From: Marek Olšák Date: Thu, 6 Oct 2016 22:04:28 +0000 (+0200) Subject: ralloc: don't memset ralloc_header, clear it manually X-Git-Tag: upstream/17.1.0~5118 X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=330482177c45dceb71e5787c1a7e7fc87308c144;p=platform%2Fupstream%2Fmesa.git ralloc: don't memset ralloc_header, clear it manually time GALLIUM_NOOP=1 ./run shaders/private/alien_isolation/ >/dev/null Before (2 takes): real 0m8.734s 0m8.773s user 0m34.232s 0m34.348s sys 0m0.084s 0m0.056s After (2 takes): real 0m8.448s 0m8.463s user 0m33.104s 0m33.160s sys 0m0.088s 0m0.076s Average change in "real" time spent: -3.4% calloc should only do 2 things compared to malloc: - check for overflow of "n * size" - call memset I'm not sure if that explains the difference. v2: clear "parent" and "next" in the caller of add_child. Reviewed-by: Edward O'Callaghan (v1) Tested-by: Edmondo Tommasina (v1) Reviewed-by: Nicolai Hähnle (v1) --- diff --git a/src/util/ralloc.c b/src/util/ralloc.c index 9526011..7023d79 100644 --- a/src/util/ralloc.c +++ b/src/util/ralloc.c @@ -122,13 +122,27 @@ ralloc_size(const void *ctx, size_t size) void * rzalloc_size(const void *ctx, size_t size) { - void *block = calloc(1, size + sizeof(ralloc_header)); + void *block = malloc(size + sizeof(ralloc_header)); ralloc_header *info; ralloc_header *parent; if (unlikely(block == NULL)) return NULL; + info = (ralloc_header *) block; + /* measurements have shown that calloc is slower (because of + * the multiplication overflow checking?), so clear things + * manually + */ + info->parent = NULL; + info->child = NULL; + info->prev = NULL; + info->next = NULL; + info->destructor = NULL; + + /* memset the allocation except for ralloc_header */ + memset(&info[1], 0, size); + parent = ctx != NULL ? get_header(ctx) : NULL; add_child(parent, info);