This is the mail archive of the libc-alpha@sources.redhat.com mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Add sparc64 tls and nptl support


On Wed, 13 Apr 2005 18:53:58 -0400
Jakub Jelinek <jakub@redhat.com> wrote:

> If arch_minimum_kernel != 2.0.10, then __LINUX_KERNEL_VERSION is never 0,
> always at least that minimum kernel version:
 ...
> On ia64, arch_minimum_kernel is not 2.0.10, therefore
> __LINUX_KERNEL_VERSION will never be 0 and I was arguing that
> on sparc64 we should do similarly.

Now I understand, ok I'll cook something up.

> Yes.  I got some failures:
> make[2]: *** [/usr/src/libc/obj/nptl/tst-align2.out] Error 1
 ...
> make[2]: *** [/usr/src/libc/obj/nptl/tst-getpid1.out] Error 1

These are due to a bug in sparc32 and sparc64 clone.S, which I have
fixed in my tree already.  We can't store values in local registers
across the clone because the kernel doesn't copy that register
window.  Instead we need to hold onto the values in global registers
across the system call.

> > It warns because the guarding ifdefs around the two different
> > possible elf_machine_type_class defines changes the two times
> > rtld.c ends up including dl-machine.h
> 
> Yeah, I know.

Are you saying that it doesn't matter that this macro is defined
correctly in the two cases?  It seems that your include guarding
patch only allows it to get set the first time, so it will only
take on that first definition, is that what we want?


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]