Re: heimdal: FTBFS on hurd-i386 (for review)
On Tue, 2014-02-11 at 18:09 +0100, Samuel Thibault wrote:
> Svante Signell, le Tue 11 Feb 2014 17:48:50 +0100, a écrit :
> > uint64_t param1;
> > uint64_t syscall;
> > };
> > +#ifdef __GNU__
> > +/* FIXME: Enter the correct values here */
> > +#define _IOT_sundevdata _IOT(_IOTS(long), 7, 0, 0, 0, 0)
>
> Why long? The structure contains uint64_t, not longs. Did you read the
> ioctl part of the porting guideline wiki page?
>
> Also, the _ILP32 case needs to be handled, too.
>
> > +#else
> > #define VIOC_SUN_SYSCALL_DEV _IOW('C', 1, struct sundevdata)
> > #endif
> > +#endif /* __GNU__ */
>
> Why not defining VIOC_SUN_SYSCALL_DEV in the GNU case? Also, take care
> of the #ifdef/#endif parenthesing, GNU does not belong to where you have
> put it.
New patch attached. Regarding the use of _IOT etc, I'm trying to follow
the same construct as you and Pino did in bug#483281
Currently building again (takes a long time)
--- a/lib/kafs/afssys.c 2013-12-07 14:36:38.000000000 +0100
+++ b/lib/kafs/afssys.c 2014-02-11 18:35:53.000000000 +0100
@@ -76,6 +76,10 @@ struct sundevdata {
uint32_t param1;
uint32_t syscall;
};
+#ifdef __GNU__
+/* FIXME: Enter the correct values here */
+#define _IOT_sundevdata _IOT(_IOTS(uint32_t), 7, 0, 0, 0, 0)
+#endif
#define VIOC_SUN_SYSCALL_DEV _IOW('C', 2, struct sundevdata)
#else
struct sundevdata {
@@ -87,6 +91,10 @@ struct sundevdata {
uint64_t param1;
uint64_t syscall;
};
+#ifdef __GNU__
+/* FIXME: Enter the correct values here */
+#define _IOT_sundevdata _IOT(_IOTS(uint64_t), 7, 0, 0, 0, 0)
+#endif
#define VIOC_SUN_SYSCALL_DEV _IOW('C', 1, struct sundevdata)
#endif
#endif /* _IOW */
Reply to: