Skip to contentSkip to conversation

slabify iocontext + request_queue

PATCH Christoph Hellwig 2 participants 3 messages May 15, 2004

lore
Share & tools
lore
#0 lore raw

this just went in:

Quoted 6 lines

From: Jens Axboe <[email protected]>

Move both request_queue and io_context allocation to a slab cache.

This is mainly a space-saving exercise. Some setups have a lot of disks and the kmalloc rounding-up can consume significant amounts of memory.

While I agree on the io_context part, slabifying request_queue is a space waste on most machines out there. The averange desktop has less than a handfull of these, and even for smaller servers it doesn't exactly look like a gain.

#1 lore raw
Quoted 6 lines · Christoph Hellwig

this just went in:

Quoted 4 lines

From: Jens Axboe <[email protected]>

Move both request_queue and io_context allocation to a slab cache.

Quoted 7 lines
Quoted 2 lines

This is mainly a space-saving exercise. Some setups have a lot of disks and the kmalloc rounding-up can consume significant amounts of memory.

While I agree on the io_context part, slabifying request_queue is a space waste on most machines out there. The averange desktop has less than a handfull of these, and even for smaller servers it doesn't exactly look like a gain.

See the thread last week on queue congestion threshold calculations, there were some numbers in there.

Signature · 1 line
Jens Axboe
#2 lore raw
Quoted 7 lines · Jens Axboe
Quoted 4 lines

While I agree on the io_context part, slabifying request_queue is a space waste on most machines out there. The averange desktop has less than a handfull of these, and even for smaller servers it doesn't exactly look like a gain.

See the thread last week on queue congestion threshold calculations, there were some numbers in there.

Maybe my math is completely off, but with slab you'd need a page at least, the kmem_cache_t and maybe a kmem_bufctl_t

So for the usual two or three queue desktops we went from 1024 or 1536 to 4096 + N.

Cutoff point is at aproximately 9 queues which I think most machines running linux won't reach.

Anyway, not that this is really important, I think we just need to question all this silent bloating a little..