Re: Re: [parisc-linux-cvs] linux-2.6.git: Changes to 'master'

On 12/22/06, Helge Deller <[email protected]> wrote:
> On Sat Dec 23 2006, Helge Deller wrote:
> >  arch/parisc/kernel/syscall_table.S |    4 ++++
> >  include/asm-parisc/unistd.h        |    6 +++++-
> >  2 files changed, 9 insertions(+), 1 deletion(-)
> >
> > New commits:
> > http://git.parisc-linux.org/?p=linux-2.6.git;a=commitdiff;h=5f0f48d1146bd6fce7b5a480625d741320b3b3cd
> > commit 5f0f48d1146bd6fce7b5a480625d741320b3b3cd
> > Author: Helge Deller <[email protected]>
> > Date:   Sat Dec 23 00:12:02 2006 +0100
> >
> >     add missing syscalls for vmsplice, move_pages, getcpu & epoll_pwait
>
> maybe people here could double-check if my ENTRY_SAME/ENTRY_COMP choice is correct ?
>
> +       ENTRY_SAME(vmsplice)
> +       ENTRY_COMP(move_pages)          /* 295 */
> +       ENTRY_SAME(getcpu)
> +       ENTRY_SAME(epoll_pwait)
>
> At least now the vmsplice01 (splice & vmsplice) test of Linux Test Project will pass as well, if people use new glibc, where PARISC's __NR_vmsplice == 294 is defined.
>
> The only still failing tests are
> fstatat01    1  FAIL  :  fstatat() Failed, errno=4 : Interrupted system call
> getcontext01    1  FAIL  :  getcontext - Sanity test :  Fail errno=251 : Function not implemented

I added the vmsplice definitions in upstream libc-ports for hppa.

The getcontext01 fail is because nobody has written get/set/swap
context for hppa. Someone needs to write them if you want the test to
pass :-)

c.