On 13.08.2015 08:36, Lucas Stach wrote:
Am Donnerstag, den 13.08.2015, 15:18 +0900 schrieb Michel Dänzer:
On 13.08.2015 15:03, Lucas Stach wrote:
Hi Christian,
this commit is causing a boot regression with v4.2-rcX on my Richland APU (CHIP_ARUBA) based laptop. I didn't have time yet to track down where exactly it is going wrong, but I bisected it down to this single commit.
I don't have the VCE firmware installed on this system, so from a quick look at the code I would expect it to drop out pretty early and just leave the VCE unconfigured, but otherwise keep things working as before. This is unfortunately not the case.
If the radeon driver is built into the kernel (or loaded from the initrd?), the attempt to load the firmware might take a long time to time out.
Gah. Thanks, I was too impatient to wait for the firmware loading to time out. In fact this is a standard Fedora kernel config, so radeon is a module, but it is built into the initrd.
So it's not really readeons fault, but one more iteration of the fact that anything involving firmware loading is just horribly inconvenient. Especially if it's firmware for an optional component.
Yeah, exactly. The timeout for loading the firmware is really long on both the standard Fedora as well as Ubuntu kernels. Not sure if that's the default or just their configuration.
While it doesn't have the highest priority for us I usually still do tests if working without firmware still works once or twice during the upstreaming process. So if you really find that the box doesn't work without the firmware leave me a note and I will fix it.
Best regards, Christian.
Regards, Lucas