linux-kernel - Re: [PATCH 2/3] tracing/stack-tracer: fix locking

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sun, 23 Nov 2008 12:07:14 +0100
From:	Ingo Molnar <mingo@...e.hu>
To:	Török Edwin <edwintorok@...il.com>
Cc:	srostedt@...hat.com, a.p.zijlstra@...llo.nl, sandmann@...mi.au.dk,
	linux-kernel@...r.kernel.org, viro@...IV.linux.org.uk
Subject: Re: [PATCH 2/3] tracing/stack-tracer: fix locking


* Török Edwin <edwintorok@...il.com> wrote:

> On 2008-11-23 13:01, Ingo Molnar wrote:
> > * Török Edwin <edwintorok@...il.com> wrote:
> >
> >   
> >> On 2008-11-23 12:52, Ingo Molnar wrote:
> >>     
> >>> * Török Edwin <edwintorok@...il.com> wrote:
> >>>
> >>>   
> >>>       
> >>>> +		down_read(&mm->mmap_sem);
> >>>> +		vma = find_vma(mm, ip);
> >>>>  		if (vma) {
> >>>>  			file = vma->vm_file;
> >>>>  			vmstart = vma->vm_start;
> >>>>  		}
> >>>> +		up_read(&mm->mmap_sem);
> >>>>  	}
> >>>>  	if (file) {
> >>>>  		ret = trace_seq_path(s, &file->f_path);
> >>>>     
> >>>>         
> >>> and now it's "file" that is held without a reference and possibly 
> >>> racy.
> >>>   
> >>>       
> >> I'll move the (file) inside the mmap_sem too, since file is part of 
> >> vma, and you need mmap_sem to access the vma nobody can modify 
> >> (file) if they don't hold mmap_sem. Am I right?
> >>     
> >
> > yeah - you can use the mm lock for that as the vma holds a reference 
> > to the file. This works as long as none of the seq functions called 
> > with this lock held can provoke a user pagefault (they shouldnt, but i 
> > havent checked in detail).
> >   
> 
> It calls d_path, that access only kernel structures AFAICT, so it 
> can't provoke a *user* pagefault. Not sure about kernel pagefaults, 
> but they shouldn't matter.

yeah - kernel pagefaults can happen in rare circumstances here but 
they are all atomic and transparent to most other parts of the kernel 
so they indeed do not matter.

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/