[OpenAFS-devel] Re: 2.6.16 linux and d_cache

[Marc Dionne]

Hi Bill,

I managed to compile and run the current CVS head on a 2.6.15-1.1977=
=20
kernel (FC5), modulo the following adjustments:

1 - Patch to acinclude.m4 that I posted earlier, to enable the=20
STRUCT_INODE_HAS_I_MUTEX flag
2 - Replacing AFSTOI by AFSTOV, and ITOAFS by VTOAFS in afs_callback.=
c.=20
(like you, don't now where these are supposed to be defined)
3 - Changing a "register int" declaration to "int" in ubik/vote.c to=
=20
work around a compiler crash
4 - Finally, compiling a new kernel with this patch:=20
http://www.ussg.iu.edu/hypermail/linux/kernel/0602.2/0936.html

I haven't used it extensively, but so far it seems functional.

Cheers,
Marc

William John Murray wrote:
> On Fri, 2006-02-24 at 00:45 +0100, William John Murray wrote:
>  =20
>>   Hello all,
>>          I am still trying to get AFS on 2.6.16 linux.
>> The latest snapshots, and 1.5.0 have a new routine in afs_callback=
.c
>> which does not compile:
>> 'struct dentry=E2=80=99 has no member named =E2=80=98d_child'
>>
>>    =20
>
> Thanks Marc...this fixes the d_child.
> There is another oddity. The routine uses AFSTOI and ITOAFS,
> but these don't seem to exist. I THINK AFSTOV and VTOAFS replace th=
em I
> think? But how come no one else has trouble with them?
>
>    Anyway, this brings me back to my problem of a couple of weeks a=
go:
>
> libafs: falsely claims to have parameter sys_call_table_addr
>
> on inserting the library...I still have no idea what to do with tha=
t.
>     Bill