lkml.org 
[lkml]   [2020]   [Apr]   [17]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe:Re: [PATCH] kmalloc index optimization(add kmalloc max size check)
From
Date


From: Christopher Lameter <cl@linux.com>
Date: 2020-04-17 23:59:00
To: Bernard Zhao <bernard@vivo.com>
Cc: Pekka Enberg <penberg@kernel.org>,David Rientjes <rientjes@google.com>,Joonsoo Kim <iamjoonsoo.kim@lge.com>,Andrew Morton <akpm@linux-foundation.org>,linux-mm@kvack.org,linux-kernel@vger.kernel.org,kernel@vivo.com
Subject: Re: [PATCH] kmalloc_index optimization(add kmalloc max size check)>On Fri, 17 Apr 2020, Bernard Zhao wrote:
>
>> kmalloc size should never exceed KMALLOC_MAX_SIZE.
>> kmalloc_index realise if size is exceed KMALLOC_MAX_SIZE, e.g 64M,
>> kmalloc_index just return index 26, but never check with OS`s max
>> kmalloc config KMALLOC_MAX_SIZE. This index`s kmalloc caches maybe
>> not create in function create_kmalloc_caches.
>> We can throw an warninginfo in kmalloc at the beginning, instead of
>> being guaranteed by the buddy alloc behind.
>
>kmalloc_index(0 already bugs if the allocation is more than 64M
>
>
>...
>
> if (size <= 64 * 1024 * 1024) return 26;
> BUG();
>
>
>You could modify that to check for KMALLOC_MAX_SIZE with some more
>conditionals but then kmalloc_index) is written so that the compiler gets
>constant folding right.
>
>If you have a patch like that then please verify that all c compilers in
>use perform correct constant folding and do not add unnecessary code.
>
>

Sorry for the misunderstanding.
What I meant was that the “if”in kmalloc_index should be consistent
with KMALLOC_MAX_SIZE. For example, if the MAX_ZONEORDER configured
by the kernel is 11, which is 4M, then size >4M should trigger BUG(). If the
configuration is smaller, e.g MAX_ZONEORDER is 9, 1M, then the size > 1M
should be BUG.
But the current code is not, kmalloc_index will only be BUG() when it exceeds 64M.




\
 
 \ /
  Last update: 2020-04-18 03:39    [W:0.129 / U:1.268 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site