Debian fixes: s390 fail to build and more...

[Hector Oron]

Hello,

s390 and s390x seem to fail to build with current CVS snapshot
(2011-09-18) on Debian system.
I have put together a patch:
* http://patch-tracker.debian.org/patch/series/view/dietlibc/0.33~cvs20110918-1/0021-fix-ftbfs-s390.diff

I have also pulled Enrico Scholz changes for ARM so we get better
coverage for ARMv7 systems:
* http://patch-tracker.debian.org/patch/series/view/dietlibc/0.33~cvs20110918-1/0020-fixes-for-ARMv7.diff

In general, there are more fixes applied, all of them you can see from:
* http://patch-tracker.debian.org/package/dietlibc/0.33~cvs20110918-1

I would appreciate if you could find time to review this patches and
apply them to CVS HEAD if you consider those are fine.

Kind regards,

[Felix von Leitner]

Post by Hector Oron
s390 and s390x seem to fail to build with current CVS snapshot
(2011-09-18) on Debian system.
* http://patch-tracker.debian.org/patch/series/view/dietlibc/0.33~cvs20110918-1/0021-fix-ftbfs-s390.diff

Will look at it as soon as I can, I'm currently somewhere in the
wilderness without proper internet access.

Post by Hector Oron
I have also pulled Enrico Scholz changes for ARM so we get better
* http://patch-tracker.debian.org/patch/series/view/dietlibc/0.33~cvs20110918-1/0020-fixes-for-ARMv7.diff

I thought I had already pulled the arm patches in a while ago?
Are there new ones?

Post by Hector Oron
* http://patch-tracker.debian.org/package/dietlibc/0.33~cvs20110918-1
I would appreciate if you could find time to review this patches and
apply them to CVS HEAD if you consider those are fine.

I generally do aim to include all patches where I can see what they do
and where the code still builds after applying them :-)

Felix

[Hector Oron]

Hello,

Post by Felix von Leitner

  s390 and s390x seem to fail to build with current CVS snapshot
(2011-09-18) on Debian system.
   * http://patch-tracker.debian.org/patch/series/view/dietlibc/0.33~cvs20110918-1/0021-fix-ftbfs-s390.diff

Will look at it as soon as I can, I'm currently somewhere in the
wilderness without proper internet access.

Thanks, there is really no hurry.

Post by Felix von Leitner

  I have also pulled Enrico Scholz changes for ARM so we get better
   * http://patch-tracker.debian.org/patch/series/view/dietlibc/0.33~cvs20110918-1/0020-fixes-for-ARMv7.diff

I thought I had already pulled the arm patches in a while ago?
Are there new ones?

Maybe he can comment, but those apply cleanly on CVS head, those might
still need some cleanup, as I don't think we have support for
WANT_ELFINFO.

Post by Felix von Leitner

   * http://patch-tracker.debian.org/package/dietlibc/0.33~cvs20110918-1
  I would appreciate if you could find time to review this patches and
apply them to CVS HEAD if you consider those are fine.

I generally do aim to include all patches where I can see what they do
and where the code still builds after applying them :-)

Thanks, :-)

[Enrico Scholz]

Post by Felix von Leitner
I thought I had already pulled the arm patches in a while ago?
Are there new ones?

https://github.com/ensc/dietlibc/commits/rebase/

lists changesets against CVS HEAD. Most important changes are:

- *jmp() fixes for ARM

- syscall optimizations; e.g.

0000e048 <kill>:
e048: e3a0c025 mov ip, #37 ; 0x25
e04c: eafff45a b b1bc <__unified_syscall_swi>

versus

0000e0dc <kill>:
e0dc: e92d40b0 push {r4, r5, r7, lr}
e0e0: e59d4010 ldr r4, [sp, #16]
e0e4: e59d5014 ldr r5, [sp, #20]
e0e8: e3a07025 mov r7, #37 ; 0x25
e0ec: ef000000 svc 0x00000000
e0f0: eafff452 b b240 <__unified_syscall>

- a working getpagesize(2) implementation (code on non i386, x86_64
and ARM is untested and probably broken)

- improvements for recent ARM architectures (e.g. 'bx lr' instead of
'mov pc,lr')

- a largely improved testsuite



Enrico