[13/23] sim/erc32: Fix a few compiler warnings
Commit Message
Minor edits to remove compiler warnings.
---
sim/erc32/interf.c | 2 +-
sim/erc32/sis.h | 1 -
2 files changed, 1 insertion(+), 2 deletions(-)
Comments
On 17 Feb 2015 08:44, Jiri Gaisler wrote:
> Minor edits to remove compiler warnings.
if you're really interested in doing this, you could add SIM_AC_OPTION_WARNINGS
to the erc32/configure.ac script ...
> --- a/sim/erc32/interf.c
> +++ b/sim/erc32/interf.c
> @@ -509,7 +509,7 @@ sim_do_command(sd, cmd)
> SIM_DESC sd;
> const char *cmd;
> {
> - exec_cmd(&sregs, cmd);
> + exec_cmd(&sregs, (char *) cmd);
> }
shouldn't exec_cmd be changed to take a const char * instead ? looks to me like
it treats it as constant ... it does a strdup(cmd) and only operates on the
result.
looking further, seems like it this func really should use buildargv/freeargv
rather than tokenize things itself ?
-mike
On 02/17/2015 12:07 PM, Mike Frysinger wrote:
> On 17 Feb 2015 08:44, Jiri Gaisler wrote:
>> Minor edits to remove compiler warnings.
>
> if you're really interested in doing this, you could add SIM_AC_OPTION_WARNINGS
> to the erc32/configure.ac script ...
>
>> --- a/sim/erc32/interf.c
>> +++ b/sim/erc32/interf.c
>> @@ -509,7 +509,7 @@ sim_do_command(sd, cmd)
>> SIM_DESC sd;
>> const char *cmd;
>> {
>> - exec_cmd(&sregs, cmd);
>> + exec_cmd(&sregs, (char *) cmd);
>> }
>
> shouldn't exec_cmd be changed to take a const char * instead ? looks to me like
> it treats it as constant ... it does a strdup(cmd) and only operates on the
> result.
I am passing cmd to strok() at one place, so const char cannot be used.
>
> looking further, seems like it this func really should use buildargv/freeargv
> rather than tokenize things itself ?
> -mike
>
I will keep this on my soon-todo list. Please bare with it this time ...
Jiri.
On 18 Feb 2015 17:21, Jiri Gaisler wrote:
> On 02/17/2015 12:07 PM, Mike Frysinger wrote:
> > On 17 Feb 2015 08:44, Jiri Gaisler wrote:
> >> Minor edits to remove compiler warnings.
> >
> > if you're really interested in doing this, you could add SIM_AC_OPTION_WARNINGS
> > to the erc32/configure.ac script ...
> >
> >> --- a/sim/erc32/interf.c
> >> +++ b/sim/erc32/interf.c
> >> @@ -509,7 +509,7 @@ sim_do_command(sd, cmd)
> >> SIM_DESC sd;
> >> const char *cmd;
> >> {
> >> - exec_cmd(&sregs, cmd);
> >> + exec_cmd(&sregs, (char *) cmd);
> >> }
> >
> > shouldn't exec_cmd be changed to take a const char * instead ? looks to me like
> > it treats it as constant ... it does a strdup(cmd) and only operates on the
> > result.
>
> I am passing cmd to strok() at one place, so const char cannot be used.
ah, i see that now yes. i'm not sure deploying the cast is correct as the API
allows for a constant to be passed in at which time this would crash.
> > looking further, seems like it this func really should use buildargv/freeargv
> > rather than tokenize things itself ?
>
> I will keep this on my soon-todo list. Please bare with it this time ...
np. in the meantime, adding a second dupe should allow for minimal code change.
-mike
@@ -509,7 +509,7 @@ sim_do_command(sd, cmd)
SIM_DESC sd;
const char *cmd;
{
- exec_cmd(&sregs, cmd);
+ exec_cmd(&sregs, (char *) cmd);
}
char **
@@ -203,7 +203,6 @@ extern int check_bpt (struct pstate *sregs);
extern void reset_all (void);
extern void sys_reset (void);
extern void sys_halt (void);
-extern int bfd_load (char *fname);
extern double get_time (void);
/* exec.c */