On Mon, Jan 24, 2022 at 4:59 AM Dongming Liu <ldming101@gmail.com> wrote:
> Maybe we can use one of the following methods to fix it:
> 1. re-bin segment to suitable segment index when called dsa_free
> 2. get_best_segment search all bins
(2) is definitely the wrong idea. The comments say:
/*
* What is the lowest bin that holds segments that *might* have n contiguous
* free pages? There is no point in looking in segments in lower bins; they
* definitely can't service a request for n free pages.
*/
#define contiguous_pages_to_segment_bin(n) Min(fls(n), DSA_NUM_SEGMENT_BINS - 1)
So it's OK for a segment to be in a bin that suggests that it has more
consecutive free pages than it really does. But it's NOT ok for a
segment to be in a bin that suggests it has fewer consecutive pages
than it really does. If dsa_free() is putting things back into the
wrong place, that's what we need to fix.
--
Robert Haas
EDB: http://www.enterprisedb.com