Re: [PATCH 11/18] lib/stackdepot: rename slab variables

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 2/1/23 13:38, Marco Elver wrote:
> On Tue, 31 Jan 2023 at 20:06, Andrey Konovalov <andreyknvl@xxxxxxxxx> wrote:
>>
>> On Tue, Jan 31, 2023 at 12:59 PM Alexander Potapenko <glider@xxxxxxxxxx> wrote:
>> >
>> > On Mon, Jan 30, 2023 at 9:50 PM <andrey.konovalov@xxxxxxxxx> wrote:
>> > >
>> > > From: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
>> > >
>> > > Give better names to slab-related global variables: change "depot_"
>> > > prefix to "slab_" to point out that these variables are related to
>> > > stack depot slabs.
>> >
>> > I started asking myself if the word "slab" is applicable here at all.
>> > The concept of preallocating big chunks of memory to amortize the
>> > costs belongs to the original slab allocator, but "slab" has a special
>> > meaning in Linux, and we might be confusing people by using it in a
>> > different sense.
>> > What do you think?
>>
>> Yes, I agree that using this word is a bit confusing.
>>
>> Not sure what be a good alternative though. "Region", "block",
>> "collection", and "chunk" come to mind, but they don't reflect the
>> purpose/usage of these allocations as good as "slab". Although it's
>> possible that my perception as affected by overly frequently looking
>> at the slab allocator internals :)
>>
>> Do you have a suggestion of a better word?
> 
> I'd vote for "pool" and "chunk(s)" (within that pool).

+1, also wasn't happy that "slab" is being used out of the usual context here :)

Thanks




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Bugtraq]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux