Maybe its just a driver which can't find a blob - by itself the GPU should have this blob loaded on it by coreboot. https://bugs.freedesktop.org/show_bug.cgi?id=101473 Alex Deucher says that "Invalid PCI ROM header signature: expecting 0xaa55, got 0x0000" message can be ignored
BTW this "0x0000" issue is not G505S specific, seems T400s also has it - https://lists.gnu.org/archive/html/libreboot/2016-05/msg00050.html There aren't many coreboot supported laptops with dual GPU perhaps thats why its not fixed yet Also, maybe it could be possible to somehow pass a copy of that vgabios blob from GRUB to a Linux kernel, so that its' driver detects this blob. If i'm not mistaken i've seen the people trying it for their Mac Pro (similar problem but got another value, maybe 0xffff), forgot how its ended though The research about GPUs is not over, we are going to continue it soon ;) 2017-12-30 2:25 GMT+03:00 awokd <[email protected]>: > On Fri, December 29, 2017 10:50 pm, Ivan Ivanov wrote: >> awokd, Thank you very much for your microcode research ;) >> >> Please remind me, which version of G505S do you have: >> integrated /// integrated + HD 8570M /// integrated + R5 M230 ? > > Mine's the integrated. Blooorp turned out to have the integrated + R5 > M230. We managed to get his working with Qubes 4.0 by adding > xen-pciback.hide(04:00.0) to the dom0 command line in GRUB so it wouldn't > attempt to initialize the M230 at all. That's why his kept crashing. > >> Last time we experimented with Mike, we were 100% sure we extracted >> vgabios'es correctly, in the best possible way, but - maybe because of >> unstable AMDGPU at Linux - >> we couldn't get our discrete GPUs working at this time, no matter what we >> tried :P > >> If you have integrated + one of the discretes: >> do you have a discrete GPU working when you are using two vgabios'es ? If >> yes, at what Linux distro and what kernel, your software setup? >> >> I'm going to re-test it soon with a more stable kernel, probably >> something like 4.14.8 at Void Linux: excellent not-systemd distro which is >> stable while having very modern packages >> >> Not Qubes OS, because at Qubes its very difficult to test any GPU, >> this OS isn't for gaming ;-) > > We troubleshot his in a similar way but with Stretch. If you load the > firmware-amd-graphics package it gave us some helpful output. In this case > it seemed like it was missing the blob (signature 0x0000 instead of what > it wanted) so we just hid it. I'm curious too to see if it would work > properly with one but unfortunately can't try on mine! > > -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/CAAaskFAkZUtREMzkvXX%3DL0yse02bXype1mdmzzHCQdQAytxb0g%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
