First Last Prev Next    No search results available      Search page      Enter new bug
Bug#: 147126
Alias:
Product:
Component:
Status: RESOLVED
Resolution: FIXED
Assigned To: Gentoo Toolchain Maintainers <toolchain@gentoo.org>
Hardware:
OS:
Version:
Priority:
Severity:
Reporter: SpanKY <vapier@gentoo.org>
Add CC:
CC:
URL:
Summary:
Status Whiteboard:
Keywords:

Filename Description Type Creator Created Size Actions
Create a New Attachment (proposed patch, testcase, etc.) View All

Bug 147126 depends on: Show dependency tree
Show dependency graph
Bug 147126 blocks:
Votes: 0    Show votes for this bug    Vote for this bug

Additional Comments: (this is where you put emerge --info)


Not eligible to see or edit group visibility for this bug.






View Bug Activity   |   Format For Printing   |   XML   |   Clone This Bug


Description:   Opened: 2006-09-10 22:31 0000
it's about time we kill off these split packages and move to a common one
(kgcc64)

ive been using kgcc64 myself on powerpc and hppa and i know Kumba has been
testing mips ... if you have any issues, please post them here

------- Comment #1 From SpanKY 2006-09-25 06:20:13 0000 -------
no feedback !?  to be sure, i'd like to have these things out of the tree by
2006.2/2007.0 ...

------- Comment #2 From Stephen Becker (RETIRED) 2006-09-25 06:27:07 0000 -------
(In reply to comment #1)
> no feedback !?  to be sure, i'd like to have these things out of the tree by
> 2006.2/2007.0 ...
> 

mips is fine here

I've also been using it on sparc, and it works fine, except that weeve will
probably still have his panties all bunched up because it doesn't provide
"sparc64-linux-gcc" to satisfy the default kernel makefile (looks like it
provides sparc64-linux-gnu-gcc instead).  In fact, when I mentioned a problem I
was having with kgcc64 at one point on sparc (turned out to be pebkac), weeve
had no idea that such a thing existed, and was thoroughly astonished that it
even had a ~sparc keyword.  So perhaps you should poke the sparc guys a bit
harder on this one...

------- Comment #3 From Markus Rothe 2006-09-25 06:28:41 0000 -------
(In reply to comment #0)
> if you have any issues, please post them here

I thought leaving no comment means 'no issues'. ;-)

Anyway.. I haven't seen issues on ppc64 so far, though I'm not using 32bit
userland that often.

------- Comment #4 From SpanKY 2006-09-25 06:34:25 0000 -------
i can have kgcc64 generate an $ARCH-linux-{gcc,cpp} wrapper as well ... i just
thought that people wanted $ARCH-linux-gnu-{gcc,cpp}

------- Comment #5 From Stephen Becker (RETIRED) 2006-09-25 07:56:13 0000 -------
(In reply to comment #4)
> i can have kgcc64 generate an $ARCH-linux-{gcc,cpp} wrapper as well ... i just
> thought that people wanted $ARCH-linux-gnu-{gcc,cpp}

That $ARCH-linux-gnu-{gcc,cpp} would require a patch or sed to the kernel
Makefile to work, in which case, you might as well stick with the default
$ARCH-$MACHINE-linux-gnu-{gcc,cpp}.

------- Comment #6 From Jason Wever (RETIRED) 2006-09-25 08:09:27 0000 -------
So on sparc64, to satisfy either the upstream makefiles or the ones in any of
our kernel ebuilds that support sparc64, we'll need a 'sparc64-linux-gcc' to
exist.  By default the makefiles check to see if gcc supports -m64, and if they
don't, it looks for sparc64-linux-gcc.  Since we don't support -m64,
gcc-sparc64 provides the latter.  I'd like to keep providing that going forward
to maintain the current behavior.

------- Comment #7 From SpanKY 2006-09-25 09:29:14 0000 -------
done then ... kgcc64 no longer provides $arch-linux-gnu-gcc, just
$arch-linux-gcc

------- Comment #8 From Jason Wever (RETIRED) 2006-09-30 16:41:38 0000 -------
I think we're pretty close on SPARC.  Only obvious thing I see now is having to
adjust the sys-kernel/linux-headers logic WRT the gcc64 use flag.

------- Comment #9 From Jeroen Roovers 2006-10-09 07:59:45 0000 -------
(In reply to comment #7)
> done then ... kgcc64 no longer provides $arch-linux-gnu-gcc, just
> $arch-linux-gcc

And what provides $arch-linux-ld?

/usr/src/linux $ make
  [...]
  CC      init/do_mounts.o
  LD      init/mounts.o
/bin/sh: hppa64-linux-ld: command not found

------- Comment #10 From SpanKY 2006-10-09 09:28:57 0000 -------
that's a binutils issue, not a gcc issue

hppa has "binutils-hppa64" because the hppa target does not yet support unified
32/64 bit hppa support

------- Comment #11 From Tom Gall 2006-11-04 12:40:18 0000 -------
I presume this is only on those architectures which run mixed where the primary
gcc is different then the gcc needed to build the kernel for whatever reason.

If so, then ppc64 likely doesn't need to be on this bug as the user space
compiler is the same as one would build the kernel with. However those that run
on a ppc64 kernel but 32 bit user space might have an interest here *if* the
gcc compiler isn't being built correctly with both 32/64 bit enabled. Further
if there is some creation of say something like a powerpc64-linux-kgcc kind of
thing, then really the ebuild on ppc64 would just need to create the symlinks
and that'd be that.

Feel free to drop by #gentoo-ppc64 to discuss.

Thanks!

Tom

------- Comment #12 From SpanKY 2006-11-04 15:35:23 0000 -------
this applies to every team that has their own "gcc-$ARCH" ... powerpc has
"sys-devel/powerpc64" and it needs to go away

------- Comment #13 From Guy Martin 2006-11-07 16:02:46 0000 -------
binutils-hppa64 fixed (thanks Spanky)
kgcc64 marked stable on hppa
moved gcc-hppa64 to kgcc64. will remove the old soon :)

------- Comment #14 From SpanKY 2006-12-31 05:20:48 0000 -------
poke poke ... if arches are OK with kgcc64, then feel free to finally make the
move and scrub the old cruft

------- Comment #15 From Gustavo Zacarias (RETIRED) 2007-01-18 12:57:08 0000 -------
We've switched.

------- Comment #16 From Joe Jezak 2007-02-21 04:18:50 0000 -------
This is for ppc64 to decide, ppc32 should be fine with it, so I'm removing us
from the CC list.

------- Comment #17 From Guy Martin 2007-02-28 17:25:10 0000 -------
I just dropped gcc-hppa64 and moved it to kgcc64.

Please re-add hppa if something is missing or else.

------- Comment #18 From Markus Rothe 2007-03-29 14:13:52 0000 -------
ppc64 done

First Last Prev Next    No search results available      Search page      Enter new bug