[PATCH] aarch64: allow adding/removing just feature flags via .arch_extension
Marcus Shawcroft
marcus.shawcroft@gmail.com
Mon Nov 3 17:44:00 GMT 2014
More information about the Binutils mailing list
Mon Nov 3 17:44:00 GMT 2014
- Previous message (by thread): Ping: [PATCH] aarch64: allow adding/removing just feature flags via .arch_extension
- Next message (by thread): [PATCH] aarch64: allow adding/removing just feature flags via .arch_extension
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi On 24 October 2014 13:40, Jan Beulich <JBeulich@suse.com> wrote: > static int > -aarch64_parse_features (char *str, const aarch64_feature_set **opt_p) > +aarch64_parse_features (char *str, const aarch64_feature_set **opt_p, > + char *ext) The new parameter here is being used as a boolean flag, represent it with a bfd_boolean rather than a char *. > { > /* We insist on extensions being added before being removed. We achieve > this by using the ADDING_VALUE variable to indicate whether we are > @@ -7267,16 +7270,18 @@ aarch64_parse_features (char *str, const > while (str != NULL && *str != 0) > { > const struct aarch64_option_cpu_value_table *opt; > - char *ext; > int optlen; > > - if (*str != '+') > + if (!ext) > { > - as_bad (_("invalid architectural extension")); > - return 0; > - } > + if (*str != '+') > + { > + as_bad (_("invalid architectural extension")); > + return 0; > + } > > - str++; > + str++; > + } > ext = strchr (str, '+'); The logic here is odd. I don;t think we want this in the case when ext != NULL. The effect is that: .arch_extension +crc ...will give a "missing extension" rather than an "unknown extension" error message. Cheers /Marcus
- Previous message (by thread): Ping: [PATCH] aarch64: allow adding/removing just feature flags via .arch_extension
- Next message (by thread): [PATCH] aarch64: allow adding/removing just feature flags via .arch_extension
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Binutils mailing list