This is the mail archive of the
elfutils-devel@sourceware.org
mailing list for the elfutils project.
Re: patch to commit soon: PR25375 debuginfod prefetching
- From: Mark Wielaard <mark at klomp dot org>
- To: "Frank Ch. Eigler" <fche at redhat dot com>, elfutils-devel at sourceware dot org
- Date: Tue, 25 Feb 2020 12:32:02 +0100
- Subject: Re: patch to commit soon: PR25375 debuginfod prefetching
- References: <20200224222946.GC26763@redhat.com>
Hi Frank,
On Mon, 2020-02-24 at 17:29 -0500, Frank Ch. Eigler wrote:
> This patch has been baking on my public servers awhile and can make a
> huge difference in performance. It's not something immediately
> obvious how or whether to test, as it's a pure performance improvement.
> Planning to push shortly.
As far as I understand it, it looks good.
One tiny question:
> @@ -2809,7 +2857,8 @@ main (int argc, char *argv[])
> fdcache_mbs = 1024; // 1 gigabyte
> else
> fdcache_mbs = sfs.f_bavail * sfs.f_bsize / 1024 / 1024 / 4; // 25% of free space
> - fdcache_fds = concurrency * 2;
> + fdcache_prefetch = 64; // guesstimate storage is this much less costly than re-decompression
> + fdcache_fds = (concurrency + fdcache_prefetch) * 2;
Here fdcache_prefetch is set and used before argp_parse () is called,
which would set it to the user supplied value (if any). Is that
intentional?
Cheers,
Mark