Radeon driver with FC20 latest kernel update

Tue Jan 20 14:46:31 UTC 2015

On Tuesday 20 January 2015 14:31:09 Andrew R Paterson wrote:
> On Tuesday 20 January 2015 15:10:15 poma wrote:
> > On 20.01.2015 14:44, Andrew R Paterson wrote:
> > > On Tuesday 20 January 2015 14:31:32 poma wrote:
> > >> On 20.01.2015 12:52, Andrew R Paterson wrote:
> > >>> Hi,
> > >>> I am running fedora 20 X86_64
> > >>> 
> > >>> I have an Intel Processor:
> > >>> Intel(R) Core(TM)2 Quad CPU Q6700 @ 2.66GHz.
> > >>> 
> > >>> I have an AMD radeon based GPU card :
> > >>> Advanced Micro Devices, Inc. [AMD/ATI] RV535 [Radeon X1650 PRO]
> > >>> (Secondary)
> > >>> (rev 9e)
> > >>> 
> > >>> I have two monitors :
> > >>> ACER Technolgies – on DVI-0 (1920x1080)
> > >>> Viewsonic Corporation on DVI-1 (1920x1200)
> > >>> 
> > >>> They are (were!) driven by the radeon module :
> > >>> 
> > >>> lsmod :-
> > >>> .....
> > >>> radeon 1496040 2
> > >>> .....
> > >>> i2c_algo_bit 13257 1 radeon
> > >>> .....
> > >>> drm_kms_helper 	89555 1 radeon
> > >>> ttm 				80807 1 radeon
> > >>> .....
> > >>> drm 				296901 5 ttm,drm_kms_helper,radeon
> > >>> 
> > >>> 
> > >>> /var/log/messages shows:
> > >>> 
> > >>> radeon 0000:01:00.0: VRAM: 512M 0x0000000000000000 -
> > >>> 0x000000001FFFFFFF
> > >>> (512M used)
> > >>> radeon 0000:01:00.0: GTT: 512M 0x0000000020000000 - 0x000000003FFFFFFF
> > >>> [drm] Detected VRAM RAM=512M, BAR=256M
> > >>> [drm] RAM width 128bits DDR
> > >>> 
> > >>> X server log shows:
> > >>> 
> > >>> [ 88.037] (II) Loader magic: 0x80dd00
> > >>> [ 88.037] (II) Module ABI versions:
> > >>> [ 88.037] X.Org ANSI C Emulation: 0.4
> > >>> [ 88.037] X.Org Video Driver: 14.1
> > >>> [ 88.037] X.Org XInput driver : 19.2
> > >>> [ 88.037] X.Org Server Extension : 7.0
> > >>> [ 88.037] (II) xfree86: Adding drm device (/dev/dri/card0)
> > >>> [ 88.040] (--) PCI:*(0:1:0:0) 1002:71c1:174b:0840 rev 158, Mem @
> > >>> 0xd0000000/268435456, 0xfdaf0000/65536, I/O @ 0x0000ae00/256, BIOS @
> > >>> 0x????????/131072
> > >>> [ 88.040] (--) PCI: (0:1:0:1) 1002:71e1:174b:0841 rev 158, Mem @
> > >>> 0xfdae0000/65536, BIOS @ 0x????????/65536
> > >>> ....
> > >>> ....
> > >>> [ 88.058] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
> > >>> [ 88.058] (II) FBDEV: driver for framebuffer: fbdev
> > >>> [ 88.058] (II) VESA: driver for VESA chipsets: vesa
> > >>> [ 88.058] (++) using VT number 1
> > >>> .....
> > >>> .....
> > >>> [ 88.059] (II) [KMS] Kernel modesetting enabled.
> > >>> [ 88.059] (WW) Falling back to old probe method for modesetting
> > >>> [ 88.059] (WW) Falling back to old probe method for fbdev
> > >>> [ 88.059] (II) Loading sub module "fbdevhw"
> > >>> [ 88.059] (II) LoadModule: "fbdevhw"
> > >>> [ 88.059] (II) Loading /usr/lib64/xorg/modules/libfbdevhw.so
> > >>> [ 88.060] (II) Module fbdevhw: vendor="X.Org Foundation"
> > >>> [ 88.060] compiled for 1.14.4, module version = 0.0.2
> > >>> [ 88.060] ABI class: X.Org Video Driver, version 14.1
> > >>> [ 88.060] (WW) Falling back to old probe method for vesa
> > >>> [ 88.060] (II) RADEON(0): Creating default Display subsection in
> > >>> Screen
> > >>> section
> > >>> "Default Screen Section" for depth/fbbpp 24/32
> > >>> [ 88.060] (==) RADEON(0): Depth 24, (--) framebuffer bpp 32
> > >>> [ 88.060] (II) RADEON(0): Pixel depth = 24 bits stored in 4 bytes (32
> > >>> bpp
> > >>> pixmaps)
> > >>> [ 88.060] (==) RADEON(0): Default visual is TrueColor
> > >>> [ 88.060] (==) RADEON(0): RGB weight 888
> > >>> [ 88.060] (II) RADEON(0): Using 8 bits per RGB (8 bit DAC)
> > >>> [ 88.060] (--) RADEON(0): Chipset: "ATI Radeon X1650" (ChipID =
> > >>> 0x71c1)
> > >>> .....
