xorg broken on Beaglebone black revision 300438

[Otacílio]

Em 23/05/2016 22:56, Ian Lepore escreveu:
> On Mon, 2016-05-23 at 21:05 -0300, Otacílio wrote:
>> Dears
>>
>> Yesterday I build a new  image with the revision 300438 to my
>> Beaglebone
>> black. I'm using the same package of scfb and configuration to xorg,
>> but, with this new revision, xcfe do not works. After open xfce the
>> mouse stucks on the top left corner and panels shows black belts and
>> squares on chars. I did a small modification on the kernel but is not
>> clear to me that this can cause this erroneous behavior.  The
>> modification, xorg.conf, and Xlog from revision r298522 (works) and
>> revision 300438 (don't works). If you agree I will open a bug.
>>
>> Index: sys/arm/arm/locore-v6.S
>> ===================================================================
>> --- sys/arm/arm/locore-v6.S    (revision 300438)
>> +++ sys/arm/arm/locore-v6.S    (working copy)
>> @@ -134,8 +134,8 @@
>>        bic    r7, #CPU_CONTROL_IC_ENABLE
>>        bic    r7, #CPU_CONTROL_BPRD_ENABLE
>>        bic    r7, #CPU_CONTROL_SW_ENABLE
>> -    orr    r7, #CPU_CONTROL_UNAL_ENABLE
>> -    orr    r7, #CPU_CONTROL_AFLT_ENABLE
>> +    bic    r7, #CPU_CONTROL_UNAL_ENABLE
>> +    bic    r7, #CPU_CONTROL_AFLT_ENABLE
>>        orr    r7, #CPU_CONTROL_VECRELOC
>>        mcr    CP15_SCTLR(r7)
>>        DSB
>> @@ -458,8 +458,8 @@
>>        bic    r0, #CPU_CONTROL_IC_ENABLE
>>        bic    r0, #CPU_CONTROL_BPRD_ENABLE
>>        bic    r0, #CPU_CONTROL_SW_ENABLE
>> -    orr    r0, #CPU_CONTROL_UNAL_ENABLE
>> -    orr    r0, #CPU_CONTROL_AFLT_ENABLE
>> +    bic    r0, #CPU_CONTROL_UNAL_ENABLE
>> +    bic    r0, #CPU_CONTROL_AFLT_ENABLE
>>        orr    r0, #CPU_CONTROL_VECRELOC
>>        mcr    CP15_SCTLR(r0)
>>        DSB
>>
> I cannot tell from your message whether you mean that this change
> (disabling alignment faults) did, or did not, fix the problem.
>
> I have been testing a change just like this today (for other reasons,
> nothing to do with beaglebone or X), and it's likely to be committed
> soon.
>
> -- Ian
>
Hello

This patch was a suggestion from Mikaël Urankar to try fix a segfault 
from webkit-gtk2. I think that this is not related with the problem but...


[]'s

-Otacílio