Re: + mm-slab-shorten-kmalloc-cache-names-for-large-sizes.patch added to -mm tree

[Randy Dunlap]

On 09/08/2018 08:54 AM, Matthew Wilcox wrote:
> On Sat, Sep 08, 2018 at 10:03:06AM -0400, Johannes Weiner wrote:
>> On Fri, Sep 07, 2018 at 10:58:16PM +0300, Alexey Dobriyan wrote:
>>> On Thu, Sep 06, 2018 at 03:48:49PM -0700, akpm@xxxxxxxxxxxxxxxxxxxx wrote:
>>>> -	{"kmalloc-1024",         1024},		{"kmalloc-2048",         2048},
>>>> -	{"kmalloc-4096",         4096},		{"kmalloc-8192",         8192},
>>>> -	{"kmalloc-16384",       16384},		{"kmalloc-32768",       32768},
>>>> -	{"kmalloc-65536",       65536},		{"kmalloc-131072",     131072},
>>>> -	{"kmalloc-262144",     262144},		{"kmalloc-524288",     524288},
>>>> -	{"kmalloc-1048576",   1048576},		{"kmalloc-2097152",   2097152},
>>>> -	{"kmalloc-4194304",   4194304},		{"kmalloc-8388608",   8388608},
>>>> -	{"kmalloc-16777216", 16777216},		{"kmalloc-33554432", 33554432},
>>>> -	{"kmalloc-67108864", 67108864}
>>>> +	{"kmalloc-1k",           1024},		{"kmalloc-2k",           2048},
>>>> +	{"kmalloc-4k",           4096},		{"kmalloc-8k",           8192},
>>>> +	{"kmalloc-16k",         16384},		{"kmalloc-32k",         32768},
>>>> +	{"kmalloc-64k",         65536},		{"kmalloc-128k",       131072},
>>>> +	{"kmalloc-256k",       262144},		{"kmalloc-512k",       524288},
>>>> +	{"kmalloc-1M",        1048576},		{"kmalloc-2M",        2097152},
>>>> +	{"kmalloc-4M",        4194304},		{"kmalloc-8M",        8388608},
>>>> +	{"kmalloc-16M",      16777216},		{"kmalloc-32M",      33554432},
>>>> +	{"kmalloc-64M",      67108864}
>>>
>>> I'd rather use KB and MB suffixes or at least capital 'K'.
>>
>> I like k and M better.
> 
> k and M work for me too.  It we were going to be anal then we should
> go with the IEC standard of KiB and MiB, but we're trying to make

But                           ^K and  ^M.

Small 'k' .. I don't know what that is.

> /proc/slabinfo a little less ugly, and so we have 16 characters to work
> with.  "dma-kmalloc-256KiB" is 18 characters.  The obvious place to lose
> two characters is the "iB" which are implicit; we know we're measuring
> bytes and the binary nature is assumed in this context of memory bytes
> (not storage bytes).  "dma-kmalloc-256k" is better.
> 


-- 
~Randy