patrick/zephyr
1
0
Fork 0
Code Issues Pull requests Actions 7 Packages Projects Releases Wiki Activity

k_mem_pool_alloc(): remove dead code

Browse source 
A loop in k_mem_pool_alloc() around z_sys_mem_pool_block_alloc() assumes
the later may return -EAGAIN with an elaborate comment about it. But
-EAGAIN is no longer returned by that function since commit 7845e1b01e
("lib/mempool: Fix spurious -ENOMEM due to agressive latency control").

Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
This commit is contained in:
Nicolas Pitre 2019-10-04 12:10:07 -04:00 committed by Andrew Boie
parent f0ddbd7eee
commit d12462ca6c
1 changed files with 3 additions and 20 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

23
kernel/mempool.c
View file

@ -61,26 +61,9 @@ int k_mem_pool_alloc(struct k_mem_pool *p, struct k_mem_block *block,
while (true) {
u32_t level_num, block_num;
/* There is a "managed race" in alloc that can fail
* (albeit in a well-defined way, see comments there)
* with -EAGAIN when simultaneous allocations happen.
* Retry exactly once before sleeping to resolve it.
* If we're so contended that it fails twice, then we
* clearly want to block.
*/
for (int i = 0; i < 2; i++) {
ret = z_sys_mem_pool_block_alloc(&p->base, size,
&level_num, &block_num,
&block->data);
if (ret != -EAGAIN) {
break;
}
}
if (ret == -EAGAIN) {
ret = -ENOMEM;
}
ret = z_sys_mem_pool_block_alloc(&p->base, size,
&level_num, &block_num,
&block->data);
block->id.pool = pool_id(p);
block->id.level = level_num;
block->id.block = block_num;