Gregoire Gentil wrote:
Philip, I found the patch. Thanks. SPI3 is working for me too but I think that there are a couple of errors: - first, in the patch you posted on the beagleboard mailing list, you don't setup CS0 and CS1 pins in u-boot. I think that you should do it.
Yeah. I wanted to get the info out quickly since there are a couple of longish threads on the Beagle list over the problem.
- secondly, you have added more mux configuration in the kernel for SPI3 that should not be SPI3 but those new ones are wrong as they are competing with some USB pins. It's the same error as David pointed you for MMC2. Nevertheless, it's still working. Why? Because I have now a strong feeling that mux configuration is not working in the kernel (at least for the beagleboard). Here are a few facts that would confirm this statement:
I thought I #if 0 ... #endif the kernel mux code? I'll double check. I should try it with un patched u-boot and only the specific pins set in the kernel.
I'm juggling too many different things atm :( I also changed the PIN MUX config option in the flag per another question. Philip
- MUX setup for USB ehci has never worked in the kernel. It's why the beagleaboard rev-C ehci patch has been transfered to u-boot. - the difference between your patch before and after it was working, is really the u-boot configuration. You haven't really changed anything in the kernel (especially in the spi driver) and as mentioned above, you have even introduced some competing muxes that should have created more trouble if the kernel mux config were working correctly. - I had two other areas where I configured the pins in kernel and it was not working. Only when I eventually did it in u-boot, it started to work. I don't know what's wrong with the pin configuration in the kernel, GrégoireOn Thu, 2009-02-19 at 09:14 -0500, Philip Balister wrote:Gregoire Gentil wrote:I posted it to the Beagle group. Let me know if you are having trouble finding it.Philip, Can you please post here or on the Beagleboard mailing list the u-boot patch? This muxpin is very tricky and I have experienced many problems when set up in the kernel while it seems to work better from u-boot - don't know why,If we come up with a better config for the expansion port, we'll clean it up and submit here. My gut feeling is having SPI interfaces on the expansion connector will be more useful then the MMC interface.Philip-- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature