Rawhide install success - NOT?

John Ellson ellson at research.att.com
Sat Nov 19 15:41:18 UTC 2005


G.Wolfe Woodbury wrote:
> On Sat, Nov 19, 2005 at 12:36:47AM -0500, John Ellson wrote:
>   
>> John Ellson wrote:
>>     
>>> G.Wolfe Woodbury wrote:
>>>       
>>>> On Fri, Nov 18, 2005 at 11:25:37AM -0800, MJang wrote:
>>>>
>>>>         
>>>>> Folks,
>>>>>
>>>>> Just reporting success with an install from today's (11-18) Rawhide on
>>>>> my HP Laptop (ze4500 series). Presumably with the freeze in effect,
>>>>> today's build is essentially the same as what we'll see from FC5 
>>>>> Test 1.
>>>>> So if you want a pre-preview.....
>>>>>
>>>>> Thanks,
>>>>> Mike
>>>>>    
>>>>>           
>>>> On the other hand....
>>>> Some module being probed during hardware initialization is munging my
>>>> video display.  This has happened with the Trident Cyberblade/i1 and an
>>>> ATI Rage XL card.
>>>>
>>>> The screen comes back from reset with vertical lines and bits of text
>>>> (not even whole letters!) between the lines.  I've moved the tridentfb
>>>> module out of the way, and that implies that it's some other module that
>>>> is the culprit.  I'm off the hunIn my case I probl.
>>>>  
>>>>         
>>> Any chance it could be hal, perhaps on an Athlon?
>>>
>>> I upgraded five i686 boxes and one Athlon yesterday, and only the 
>>> Athlon had problems.
>>> I couldn't get X to start.  I tried three different graphics cards
>>> without success.  Eventually I downgraded hal and that fixed it.
>>>
>>> I don't recall why I suspected hal.  Something in /var/log/messages I 
>>> think.
>>>
>>> John
>>>
>>>       
>> It wasn't hal I downgraded, it was udev.  From udev-075-2 to udev-075-1.
>>
>> Sorry for the false accusations against hal.
>>     
>
> OP Here,  No, it is a Coppermine Celeron, it is a VIA chipset.  I've
> used two video sets (Trident Cyberblade/i1(agp)  and  ATI Rage XL(pci))
> to make sure it wasn't the video card _per_se_.  Can someone point me at
> a copy of Udev-075-1?
>
> --Wolfe
>
>   
I did more testing and generated bug #173720

The problem seems to be that udev-075-2 is imcompatible with older kernels.
I don't know where the transition is, but it works with
kernel-2.6.14-1.1688_FC5 but not with kernel-2.6.14-1.1632_FC5

1632 was the last kernel with working v4l drivers for tvtime so I'm 
sticking with
kernel-2.6.14-1.1632_FC5 and udev-075-1
Everything else is current, including modular X.

John




More information about the test mailing list