You are not logged in.

Pages: 1

#1 2015-09-06 13:26:06

niam
Member
Registered: 2015-09-06
Posts: 13

Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Hello ALL.
Please help to resolve (actually would like to update kernel) the following situation:

After upgrading from 3.19.2-pf to 4.1.4-pf or linux 4.2-3 I've got during booting via grube the following message on the screen: radeon 0000:01:00.0 Fatal error during GPU init
and no X (xf86-video-ati) and framebuffer in console. 
With 3.19.2-pf everything is FINE!

Could someone help what was changed in 4.1.x  kernel and what should be done additionally to make it working?

Some additional diagnostic information:
kernels details that were used

WORKING: 
uname -a
Linux 3.19.2-pf #1 SMP PREEMPT Sat Mar 14 20:31:07 EET 2015 x86_64 GNU/Linux

NOT WORKING: (for both kernels the same story = no radeondrmfb)
file -s /boot/vmlinuz-linux
/boot/vmlinuz-linux: Linux kernel x86 boot executable bzImage, version 4.2.0-3-ARCH (builduser@foutrelis) #1 SMP PREEMPT Fri Sep 4 20:, RO-rootFS, swap_dev 0x4, Normal VGA


file -s /boot/vmlinuz-linux-pf 
/boot/vmlinuz-linux-pf: Linux kernel x86 boot executable bzImage, version 4.1.4-pf (hzy@arch) #1 SMP PREEMPT Tue Sep 1 21:49:23 CST 2015, RO-rootFS, swap_dev 0x4, Normal VGA
for the case of 3.19.2:   (if 4.1.4 pf = kernel is 4.1.4)

