct( union slab_flags_u, s ) ), TP_fast_assign( @@ -32,51 +34,57 @@ DECLARE_EVENT_CLASS(kmem_alloc, __entry->ptr = ptr; __entry->bytes_req = bytes_req; __entry->bytes_alloc = bytes_alloc; - __entry->gfp_flags = (__force unsigned long)gfp_flags; + __entry->gfp.flags = gfp_flags; + __entry->s.sflags = s ? s->flags : 0; ), - TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s", + TP_printk("call_site=%pS ptr=%p bytes_req=%zu bytes_alloc=%zu gfp_flags=%s accounted=%s", (void *)__entry->call_site, __entry->ptr, __entry->bytes_req, __entry->bytes_alloc, - show_gfp_flags(__entry->gfp_flags)) + show_gfp_flags(__entry->gfp.ulong), + ((__entry->gfp.flags & __GFP_ACCOUNT) || + (__entry->s.sflags & SLAB_ACCOUNT)) ? "true" : "false") ); Thank you, Vasily Averin
next prev parent reply other threads:[~2022-05-19 16:29 UTC|newest] Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top 2022-05-18 6:24 [PATCH v3] tracing: add 'accounted' entry into output of allocation tracepoints Vasily Averin 2022-05-18 15:09 ` Shakeel Butt 2022-05-18 15:45 ` Vasily Averin 2022-05-18 20:04 ` Steven Rostedt 2022-05-18 20:04 ` Steven Rostedt 2022-05-19 11:35 ` Vasily Averin 2022-05-19 11:35 ` Vasily Averin 2022-05-19 14:03 ` Steven Rostedt 2022-05-19 14:03 ` Steven Rostedt 2022-05-19 16:29 ` Vasily Averin [this message] 2022-05-19 16:29 ` Vasily Averin 2022-05-19 16:32 ` Steven Rostedt 2022-05-19 16:32 ` Steven Rostedt 2022-05-21 18:32 ` Vasily Averin 2022-05-21 18:32 ` Vasily Averin 2022-05-21 18:36 ` [PATCH v4] " Vasily Averin 2022-05-22 3:51 ` Hyeonggon Yoo 2022-05-22 4:33 ` Vasily Averin 2022-05-22 5:19 ` Hyeonggon Yoo 2022-05-22 5:42 ` Shakeel Butt 2022-05-22 18:53 ` Vasily Averin 2022-05-22 20:09 ` Steven Rostedt 2022-05-23 4:03 ` Vasily Averin 2022-05-23 13:12 ` Vlastimil Babka 2022-05-30 7:47 ` [PATCH v5] " Vasily Averin 2022-05-30 8:25 ` Muchun Song 2022-05-31 11:46 ` Hyeonggon Yoo 2022-05-31 16:58 ` Vasily Averin 2022-06-03 3:21 ` [PATCH mm v6] mm/tracing: " Vasily Averin 2022-06-03 3:21 ` Vasily Averin 2022-06-15 9:41 ` Vlastimil Babka 2022-06-15 9:41 ` Vlastimil Babka 2022-05-25 1:34 ` [PATCH v4] tracing: " Roman Gushchin 2022-05-25 7:33 ` Hyeonggon Yoo 2022-05-25 8:24 ` Vasily Averin
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=e018be81-f4f2-a26f-7c5a-7adddd9c56c4@openvz.org \ --to=vvs@openvz.org \ --cc=42.hyeyoo@gmail.com \ --cc=YoPOhRctb8wwbmY5@carbon \ --cc=akpm@linux-foundation.org \ --cc=cl@linux.com \ --cc=iamjoonsoo.kim@lge.com \ --cc=kernel@openvz.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-mm@kvack.org \ --cc=mhocko@suse.com \ --cc=mingo@redhat.com \ --cc=penberg@kernel.org \ --cc=rientjes@google.com \ --cc=roman.gushchin@linux.dev \ --cc=rostedt@goodmis.org \ --cc=shakeelb@google.com \ --cc=songmuchun@bytedance.com \ --cc=vbabka@suse.cz \ --cc=willy@infradead.org \ /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: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.