linux-kernel - Re: [PATCH v5 4/4] mm/sparse: Optimize memmap allocation during sparse

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]

Message-ID: <b8d5b9cb-ca09-4bcc-0a31-3db1232fe787@oracle.com>
Date:   Fri, 29 Jun 2018 14:01:14 -0400
From:   Pavel Tatashin <pasha.tatashin@...cle.com>
To:     Dave Hansen <dave.hansen@...el.com>
Cc:     bhe@...hat.com, LKML <linux-kernel@...r.kernel.org>,
        Andrew Morton <akpm@...ux-foundation.org>, pagupta@...hat.com,
        Linux Memory Management List <linux-mm@...ck.org>,
        kirill.shutemov@...ux.intel.com
Subject: Re: [PATCH v5 4/4] mm/sparse: Optimize memmap allocation during
 sparse_init()

On 06/29/2018 01:52 PM, Dave Hansen wrote:
> On 06/29/2018 10:48 AM, Pavel Tatashin wrote:
>> Here is example:
>> Node1:
>> map_map[0] -> Struct pages ...
>> map_map[1] -> NULL
>> Node2:
>> map_map[2] -> Struct pages ...
>>
>> We always want to configure section from Node2 with struct pages from
>> Node2. Even, if there are holes in-between. The same with usemap.
> 
> Right...  But your example consumes two mem_map[]s.
> 
> But, from scanning the code, we increment nr_consumed_maps three times.
> Correct?

Correct: it should be incremented on every iteration of the loop. No matter if the entries contained valid data or NULLs. So we increment in three places:

if map_map[] has invalid entry, increment, continue
if usemap_map[] has invalid entry, increment, continue
at the end of the loop, everything was valid we increment it

This is done so nr_consumed_maps does not get out of sync with the current pnum. pnum does not equal to nr_consumed_maps, as there are may be holes in pnums, but there is one-to-one correlation.

Pavel