pacman  -Qv | grep linux
archlinux-keyring 20150605-1
archlinuxcn-keyring 20150516-1
lib32-util-linux 2.26.2-1
libutil-linux 2.26.2-1
linux 4.2-3
linux-api-headers 4.1.4-1
linux-firmware 20150904.6ebf5d5-1
linux-headers 4.2-3
linux-pf 3.19.2-1
linux-pf-headers 3.19.2-1
syslinux 6.03-3
util-linux 2.26.2-1
cat journalctlb  | grep -Ei 'radeon|drm'
вер 06 16:00:59 h4os kernel: [drm] Initialized drm 1.1.0 20060810
вер 06 16:00:59 h4os kernel: [drm] radeon kernel modesetting enabled.
вер 06 16:00:59 h4os kernel: [drm] initializing kernel modesetting (RV610 0x1002:0x94C9 0x1043:0x15B2).
вер 06 16:00:59 h4os kernel: [drm] register mmio base: 0xFDEF0000
вер 06 16:00:59 h4os kernel: [drm] register mmio size: 65536
вер 06 16:00:59 h4os kernel: radeon 0000:01:00.0: VRAM: 256M 0x0000000000000000 - 0x000000000FFFFFFF (256M used)
вер 06 16:00:59 h4os kernel: radeon 0000:01:00.0: GTT: 512M 0x0000000010000000 - 0x000000002FFFFFFF
вер 06 16:00:59 h4os kernel: [drm] Detected VRAM RAM=256M, BAR=0M
вер 06 16:00:59 h4os kernel: [drm] RAM width 64bits DDR
вер 06 16:00:59 h4os kernel: radeon 0000:01:00.0: Fatal error during GPU init
вер 06 16:00:59 h4os kernel: [drm] radeon: finishing device.
вер 06 16:00:59 h4os kernel: [drm] radeon: ttm finalized
вер 06 16:00:59 h4os kernel: radeon: probe of 0000:01:00.0 failed with error -12
lspci -v | grep -A 12 VGA
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV610/M72-S [Mobility Radeon HD 2400] (prog-if 00 [VGA controller])
	Subsystem: ASUSTeK Computer Inc. Device 15b2
	Physical Slot: 16
	Flags: bus master, fast devsel, latency 0, IRQ 31
	Memory at c0000000 (32-bit, prefetchable) [size=256M]
	I/O ports at 9000 [size=256]
	Memory at fdef0000 (32-bit, non-prefetchable) [size=64K]
	Expansion ROM at fdec0000 [disabled] [size=128K]
	Capabilities: [50] Power Management version 3
	Capabilities: [58] Express Legacy Endpoint, MSI 00
	Capabilities: [a0] MSI: Enable+ Count=1/1 Maskable- 64bit+
	Capabilities: [100] Vendor Specific Information: ID=0001 Rev=1 Len=010 <?>
	Kernel driver in use: radeon
cat dmesg  | grep -Ei 'radeon|drm'
[    0.683887] [drm] Initialized drm 1.1.0 20060810
[    0.692813] [drm] radeon kernel modesetting enabled.
[    0.694141] [drm] initializing kernel modesetting (RV610 0x1002:0x94C9 0x1043:0x15B2).
[    0.694156] [drm] register mmio base: 0xFDEF0000
[    0.694157] [drm] register mmio size: 65536
[    0.694276] radeon 0000:01:00.0: VRAM: 256M 0x0000000000000000 - 0x000000000FFFFFFF (256M used)
[    0.694279] radeon 0000:01:00.0: GTT: 512M 0x0000000010000000 - 0x000000002FFFFFFF
[    0.694281] [drm] Detected VRAM RAM=256M, BAR=0M
[    0.694282] [drm] RAM width 64bits DDR
[    0.694356] radeon 0000:01:00.0: Fatal error during GPU init
[    0.694918] [drm] radeon: finishing device.
[    0.702587] [drm] radeon: ttm finalized
[    0.702776] radeon: probe of 0000:01:00.0 failed with error -12
cat Xorg.0.log  | grep -Ei 'EE|WW|II'
[   331.479] Current Operating System: Linux h4os 4.1.4-pf #1 SMP PREEMPT Tue Sep 1 21:49:23 CST 2015 x86_64
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[   331.480] (==) No Layout section.  Using the first Screen section.
[   331.480] (==) No screen section available. Using defaults.
[   331.480] (**) |-->Screen "Default Screen Section" (0)
[   331.480] (==) No device specified for screen "Default Screen Section".
[   331.480] (==) No monitor specified for screen "Default Screen Section".
[   331.480] (II) The server relies on udev to provide the list of input devices.
[   331.480] (II) Loader magic: 0x817d60
[   331.480] (II) Module ABI versions:
[   331.482] (II) systemd-logind: took control of session /org/freedesktop/login1/session/c22
[   331.484] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[   331.484] (II) LoadModule: "glx"
[   331.485] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[   331.486] (II) Module glx: vendor="X.Org Foundation"
[   331.487] (II) LoadModule: "radeon"
[   331.487] (II) Loading /usr/lib/xorg/modules/drivers/radeon_drv.so
[   331.487] (II) Module radeon: vendor="X.Org Foundation"
[   331.487] (II) RADEON: Driver for ATI Radeon chipsets:
	KAVERI, HAWAII, HAWAII, HAWAII, HAWAII, HAWAII, HAWAII, HAWAII,
	HAWAII, HAWAII, HAWAII, HAWAII, HAWAII
[   331.495] (--) controlling tty is VT number 1, auto-enabling KeepTty
[   331.495] (II) [KMS] drm report modesetting isn't supported.
[   331.495] (EE) Screen 0 deleted because of no matching config section.
[   331.495] (II) UnloadModule: "radeon"
[   331.495] (EE) Device(s) detected, but none match those in the config file.
[   331.495] (EE) 
[   331.495] (EE) no screens found(EE) 
[   331.495] (EE) 
[   331.495] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   331.495] (EE) 
[   331.760] (EE) Server terminated with error (1). Closing log file.

Last edited by niam (2015-09-06 17:46:03)

Offline

#2 2015-09-06 16:03:10

ewaller
Administrator
From: Pasadena, CA
Registered: 2009-07-13
Posts: 19,796

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Post the output of uname -a

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

Offline

#3 2015-09-06 17:53:24

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

ewaller wrote:

Post the output of uname -a

added to the original post.
Actually my preconclusion - there is no init in kernel radeondrmfb.
Probably like it is described in Gentoo - conflict between vesafb and radeondrmfb.
BUT the strange thing is that there is such sitation on stock arch kernel (not pf) too.

Offline

#4 2015-09-06 18:07:46

ewaller
Administrator
From: Pasadena, CA
Registered: 2009-07-13
Posts: 19,796

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Based on my limited interaction with you, you seem to know what you are doing.   I have no access to Radeon hardware, so I cannot be of much help beyond the generic.

When you boot from different kernels, I assume that all of the appropriate kernel modules are in  place under /lib/modules for the various kernels.
Also, you have some method of ensuring each kernel gets the correct initrd during boot?

Nothing is too wonderful to be true, if it be consistent with the laws of nature -- Michael Faraday
Sometimes it is the people no one can imagine anything of who do the things no one can imagine. -- Alan Turing
---
How to Ask Questions the Smart Way

Offline

#5 2015-09-07 07:40:04

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Yes initramfs is correct for each kernel and all of them have radeon.ko module.

Strange thing is that after simple pacman -Syy && pacman -Suy I've got situation when no X starting due described above problem. This is the reason why I am reporting here. My researching brings me to the kernel problems. I will try to recompile my own kernel but this is not normal for standard kernel!

Offline

#6 2015-09-07 11:16:26

Lone_Wolf
Forum Moderator
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,922

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Please post the kernel command lines you use .
Some details about your bootloader and a FULL dmesg / journalctl log from a failed boot may also help.

For clarification : when you tried arch kernel 4.2-3 , you had whole testing repo enabled and it was first repo in pacman.conf ?

Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#7 2015-09-07 13:06:44

R00KIE
Forum Fellow
From: Between a computer and a chair
Registered: 2008-09-14
Posts: 4,734

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

I'm using the same card here and generally I have no problems(1). Like Lone_Wolf says, post the full dmesg, and xorg log as it may provide more clues.

(1)
Are you using a compositing window manager? I'm asking because I can make things hang under certain conditions. I can manage to make the card hang if I use a compositing manager (in my case XFCE's compositor) and if I'm using glamor either by specifying it on a xorg .conf file or if using the modesetting driver (which uses glamor).

Sometimes things go a bit wonky when watching videos with mplayer using vdpau but I have not seen the machine/card hang when this happens.

My card:

# lspci -v -s 01:00.0
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RV610/M72-S [Mobility Radeon HD 2400] (prog-if 00 [VGA controller])
	Subsystem: Packard Bell B.V. Device c109
	Flags: bus master, fast devsel, latency 0, IRQ 28
	Memory at d0000000 (32-bit, prefetchable) [size=128M]
	I/O ports at 8800 [size=256]
	Memory at fd8f0000 (32-bit, non-prefetchable) [size=64K]
	Expansion ROM at fd8c0000 [disabled] [size=128K]
	Capabilities: [50] Power Management version 3
	Capabilities: [58] Express Legacy Endpoint, MSI 00
	Capabilities: [a0] MSI: Enable+ Count=1/1 Maskable- 64bit+
	Capabilities: [100] Vendor Specific Information: ID=0001 Rev=1 Len=010 <?>
	Kernel driver in use: radeon
	Kernel modules: radeon

R00KIE
Tm90aGluZyB0byBzZWUgaGVyZSwgbW92ZSBhbG9uZy4K

Offline

#8 2015-09-07 16:02:15

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Lone_Wolf wrote:

Please post the kernel command lines you use .
Some details about your bootloader and a FULL dmesg / journalctl log from a failed boot may also help.

For clarification : when you tried arch kernel 4.2-3 , you had whole testing repo enabled and it was first repo in pacman.conf ?

Yes, when I  tried arch kernel 4.2-3 , I had whole testing repo enabled and it was first repo in pacman.conf.

For kernel command line -nothing special, all that grub automatically generates.
cat /proc/cmdline
BOOT_IMAGE=/boot/vmlinuz-linux-pf root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw quiet

cat /boot/grub/grub.cfg
#
# DO NOT EDIT THIS FILE
#
# It is automatically generated by grub-mkconfig using templates
# from /etc/grub.d and settings from /etc/default/grub
#

### BEGIN /etc/grub.d/00_header ###
insmod part_gpt
insmod part_msdos
if [ -s $prefix/grubenv ]; then
  load_env
fi
if [ "${next_entry}" ] ; then
   set default="${next_entry}"
   set next_entry=
   save_env next_entry
   set boot_once=true
else
   set default="0"
fi

if [ x"${feature_menuentry_id}" = xy ]; then
  menuentry_id_option="--id"
else
  menuentry_id_option=""
fi

export menuentry_id_option

if [ "${prev_saved_entry}" ]; then
  set saved_entry="${prev_saved_entry}"
  save_env saved_entry
  set prev_saved_entry=
  save_env prev_saved_entry
  set boot_once=true
fi

function savedefault {
  if [ -z "${boot_once}" ]; then
    saved_entry="${chosen}"
    save_env saved_entry
  fi
}

function load_video {
  if [ x$feature_all_video_module = xy ]; then
    insmod all_video
  else
    insmod efi_gop
    insmod efi_uga
    insmod ieee1275_fb
    insmod vbe
    insmod vga
    insmod video_bochs
    insmod video_cirrus
  fi
}

