[igt-dev] [Intel-gfx] [PATCH i-g-t 3/5] i915: Exercise preemption timeout controls in sysfs

[Chris Wilson]

Quoting Andi Shyti (2020-02-29 12:45:27)
> > > > > > +     char buf[512];
> > > > > > +     int len;
> > > > > > +
> > > > > > +     lseek(engines, 0, SEEK_SET);
> > > > > > +     while ((len = syscall(SYS_getdents64, engines, buf, sizeof(buf))) > 0) {
> > > > > > +             void *ptr = buf;
> > > > > > +
> > > > > > +             while (len) {
> > > > > > +                     struct linux_dirent64 {
> > > > > > +                             ino64_t        d_ino;
> > > > > > +                             off64_t        d_off;
> > > > > > +                             unsigned short d_reclen;
> > > > > > +                             unsigned char  d_type;
> > > > > > +                             char           d_name[];
> > > > > > +                     } *de = ptr;
> > > > > 
> > > > > what is the need for having your own linux_dirent64?
> > > > 
> > > > fdopendir() takes ownership of the fd, preventing reuse. And
> > > > fdopendir(dup()) is getting ridiculous.
> > > 
> > > why not using dirent64?
> > 
> > It's still the same problem that it takes a DIR, assuming ownership of
> > the fd. Why using linux_dirent64 because the manpage says so -- if you
> > are going to use the syscall, you need to match it's calling
> > conventions, not a middleman's.
> 
> I understand, but in bits/dirent.h there is, with some
> assumptions, this part:
> 
> #ifdef __USE_LARGEFILE64
> struct dirent64
>   {
>     __ino64_t d_ino;
>     __off64_t d_off;
>     unsigned short int d_reclen;
>     unsigned char d_type;
>     char d_name[256];           /* We must not include limits.h! */
>   };
> #endif
> 
> why redefine a struct linux_dirent64?

Because the manpage didn't mention that!
-Chris