dri-devel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Tao Cui <cui.tao@linux.dev>
To: Steven Rostedt <rostedt@goodmis.org>,
	Vincent Donnefort <vdonnefort@google.com>
Cc: cui.tao@linux.dev, akpm@linux-foundation.org,
	dgilbert@interlog.com, mkp@kernel.org, loberman@redhat.com,
	jmeneghi@redhat.com, kraxel@redhat.com,
	vivek.kasireddy@intel.com, mst@redhat.com, jasowangio@gmail.com,
	eperezma@redhat.com, stefanha@redhat.com, sgarzare@redhat.com,
	mhiramat@kernel.org, jgg@ziepe.ca, kevin.tian@intel.com,
	axboe@kernel.dk, linux-mm@kvack.org,
	James.Bottomley@hansenpartnership.com, sumit.semwal@linaro.org,
	christian.koenig@amd.com, phasta@kernel.org, joro@8bytes.org,
	will@kernel.org, linux-scsi@vger.kernel.org,
	dri-devel@lists.freedesktop.org, linux-media@vger.kernel.org,
	linaro-mm-sig@lists.linaro.org, kvm@vger.kernel.org,
	virtualization@lists.linux.dev,
	linux-trace-kernel@vger.kernel.org, iommu@lists.linux.dev,
	io-uring@vger.kernel.org, linux-kernel@vger.kernel.org,
	bobbyeshleman@meta.com, mathieu.desnoyers@efficios.com,
	robin.murphy@arm.com, mason@kernel.org, aesteve@redhat.com,
	tjmercier@google.com, harry@kernel.org, vbabka@suse.cz,
	cuitao@kylinos.cn
Subject: Re: [PATCH 4/7] tracing: account ring buffer pages and trace_array metadata to memcg
Date: Fri, 20 Nov 2026 07:47:56 +0800	[thread overview]
Message-ID: <583bde38-c84d-41a3-a65c-d1365ea04f07@linux.dev> (raw)
In-Reply-To: <20261111181447.3c483a50@fedora>

Hello, Steven, Vincent,

在 2026/11/12 07:14, Steven Rostedt 写道:
> On Wed, 11 Nov 2026 16:47:38 +0000
> Vincent Donnefort <vdonnefort@google.com> wrote:
> 
>> On Fri, Nov 06, 2026 at 09:18:26PM +0800, Tao Cui wrote:
>>> From: Tao Cui <cuitao@kylinos.cn>
>>>
>>> Creating an instance immediately allocates two full sets of per-cpu
>>> ring buffers, array buffer plus snapshot, about 1.375MB per cpu per set
>>> on the default config, plus the trace_array and its name.  None of it
>>> is charged: neither trace.c nor ring_buffer.c passes __GFP_ACCOUNT,
>>> while user_events in the same directory accounts everything it
>>> allocates.
> 
> As the name suggests, "user_events" are for users and are special.
> 
>>>
>>> Instance creation is currently gated by tracefs file permissions
>>> alone, so a
>>> rootful container with the host tracefs bound in can create instances
>>> until the host runs out of memory.
> 
> You mean, "works as expected"?
> 
>>  Charge the ring buffer pages and
>>> the trace_array metadata to the creating cgroup.
>>>
>>> Signed-off-by: Tao Cui <cuitao@kylinos.cn>
>>> ---
>>>  kernel/trace/ring_buffer.c | 8 +++++++-
>>>  kernel/trace/trace.c       | 6 +++---
>>>  2 files changed, 10 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c
>>> index 1fd28a72e6ee..676c059c112c 100644
>>> --- a/kernel/trace/ring_buffer.c
>>> +++ b/kernel/trace/ring_buffer.c
>>> @@ -441,7 +441,13 @@ static struct buffer_data_page *alloc_cpu_data(int cpu, int order)
>>>  	 * gracefully without invoking oom-killer and the system is not
>>>  	 * destabilized.
>>>  	 */
>>> -	mflags = GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_COMP | __GFP_ZERO;
>>> +	/*
>>> +	 * The subbuffers are charged to the memcg of whoever caused the
>>> +	 * allocation: the creator of a trace instance, or root at boot for
>>> +	 * the global trace.
>>> +	 */
>>> +	mflags = GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_COMP | __GFP_ZERO |
>>> +		  __GFP_ACCOUNT;
>>>  
>>>  	page = alloc_pages_node(cpu_to_node(cpu), mflags, order);
>>>  	if (!page)  
>>
>> Hum, by default tracefs is root only. 
>>
>> And even if made accessible to other users: 
>>
>>   1. The life span of a trace instance is not tied to the caller.
>>   2. The ring buffer is shared by multiple users.
> 
> Right. Is there a use case out there that allows non-admin to create an
> instance? Tracing ring buffers can expose a lot and should not be given
> to unprivileged users.
> 
Not that I can point to.  Instances are admin-only by default, and the permission layer is the right place for that.

> What exactly is this trying to achieve?
>
The series closes uncharged kernel allocations that scale with user input.  For tracing I misread the tracefs permission check as a gap rather than the intended boundary.  Combined with Vincent's points on lifetime and sharing, I'll drop patch.
 
Thanks,
Tao
> -- Steve
> 
> 
>>
>> Overall I am not convinced charging the creator makes much sense.
>>
>> Eventually a few allocations for iterators could be GFP_ACCOUNT.
>>
> 


  reply	other threads:[~2026-11-19 23:48 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-06 13:18 [PATCH 0/7] memcg: account the remaining uncharged kernel allocations Tao Cui
2026-11-06 13:18 ` [PATCH 1/7] scsi: sg and st: account the command buffer memory to memcg Tao Cui
2026-11-07 13:27   ` sashiko-bot
2026-11-09  2:25     ` Tao Cui
2026-11-06 13:18 ` [PATCH 2/7] dma-buf/udmabuf: account the page array metadata " Tao Cui
2026-11-06 13:23   ` Christian König
2026-11-06 17:25     ` T.J. Mercier
2026-11-09  2:32       ` Tao Cui
2026-11-09 10:56       ` Christian König
2026-11-09 13:44         ` Albert Esteve
2026-11-06 13:18 ` [PATCH 3/7] vhost: account the per-device kernel allocations " Tao Cui
2026-11-19 15:12   ` Michael S. Tsirkin
2026-11-19 23:53     ` Tao Cui
2026-11-06 13:18 ` [PATCH 4/7] tracing: account ring buffer pages and trace_array metadata " Tao Cui
2026-11-07 13:27   ` sashiko-bot
2026-11-09  2:35     ` Tao Cui
2026-11-11 16:47   ` Vincent Donnefort
2026-11-11 23:14     ` Steven Rostedt
2026-11-19 23:47       ` Tao Cui [this message]
2026-11-06 13:18 ` [PATCH 5/7] iommufd: account the hw_queue page array " Tao Cui
2026-11-16 19:52   ` Jason Gunthorpe
2026-11-16 20:33     ` Jason Gunthorpe
2026-11-06 13:18 ` [PATCH 6/7] io_uring: account the ring context and SQPOLL thread data " Tao Cui
2026-11-07 13:27   ` sashiko-bot
2026-11-09  2:37     ` Tao Cui
2026-11-06 13:18 ` [PATCH 7/7] io_uring: account the remaining per-object allocations " Tao Cui

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=583bde38-c84d-41a3-a65c-d1365ea04f07@linux.dev \
    --to=cui.tao@linux.dev \
    --cc=James.Bottomley@hansenpartnership.com \
    --cc=aesteve@redhat.com \
    --cc=akpm@linux-foundation.org \
    --cc=axboe@kernel.dk \
    --cc=bobbyeshleman@meta.com \
    --cc=christian.koenig@amd.com \
    --cc=cuitao@kylinos.cn \
    --cc=dgilbert@interlog.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=eperezma@redhat.com \
    --cc=harry@kernel.org \
    --cc=io-uring@vger.kernel.org \
    --cc=iommu@lists.linux.dev \
    --cc=jasowangio@gmail.com \
    --cc=jgg@ziepe.ca \
    --cc=jmeneghi@redhat.com \
    --cc=joro@8bytes.org \
    --cc=kevin.tian@intel.com \
    --cc=kraxel@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=linaro-mm-sig@lists.linaro.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=linux-trace-kernel@vger.kernel.org \
    --cc=loberman@redhat.com \
    --cc=mason@kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mhiramat@kernel.org \
    --cc=mkp@kernel.org \
    --cc=mst@redhat.com \
    --cc=phasta@kernel.org \
    --cc=robin.murphy@arm.com \
    --cc=rostedt@goodmis.org \
    --cc=sgarzare@redhat.com \
    --cc=stefanha@redhat.com \
    --cc=sumit.semwal@linaro.org \
    --cc=tjmercier@google.com \
    --cc=vbabka@suse.cz \
    --cc=vdonnefort@google.com \
    --cc=virtualization@lists.linux.dev \
    --cc=vivek.kasireddy@intel.com \
    --cc=will@kernel.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: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox