This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
RE: PATCH: Speed up ELF section merge
- From: "Dave Korn" <dave dot korn at artimi dot com>
- To: "'H. J. Lu'" <hjl at lucon dot org>,<binutils at sources dot redhat dot com>
- Date: Fri, 6 May 2005 15:38:57 +0100
- Subject: RE: PATCH: Speed up ELF section merge
----Original Message----
>From: H. J. Lu
>Sent: 02 May 2005 03:29
> Here is the new one.
>
>
> H.J.
> ----
> 2005-05-01 H.J. Lu <hongjiu.lu@intel.com>
>
> * merge.c (sec_merge_init): Call bfd_hash_table_init_n with
> hash table size 16699 instead of bfd_hash_table_init.
>
> --- bfd/merge.c.hash 2005-04-15 18:51:49.000000000 -0700
> +++ bfd/merge.c 2005-05-01 19:25:42.000000000 -0700
> @@ -241,7 +241,8 @@ sec_merge_init (unsigned int entsize, bf
> if (table == NULL)
> return NULL;
>
> - if (! bfd_hash_table_init (&table->table, sec_merge_hash_newfunc))
> + if (! bfd_hash_table_init_n (&table->table, sec_merge_hash_newfunc,
> + 16699))
> {
> free (table);
> return NULL;
H.J., IIUIC this patch can't do anything to support the ld command line
options --hash-size=<NUMBER> and --reduce-memory-overheads; wouldn't it be
worthwhile providing a means of controlling whether it uses a large hash
table or not?
cheers,
DaveK
--
Can't think of a witty .sigline today....