> > >>> ________________________________________________
> > >>> All worked well up to kernel 3.17.7-200.fc20.x86_64.
> > >>> ________________________________________________
> > >>> 
> > >>> The recent update to kernel 3.17.8-200.fc20.x86_64
> > >>> results in radeon.ko failing to load.
> > >>> 
> > >>> /var/log/messages shows:
> > >>> 
> > >>> [drm:radeon_init] *ERROR* No UMS support in radeon module!
> > >>> 
> > >>> %modrpobe radeon gives same message.
> > >>> 
> > >>> My monitors are then in low res mirorred mode (ugh!)
> > >>> 
> > >>> X server log shows:
> > >>> ....
> > >>> [ 13345.844] (II) Loader magic: 0x80dd00
> > >>> [ 13345.844] (II) Module ABI versions:
> > >>> [ 13345.844] X.Org ANSI C Emulation: 0.4
> > >>> [ 13345.844] X.Org Video Driver: 14.1
> > >>> [ 13345.844] X.Org XInput driver : 19.2
> > >>> [ 13345.844] X.Org Server Extension : 7.0
> > >>> [ 13345.844] (II) xfree86: Adding drm device (/dev/dri/card0)
> > >>> [ 13345.844] setversion 1.4 failed
> > >>> [ 13345.848] (--) PCI:*(0:1:0:0) 1002:71c1:174b:0840 rev 158, Mem @
> > >>> 0xd0000000/268435456, 0xfdaf0000/65536, I/O @ 0x0000ae00/256, BIOS @
> > >>> 0x????????/131072
> > >>> [ 13345.848] (--) PCI: (0:1:0:1) 1002:71e1:174b:0841 rev 158, Mem @
> > >>> 0xfdae0000/65536, BIOS @ 0x????????/65536
> > >>> .......
> > >>> 
> > >>> I understand UMS is User ModeSetting – which according to a quick
> > >>> browse
> > >>> of
> > >>> the source for the radeon module means CONFIG_DRM_RADEON_UMS is not
> > >>> set.
> > >>> Then I came across comments to the effect that this has been
> > >>> deprecated
> > >>> (a
> > >>> mysterious word akin to legal-speak!) presumably replaced by KMS.
> > >>> 
> > >>> Has anybody got any clues how I can get the radeon driver loading and
> > >>> working again?
> > >>> Or even if its a bug whether it will soon be fixed?
> > >>> 
> > >>> Many Thanks
> > >>> 
> > >>> Andy
> > >> 
> > >> $ grep modeset /etc/modprobe.d/* /proc/cmdline
> > > 
> > > Hi poma,
> > > 
> > > grep modeset /etc/modprobe.d/* =
> > > /etc/modprobe.d/radeon-kms.conf:options radeon modeset=0
> > > 
> > > grep modeset /proc/cmdline =
> > > nothing!
> > > 
> > > Thanks
> > 
> > # sed -i s/0/1/ /etc/modprobe.d/radeon-kms.conf
> 
> Hi poma
> 
> that did it thanks
> 
> /etc/modprobe.d/radeon-kms.conf:options radeon modeset=1
> 
> I mind you I still get the initial:
>  [drm:radeon_init] *ERROR* No UMS support in radeon module!
> And the plymouth splash screen goes from one mode to another and ain't what
> it used to be!) - but that doesn't matter. On loading, the kernel sorts
> things out.
> 
> A assume if I now reboot into the previous kernel, that wont work - I'll
> just do it out of curiosity - but no probs - I'll now stick with
> 3.17.8-200.fc20.x86_64.
> 
> Thanks
> Andy
rebooted into previous kernel - no probs!
I assume thats because previous kernel supported both ums and kms????
Regards
Andy