| 1 |
On 25/12/2014 22:24, Jean-Marc Beaune wrote: |
| 2 |
> Hi, |
| 3 |
> |
| 4 |
> The boot process issue has been solved by setting up PATA properly in |
| 5 |
> the kernel config. |
| 6 |
> |
| 7 |
|
| 8 |
Just watch out, I *think* this will bite you again when you upgrade to a |
| 9 |
newer kernel? It's *NOT* the PATA which is the problem, but some |
| 10 |
bug/interaction between GPIO setup and the pata driver. The key is |
| 11 |
actually to disable GPIO-Keys (from memory) which seems to break the |
| 12 |
pata module. I think it only became a problem somewhere in the early |
| 13 |
3.x ish era where a platform module was added to emulate the Alix button |
| 14 |
as a "button". |
| 15 |
|
| 16 |
Anyway, I don't recall the details now, but like you I briefly fixed it |
| 17 |
by fiddling with PATA modules before I realised it wasn't really the |
| 18 |
issue... I speculate it's something to do with the polling of the GPIOs |
| 19 |
which is leaving the structure in some state which breaks the PATA |
| 20 |
module...? |
| 21 |
|
| 22 |
I haven't chased down details though. I concede I have read the |
| 23 |
datasheets on the CS533x and mostly understand the basics of the |
| 24 |
interrupt setup, but I don't really understand enough about how linux |
| 25 |
deals with interrupts to resolve this.. My goal was to get GPIO |
| 26 |
interrupts working in order to get a GPS 1PPS working... I think I can |
| 27 |
concede I have mostly lost interest though and it's not going to get |
| 28 |
finished... We do at least have platform setup modules for the various |
| 29 |
architectures though, so it's a start... |
| 30 |
|
| 31 |
Good luck! |
| 32 |
|
| 33 |
Ed W |
Archives