Re: GLibc Patch

Re: GLibc Patch

[Dan Kegel]

[+crossgcc]

Hi Neil,
I'm off to other things, and haven't had time to do anything with
crosstool for some time, so I'm cc'ing your message to the
crossgcc mailing list.

Thanks,
Dan


On Tue, May 26, 2009 at 9:31 AM, Neil Gierman <ngierman@roadrunn.com> wrote:
> I am using crosstool-0.43 to build a linux-elf cross compiler on cygwin (so
> my windows system can build linux elf binaries). I have run into an issue
> that I have patched and wanted to let you know so you can include if
> necessary. I am getting the glibc failure about junk at end of file when
> processing setjmp.S. From messages like
> http://www.cygwin.com/ml/crossgcc/2004-10/msg00170.html, I have patched my
> glibc-2.3.6/sysdeps/unix/sysv/linux/i386/sysdep.h to undef __i686 if it is
> defined. The same patch works for both glibc 2.3.6 and 2.3.5. It would be
> nice if you could include this in the next crosstool build.
>
>
>
>
>
> --- C:\Users\giermang\Desktop\sysdep.h.old 2009-05-26 11:27:47.000000000
> -0500
> +++ C:\Users\giermang\Desktop\sysdep.h.new 2009-05-26 11:27:32.000000000
> -0500
> @@ -26,12 +26,17 @@
> #include <bp-sym.h>
> #include <bp-asm.h>
> /* Defines RTLD_PRIVATE_ERRNO and USE_DL_SYSINFO.  */
> #include <dl-sysdep.h>
> #include <tls.h>
>
> +/* gcc-3.4 seems to expand __i686 as a predefined preprocessor symbol,
> which we don't want here */
> +#ifdef __i686
> +#undef __i686
> +#endif
> +
> /* For Linux we can use the system call table in the header file
>          /usr/include/asm/unistd.h
>     of the kernel.  But these symbols do not follow the SYS_* syntax
>     so we have to redefine the `SYS_ify' macro here.  */
> #undef SYS_ify
> #define SYS_ify(syscall_name)   __NR_##syscall_name

--
For unsubscribe information see http://sourceware.org/lists.html#faq

Re: GLibc Patch

[Khem Raj]

On Tue, May 26, 2009 at 10:16 AM, Dan Kegel<dank@kegel.com> wrote:
> [+crossgcc]
>
> Hi Neil,
> I'm off to other things, and haven't had time to do anything with
> crosstool for some time, so I'm cc'ing your message to the
> crossgcc mailing list.
>
> Thanks,
> Dan
>
>
> On Tue, May 26, 2009 at 9:31 AM, Neil Gierman <ngierman@roadrunn.com> wrote:
>> I am using crosstool-0.43 to build a linux-elf cross compiler on cygwin (so
>> my windows system can build linux elf binaries). I have run into an issue
>> that I have patched and wanted to let you know so you can include if
>> necessary. I am getting the glibc failure about junk at end of file when
>> processing setjmp.S. From messages like
>> http://www.cygwin.com/ml/crossgcc/2004-10/msg00170.html, I have patched my
>> glibc-2.3.6/sysdeps/unix/sysv/linux/i386/sysdep.h to undef __i686 if it is
>> defined. The same patch works for both glibc 2.3.6 and 2.3.5. It would be
>> nice if you could include this in the next crosstool build.
>>
>>
>>
>>
>>
>> --- C:\Users\giermang\Desktop\sysdep.h.old 2009-05-26 11:27:47.000000000
>> -0500
>> +++ C:\Users\giermang\Desktop\sysdep.h.new 2009-05-26 11:27:32.000000000
>> -0500
>> @@ -26,12 +26,17 @@
>> #include <bp-sym.h>
>> #include <bp-asm.h>
>> /* Defines RTLD_PRIVATE_ERRNO and USE_DL_SYSINFO.  */
>> #include <dl-sysdep.h>
>> #include <tls.h>
>>
>> +/* gcc-3.4 seems to expand __i686 as a predefined preprocessor symbol,
>> which we don't want here */
>> +#ifdef __i686
>> +#undef __i686
>> +#endif
>> +
>> /* For Linux we can use the system call table in the header file
>>          /usr/include/asm/unistd.h
>>     of the kernel.  But these symbols do not follow the SYS_* syntax
>>     so we have to redefine the `SYS_ify' macro here.  */
>> #undef SYS_ify
>> #define SYS_ify(syscall_name)   __NR_##syscall_name
>
> --
> For unsubscribe information see http://sourceware.org/lists.html#faq
>
>

Does this patch help
http://www.eglibc.org/archives/patches/msg00073.html

--
For unsubscribe information see http://sourceware.org/lists.html#faq

RE: GLibc Patch

[Neil Gierman]

> -----Original Message-----
> From: Khem Raj [mailto:raj.khem@gmail.com]
> 
> Does this patch help
> http://www.eglibc.org/archives/patches/msg00073.html


Yes. That patch also solves my error.

-Neil



--
For unsubscribe information see http://sourceware.org/lists.html#faq