| 1 |
Hardave Riar wrote: |
| 2 |
> Jason Pepas wrote: |
| 3 |
> |
| 4 |
>> and the profile like so: |
| 5 |
>> |
| 6 |
>> make.profile -> /usr/portage/profiles/default-linux/mips/2005.0 |
| 7 |
> |
| 8 |
> |
| 9 |
> That's the wrong profile for uclibc stages, this will try to replace |
| 10 |
> uclibc with glibc and that's a bad thing. You should probablly be using |
| 11 |
> /usr/portage/profile/uclibc/mips. Using the uclibc stages it should not |
| 12 |
> try to build glibc like it is below. |
| 13 |
> |
| 14 |
>> any ideas? I see the mips project website was just updated today, and I |
| 15 |
>> just realized last night that the cobalt images are little endian (the |
| 16 |
>> asus wl500g is little endian), so I'll try chrooting into one of those |
| 17 |
>> images next. |
| 18 |
> |
| 19 |
> |
| 20 |
> Don't bother using the cobalt stages, the wl500g has no where near |
| 21 |
> enough memory to run a glibc system. |
| 22 |
|
| 23 |
Not to mention, the Cobalt stages are MIPS IV, and therefore includes |
| 24 |
some 64-bit instructions, the WL500g will be MIPS32, and does not |
| 25 |
understand 64-bit instructions. |
| 26 |
|
| 27 |
So you can expect illegal instructions left, right & centre. |
| 28 |
|
| 29 |
As Stephen has pointed out too, messing with CHOST is asking for it... I |
| 30 |
speak from experience. :-) |
| 31 |
(tried it on my x86-based laptop ... boy did I have fun finding the |
| 32 |
pieces and putting them back together again :-/) |
| 33 |
-- |
| 34 |
____ _ Stuart Longland (a.k.a Redhatter) |
| 35 |
/ _ \ ___ ___ __| |__ __ __ Gentoo Linux/MIPS Cobalt and Docs |
| 36 |
- (_) \ / \ ; \(__ __)/ \ / \ Developer |
| 37 |
\ // O _| / /\ \ | | | /\ | /\ | |
| 38 |
/ / \ /__| / \ \ | | | \/ | \/ | |
| 39 |
(___/ \____/|_; |_| \_/ \__/ \__/ http://dev.gentoo.org/~redhatter |
Archives