ref: move shmem lock init #271
Conversation
josecm
force-pushed
the
ref/shmem_lock_init
branch
from
d4b74ea to
538de6c
Compare
There was a problem hiding this comment.
I'm not understanding the goal here to be honest.
Isn't the end result the same?
I feel like it might be related to separate allocation from initialization, but I'm not sure.
If anything I'd consider moving the .cpu_masters = 0 to the alloc function to avoid iterating the array again.
Signed-off-by: Jose Martins <[email protected]>
josecm
dismissed stale reviews from miguelafsilva5 and danielRep
via
47863df
josecm
force-pushed
the
ref/shmem_lock_init
branch
from
538de6c to
47863df
Compare
I just think it makes more sense for the lock and the cpu_masters to be initialized in the init function and not in the alloc function. The allocation function should just take care of the allocation. I also don't the iterating this array is performance sensitive (only done once at init, and most likely only a handlful of shmem regions per configuration) |
No description provided.