Hi Joseph, please provide the output of lspci -vvvv when the driver is not loaded as an attachment. Additional to that a full dmesg as a text file attachment with the crash would be helpful, you should still be able to ssh into the system even when the driver crashed. That looks strongly like the BIOS didn't assigned a PCI BAR to the GPUs. The drivers crashes because we assume that the BIOS is working correctly and assigned PCI BARs to all GPUs. Another possibility is that something went wrong while resizing the BARs. In this case it would be a driver/PCI subsystem problem. Regards, Christian. Am 15.02.2018 um 23:26 schrieb Joseph Wang: > I'm attempting to use amdgpu with the B250 mining expert motherboard > for GPU mining of ethereum. The MB can handle 19 cards. When I have > more than six cards, X11 stops working but I can still handle text > console. At nine cards, I get the attached kernel oops on boot. > > This is running with the drm-next-4.17-wip branch. > > Also would love to get this working with rocm > > > _______________________________________________ > amd-gfx mailing list > amd-gfx at lists.freedesktop.org > https://lists.freedesktop.org/mailman/listinfo/amd-gfx -------------- next part -------------- An HTML attachment was scrubbed... URL: <https://lists.freedesktop.org/archives/amd-gfx/attachments/20180216/af64d383/attachment.html>
