This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH 5/8] Move i486/pthread_spin_trylock.S to pthread_spin_trylock.S
- From: Roland McGrath <roland at hack dot frob dot com>
- To: Torvald Riegel <triegel at redhat dot com>
- Cc: "H.J. Lu" <hjl dot tools at gmail dot com>, Joseph Myers <joseph at codesourcery dot com>, GNU C Library <libc-alpha at sourceware dot org>
- Date: Wed, 26 Aug 2015 14:01:02 -0700 (PDT)
- Subject: Re: [PATCH 5/8] Move i486/pthread_spin_trylock.S to pthread_spin_trylock.S
- Authentication-results: sourceware.org; auth=none
- References: <20150826173721 dot GF21261 at intel dot com> <alpine dot DEB dot 2 dot 10 dot 1508261800370 dot 13146 at digraph dot polyomino dot org dot uk> <CAMe9rOp7zoMCsMLY2WoEucLKkrR6oJZg9XDJa2S-OJmfA08C8A at mail dot gmail dot com> <1440622616 dot 30828 dot 46 dot camel at localhost dot localdomain>
> Could you instead please try to replace the custom asm implementation
> with a C implementation, preferably the generic one in
> nptl/pthread_spin_trylock.c? And, if necessary, and improve the latter
> if there is a significant performance difference for uncontended locks?
As I just said in another thread, meaningful changes such as that should
not be conflated with the mechanical directory restructuring HJ is doing
now. Of course, doing removals first reduces the number of renamings
required. But the directory restructuring is entirely mechanical and so
its review can consist of verifying that nothing materially changed in the
build whatsoever. Meaningful changes require meaningful review.
Directory restructuring changes, when generally desireable, should never be
gated on unrelated meaningful changes. So if HJ wants to do those
meaningful improvements first, that's fine. But nobody should be objecting
to the directory restructuring changes on the grounds that some other
meaningful change is also desireable.