This is the first of a series of 7 invasive patches for the -mm tree, which
are to be reviewed (not only by UML folks), and possibly merged for the
2.6.13 cycle.
The first one splits the i386 syscall table out of entry.S, without any real
change for them (the file is included in the old place); if there are any
syscall table changes, please be careful; this maybe means place *this* one
before the others (or even merge now the i386 code movement part); I don't
want that this patch is modified and that subtle bugs are introduced. The
code movement is a really trivial code movement, however (no hidden changes).
This is needed to enable us to include the i386 syscall table, so that we are
sure that they match. I already handled the real differences between i386 and
UML (see the patch).
uml-fix-syscall-table-i386.patch
uml: fix syscall table by including $(SUBARCH)'s one, for i386
uml-quick-fix-syscall-table-x86_64.patch
uml: quick fix syscall table for x86_64
uml-fix-syscall-table-rem-for-x86-64.patch
uml: fix syscall table by including $(SUBARCH)'s one, for x86-64
uml-console-redo-locking.patch
uml: redo console locking
[ It's invasive, but seems to work well; wide testing is needed ]
uml-hostfs-avoid-buffers.patch
uml - hostfs: avoid buffers
[ This is something which should work, even because you said that; I've
partially verified this myself ]
uml-commentary-about-forking-flag.patch
uml: commentary about forking flag
[ Jeff, give a look and possibly correct it. ]
uml-ubd-handle-readonly.patch
uml ubd: handle readonly status
[ I cc'ed Jens Axboe ]
They are not to be confused with the previous series of 6 patches I also sent
today, which are mostly little or already tested patches (list below); the
explicitly marked one are urgent for 2.6.12, while for the other just take
your choice (they should be trivial and have been tested):
[LIST of patches TO MERGE]:
uml-workaround-sed-behaviour.patch
uml: workaround old problematic sed behaviour [compile-fix, for 2.6.12]
uml-nfsd-syscall.patch
uml: add nfsd syscall when nfsd is modular [for 2.6.12]
uml-crypto-i586.patch
uml: support AES i586 crypto driver [for 2.6.12]
uml-Makefile-avoid-rebuild.patch
uml kbuild: avoid useless rebuilds [for 2.6.13]
uml-inline-empty.patch
uml: inline empty proc
uml-move-va-copy-conditional.patch
uml: move va_copy conditional def
Thanks a lot for the attention and regards.
--
Paolo Giarrusso, aka Blaisorblade
Skype user "PaoloGiarrusso"
Linux registered user n. 292729
http://www.user-mode-linux.org/~blaisorblade
Blaisorblade wrote:
> This is the first of a series of 7 invasive patches for the -mm tree, which
> are to be reviewed (not only by UML folks), and possibly merged for the
> 2.6.13 cycle.
>
> The first one splits the i386 syscall table out of entry.S, without any real
> change for them (the file is included in the old place); if there are any
> syscall table changes, please be careful; this maybe means place *this* one
> before the others (or even merge now the i386 code movement part); I don't
> want that this patch is modified and that subtle bugs are introduced. The
> code movement is a really trivial code movement, however (no hidden changes).
>
> This is needed to enable us to include the i386 syscall table, so that we are
> sure that they match. I already handled the real differences between i386 and
> UML (see the patch).
I love this idea and it fits well to UML/s390, where syscall table already
lives in a separate source.
As change and grow of syscall tables is not really coordinated between different
arches, this change will simplify maintenance of UML-subarches a lot.
Bodo