[oi-dev] Error from the vesa driver (Was: libdrm-2.4.75 is too old)
Gary Mills
gary_mills at fastmail.fm
Thu Jun 4 02:43:15 UTC 2020
On Mon, Jun 01, 2020 at 09:20:30AM -0500, Gary Mills wrote:
> I just put together a desktop system using this CPU:
>
> Intel Core i3-9100 Coffee Lake 4-Core 3.6 GHz (4.2 GHz Turbo)
> LGA 1151 (300 Series) 65W BX80684I39100 Desktop Processor
> Intel UHD Graphics 630
[...]
> The Xorg log (/var/log/Xorg.0.log) shows that the vesa driver has
> failed. Errors are prefixed with EE:
[...]
> [ 522.285] (II) Loading sub module "int10"
> [ 522.285] (II) LoadModule: "int10"
> [ 522.285] (II) Loading /usr/lib/xorg/modules/amd64/libint10.so
> [ 522.285] (II) Module int10: vendor="X.Org Foundation"
> [ 522.285] compiled for 1.19.7, module version = 1.0.0
> [ 522.285] ABI class: X.Org Video Driver, version 23.0
> [ 522.285] (II) VESA(0): initializing int10
> [ 522.285] (EE) VESA(0): V_BIOS address 0x90000 out of range
[...]
> [ 522.923] (EE) Server terminated with error (1). Closing log file.
>
> The Intel video device looks like this in `scanpci -v':
>
> pci bus 0x0000 cardnum 0x02 function 0x00: vendor 0x8086 device 0x3e91
> Intel Corporation 8th Gen Core Processor Gaussian Mixture Model
> CardVendor 0x1043 card 0x8694 (ASUSTeK Computer Inc., Card unknown)
> STATUS 0x0010 COMMAND 0x0007
> CLASS 0x03 0x00 0x00 REVISION 0x00
> BIST 0x00 HEADER 0x00 LATENCY 0x00 CACHE 0x10
> BASE0 0x00000000a0000000 SIZE 16777216 MEM64
> BASE2 0x0000000090000000 SIZE 268435456 MEM64 PREFETCHABLE
> BASE4 0x00004000 SIZE 64 I/O
> MAX_LAT 0x00 MIN_GNT 0x00 INT_PIN 0x01 INT_LINE 0xff
It's actually the int10 sub-module of the vesa driver that failed.
Xorg always tries two video drivers. The first is a vendor-specific
driver. If that driver rejects the video device, Xorg falls back to
the vesa driver. This driver should always work because it uses a
standard BIOS built-in to the video device. In this case, the vesa
driver fails, leaving no working driver, and of course no graphics.
As a test, I installed an old Nvidia card (GT-730) in the system.
Everything worked. I got the graphic desktop including the GUI
installer. In this case, the nvidia driver was able to support the
card. The vesa driver was not needed at all.
I'm guessing that the vesa driver does not handle 64-bit PCI-E
addresses correctly. Certainly 31-bit variables are too short. Does
anyone know if this problem has been fixed in a later version of the
int10 sub-module? Fixing this problem should make the vesa driver
work with the Intel built-in graphics device, or with other devices
that use long addresses.
--
-Gary Mills- -refurb- -Winnipeg, Manitoba, Canada-
More information about the oi-dev
mailing list