Anaconda and Linux kernel 2.6.0-test9

[W. Michael Petullo]

Has anyone been able to compile anaconda 9.2 against 2.6.0-test9
kernel headers? I am running into kernel header-related problems.
I've seen that other software projects are facing similar difficulty
with 2.6. I imagine it has something to do with what code is protected
by #ifdef __KERNEL__'s as the kernel itself compiles fine. I get the
following errors:

[***@imp anaconda-9.2]$ make
for d in isys loader2 po stubs textw utils scripts bootdisk installclasses iw pixmaps isomd5sum command-stubs fonts; do make -C $d; [ $? = 0 ] || exit 1; done
make[1]: Entering directory `/home/mike/development/anaconda-9.2-cryptoroot/anaconda-9.2/isys'
for d in gzlib; do make -C $d; done
make[2]: Entering directory `/home/mike/development/anaconda-9.2-cryptoroot/anaconda-9.2/isys/gzlib'
make[2]: Nothing to be done for `all'.
make[2]: Leaving directory `/home/mike/development/anaconda-9.2-cryptoroot/anaconda-9.2/isys/gzlib'
cc -c -ffunction-sections -I/usr/include/python2.2 -I.. -Wall -Os -g -DHAVE_NFS -D_LARGEFILE_SOURCE=1 -D_LARGEFILE64_SOURCE=1 -D_FILE_OFFSET_BITS=64 -Werror -fPIC -o isys.lo isys.c
In file included from /usr/include/linux/blockgroup_lock.h:8,
from /usr/include/linux/ext2_fs_sb.h:19,
from /usr/include/linux/ext2_fs.h:20,
from isys.c:4:
/usr/include/linux/spinlock.h: In function `bit_spin_lock':
/usr/include/linux/spinlock.h:413: invalid type argument of `->'
/usr/include/linux/spinlock.h: In function `bit_spin_trylock':
/usr/include/linux/spinlock.h:436: invalid type argument of `->'
/usr/include/linux/spinlock.h: In function `bit_spin_unlock':
/usr/include/linux/spinlock.h:451: invalid type argument of `->'
/usr/include/linux/spinlock.h:451: `TIF_NEED_RESCHED' undeclared (first use in this function)
/usr/include/linux/spinlock.h:451: (Each undeclared identifier is reported only once
/usr/include/linux/spinlock.h:451: for each function it appears in.)
/usr/include/linux/spinlock.h: In function `bit_spin_is_locked':
/usr/include/linux/spinlock.h:462: invalid type argument of `->'
In file included from /usr/include/linux/ext2_fs_sb.h:20,
from /usr/include/linux/ext2_fs.h:20,
from isys.c:4:
/usr/include/linux/percpu_counter.h: In function `percpu_counter_mod':
/usr/include/linux/percpu_counter.h:75: invalid type argument of `->'
/usr/include/linux/percpu_counter.h:77: invalid type argument of `->'
/usr/include/linux/percpu_counter.h:77: `TIF_NEED_RESCHED' undeclared (first use in this function)
In file included from /usr/include/linux/ext2_fs.h:20,
from isys.c:4:
/usr/include/linux/ext2_fs_sb.h: At top level:
/usr/include/linux/ext2_fs_sb.h:40: parse error before "uid_t"
/usr/include/linux/ext2_fs_sb.h:48: parse error before "s_next_generation"
/usr/include/linux/ext2_fs_sb.h:50: parse error before '*' token
/usr/include/linux/ext2_fs_sb.h:55: parse error before '}' token
In file included from /usr/include/linux/ext3_fs.h:20,
from isys.c:5:
/usr/include/linux/ext3_fs_i.h:80: field `i_orphan' has incomplete type
/usr/include/linux/ext3_fs_i.h:97: parse error before "loff_t"
/usr/include/linux/ext3_fs_i.h:111: parse error before '}' token
In file included from /usr/include/linux/ext3_fs.h:21,
from isys.c:5:
/usr/include/linux/ext3_fs_sb.h:44: parse error before "uid_t"
/usr/include/linux/ext3_fs_sb.h:52: parse error before "s_next_generation"
/usr/include/linux/ext3_fs_sb.h:53: parse error before "s_hash_seed"
/usr/include/linux/ext3_fs_sb.h:55: parse error before '*' token
/usr/include/linux/ext3_fs_sb.h:71: parse error before '}' token
In file included from isys.c:5:
/usr/include/linux/ext3_fs.h:618: parse error before "u32"
/usr/include/linux/ext3_fs.h:621: parse error before '*' token
/usr/include/linux/ext3_fs.h:622: parse error before '}' token
In file included from /usr/include/linux/timer.h:5,
from /usr/include/linux/workqueue.h:8,
from /usr/include/linux/fb.h:5,
from isys.c:38:
/usr/include/linux/list.h:576:2: #warning "don't include kernel headers in userspace"
In file included from /usr/include/linux/workqueue.h:8,
from /usr/include/linux/fb.h:5,
from isys.c:38:
/usr/include/linux/timer.h:11: field `entry' has incomplete type
/usr/include/linux/timer.h:21: confused by earlier errors, bailing out
make[1]: *** [isys.lo] Error 1
make[1]: Leaving directory `/home/mike/development/anaconda-9.2-cryptoroot/anaconda-9.2/isys'
make: *** [subdirs] Error 1

