This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: patch / RFD: speed up linker map file generation
- From: Ian Lance Taylor <ian at wasabisystems dot com>
- To: Ben Elliston <bje at au dot ibm dot com>
- Cc: binutils at sources dot redhat dot com
- Date: 08 Jun 2004 22:25:06 -0400
- Subject: Re: patch / RFD: speed up linker map file generation
- References: <200405051548.i45Fmm901764@linsvr1.uk.superh.com><40A8F8CD.6050004@redhat.com> <87y8mxh4nl.fsf@au.ibm.com>
Ben Elliston <bje@au.ibm.com> writes:
> Nick Clifton <nickc@redhat.com> writes:
>
> > There should definitely be an option. There are users who run these
> > tools on small machines who will want the old behaviour. I would
> > suggest making it a generalised switch (eg
> > --reduce-memory-overheads) that can be used to trigger other (as yet
> > unimplemented) tweaks to the linker's behaviour.
Note that the linker also has the --no-keep-memory option. I don't
sure why we need a separate --reduce-memory-overheads option, although
I concede that --no-keep-memory is probably not the best name.
> There are plenty of applications that detect the amount of available
> working memory and adjust their behaviour accordingly. Can't the
> linker do the same? It seems unfortunate that we mere humans must
> instruct the program to do something that it can probably figure out
> better itself.
We would need a semi-portable way of figuring out how much space we
could reasonably allocate. Given that, it's a mere matter of
programming.
(Simply running along casually until malloc returns NULL causes
massive thrashing along the way.)
Ian