linux-kernel - Re: [External] Re: [PATCH v3 08/21] mm/vmemmap: Initialize page table lock for vmemmap

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:   Tue, 10 Nov 2020 13:17:11 +0800
From:   Muchun Song <songmuchun@...edance.com>
To:     Oscar Salvador <osalvador@...e.de>
Cc:     Jonathan Corbet <corbet@....net>,
        Mike Kravetz <mike.kravetz@...cle.com>,
        Thomas Gleixner <tglx@...utronix.de>, mingo@...hat.com,
        bp@...en8.de, x86@...nel.org, hpa@...or.com,
        dave.hansen@...ux.intel.com, luto@...nel.org,
        Peter Zijlstra <peterz@...radead.org>, viro@...iv.linux.org.uk,
        Andrew Morton <akpm@...ux-foundation.org>, paulmck@...nel.org,
        mchehab+huawei@...nel.org, pawan.kumar.gupta@...ux.intel.com,
        Randy Dunlap <rdunlap@...radead.org>, oneukum@...e.com,
        anshuman.khandual@....com, jroedel@...e.de,
        Mina Almasry <almasrymina@...gle.com>,
        David Rientjes <rientjes@...gle.com>,
        Matthew Wilcox <willy@...radead.org>,
        Michal Hocko <mhocko@...e.com>,
        Xiongchun duan <duanxiongchun@...edance.com>,
        linux-doc@...r.kernel.org, LKML <linux-kernel@...r.kernel.org>,
        Linux Memory Management List <linux-mm@...ck.org>,
        linux-fsdevel <linux-fsdevel@...r.kernel.org>
Subject: Re: [External] Re: [PATCH v3 08/21] mm/vmemmap: Initialize page table
 lock for vmemmap

On Tue, Nov 10, 2020 at 2:11 AM Oscar Salvador <osalvador@...e.de> wrote:
>
> On Sun, Nov 08, 2020 at 10:11:00PM +0800, Muchun Song wrote:
> > In the register_page_bootmem_memmap, the slab allocator is not ready
> > yet. So when ALLOC_SPLIT_PTLOCKS, we use init_mm.page_table_lock.
> > otherwise we use per page table lock(page->ptl). In the later patch,
> > we will use the vmemmap page table lock to guard the splitting of
> > the vmemmap huge PMD.
>
> I am not sure about this one.
> Grabbing init_mm's pagetable lock for specific hugetlb operations does not
> seem like a good idea, and we do not know how contented is that one.

These APIs are used to guard the operations on vmemmap page tables.
For now, it is only for specific hugetlb operations. But maybe in the future,
someone also wants to modify the vmemmap page tables, he also can
use these APIs. Yeah, we do not know how contented is init_mm's pagetable
lock. Grabbing this one may not be a good idea.

>
> I think a better fit would be to find another hook to initialize
> page_table_lock at a later stage.
> Anyway, we do not need till we are going to perform an operation
> on the range, right?

Yeah. You are right.

>
> Unless I am missing something, this should be doable in hugetlb_init.
>
> hugetlb_init is part from a init_call that gets called during do_initcalls.
> At this time, slab is fully operative.

If we initialize the page_table_lock in the hugetlb_init, we need to
walk the vmemmap page tables again. But the vmemmap pages
size is small, maybe the overhead of this is also small. And doing
this in hugetlb_init can make the code cleaner. Thanks very much.


>
> start_kernel
>  kmem_cache_init_late
>  kmem_cache_init_late
>  ...
>  arch_call_rest_init
>   rest_init
>    kernel_init_freeable
>     do_basic_setup
>      do_initcalls
>       hugetlb_init
>
> --
> Oscar Salvador
> SUSE L3



--
Yours,
Muchun