[Jeremy Katz]

Post by W. Michael Petullo
Has anyone been able to compile anaconda 9.2 against 2.6.0-test9
kernel headers? I am running into kernel header-related problems.
I've seen that other software projects are facing similar difficulty
with 2.6. I imagine it has something to do with what code is protected
by #ifdef __KERNEL__'s as the kernel itself compiles fine. I get the

/usr/include/linux should be the sanitized headers that are in the
glibc-kernheaders package. Just using the ones from the kernel tree is
likely to be a bad idea.

And I guarantee that anaconda has a fair bit of work to go to work with
2.6... continuing to work on that some more later this afternoon I
hope.

Cheers,

Jeremy

[m***@flyn.org]

Post by Jeremy Katz

Post by W. Michael Petullo
Has anyone been able to compile anaconda 9.2 against 2.6.0-test9
kernel headers? I am running into kernel header-related problems.
I've seen that other software projects are facing similar difficulty
with 2.6. I imagine it has something to do with what code is protected
by #ifdef __KERNEL__'s as the kernel itself compiles fine. I get the

/usr/include/linux should be the sanitized headers that are in the
glibc-kernheaders package. Just using the ones from the kernel tree is
likely to be a bad idea.
And I guarantee that anaconda has a fair bit of work to go to work with
2.6... continuing to work on that some more later this afternoon I
hope.

Is there a glibc-kernheaders package created from 2.6 yet? I can not find
one. I understand that kernel developers have stated that use-space programs
should not include kernel headers but I don't understand who sanitizes them
for use in /usr/include/linux etc. Is this the work of the glibc team? Red
Hat?

[Alexandre Oliva]

Post by m***@flyn.org
Is there a glibc-kernheaders package created from 2.6 yet?

This would only be necessary if there had been any kernel ABI
changes. Do you know of any?

BTW, glibc-kernheaders is actually a misnomer: it exposes the kernel
ABI, and has nothing to do with glibc, other than the fact that glibc
uses it.

[m***@flyn.org]

Post by Alexandre Oliva

Post by m***@flyn.org
Is there a glibc-kernheaders package created from 2.6 yet?

This would only be necessary if there had been any kernel ABI
changes. Do you know of any?
BTW, glibc-kernheaders is actually a misnomer: it exposes the kernel
ABI, and has nothing to do with glibc, other than the fact that glibc
uses it.

Yes, I want to add encrypted root filesystem support to anaconda unsing
loopback encryption. In order to do this, I need 2.6 (or kerneli.org's 2.4
patch) because loop.h has cryptoloop-related additions. I suppose I could
just apply the relevant portions of the 2.4 cryptoloop patch to
glibc-kernheaders. I would like to see this cryptoloop stuff upstream,
though.

Hopefully I'm on the right track now, thanks.

--
Mike