if [ x$feature_default_font_path = xy ] ; then
   font=unicode
else
insmod part_msdos 
insmod reiserfs
set root='hd0,msdos5'
if [ x$feature_platform_search_hint = xy ]; then
  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos5 --hint-efi=hd0,msdos5 --hint-baremetal=ahci0,msdos5  a3693671-f526-4814-80ab-90d8d5ff633e
else
  search --no-floppy --fs-uuid --set=root a3693671-f526-4814-80ab-90d8d5ff633e
fi
    font="/usr/share/grub/unicode.pf2"
fi

if loadfont $font ; then
  set gfxmode=auto
  load_video
  insmod gfxterm
  set locale_dir=$prefix/locale
  set lang=uk_UA
  insmod gettext
fi
terminal_input console
terminal_output gfxterm
if [ x$feature_timeout_style = xy ] ; then
  set timeout_style=menu
  set timeout=5
# Fallback normal timeout code in case the timeout_style feature is
# unavailable.
else
  set timeout=5
fi
### END /etc/grub.d/00_header ###

### BEGIN /etc/grub.d/10_linux ###
menuentry 'Arch LinuxPF' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-simple-a3693671-f526-4814-80ab-90d8d5ff633e' {
	load_video
	set gfxpayload=keep
	insmod gzio
	insmod part_msdos 
	insmod reiserfs
	set root='hd0,msdos5'
	if [ x$feature_platform_search_hint = xy ]; then
	  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos5 --hint-efi=hd0,msdos5 --hint-baremetal=ahci0,msdos5  a3693671-f526-4814-80ab-90d8d5ff633e
	else
	  search --no-floppy --fs-uuid --set=root a3693671-f526-4814-80ab-90d8d5ff633e
	fi
	echo	'Завантаження Linux linux-pf…'
	linux	/boot/vmlinuz-linux-pf root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw  quiet
	echo	'Завантаження початкового диска у оперативній пам’яті…'
	initrd	/boot/intel-ucode.img /boot/initramfs-linux-pf.img
}
submenu 'Додаткові параметри для Arch Linux' $menuentry_id_option 'gnulinux-advanced-a3693671-f526-4814-80ab-90d8d5ff633e' {
	menuentry 'Arch Linux, з Linux linux-pf' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-linux-pf-advanced-a3693671-f526-4814-80ab-90d8d5ff633e' {
		load_video
		set gfxpayload=keep
		insmod gzio
		insmod part_msdos 
		insmod reiserfs
		set root='hd0,msdos5'
		if [ x$feature_platform_search_hint = xy ]; then
		  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos5 --hint-efi=hd0,msdos5 --hint-baremetal=ahci0,msdos5  a3693671-f526-4814-80ab-90d8d5ff633e
		else
		  search --no-floppy --fs-uuid --set=root a3693671-f526-4814-80ab-90d8d5ff633e
		fi
		echo	'Завантаження Linux linux-pf…'
		linux	/boot/vmlinuz-linux-pf root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw  quiet
		echo	'Завантаження початкового диска у оперативній пам’яті…'
		initrd	/boot/intel-ucode.img /boot/initramfs-linux-pf.img
	}
	menuentry 'Arch Linux, with Linux linux-pf (fallback initramfs)' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-linux-pf-fallback-a3693671-f526-4814-80ab-90d8d5ff633e' {
		load_video
		set gfxpayload=keep
		insmod gzio
		insmod part_msdos 
		insmod reiserfs
		set root='hd0,msdos5'
		if [ x$feature_platform_search_hint = xy ]; then
		  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos5 --hint-efi=hd0,msdos5 --hint-baremetal=ahci0,msdos5  a3693671-f526-4814-80ab-90d8d5ff633e
		else
		  search --no-floppy --fs-uuid --set=root a3693671-f526-4814-80ab-90d8d5ff633e
		fi
		echo	'Завантаження Linux linux-pf…'
		linux	/boot/vmlinuz-linux-pf root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw  quiet
		echo	'Завантаження початкового диска у оперативній пам’яті…'
		initrd	/boot/intel-ucode.img /boot/initramfs-linux-pf-fallback.img
	}
	menuentry 'Arch Linux, з Linux linux' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-linux-advanced-a3693671-f526-4814-80ab-90d8d5ff633e' {
		load_video
		set gfxpayload=keep
		insmod gzio
		insmod part_msdos 
		insmod reiserfs
		set root='hd0,msdos5'
		if [ x$feature_platform_search_hint = xy ]; then
		  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos5 --hint-efi=hd0,msdos5 --hint-baremetal=ahci0,msdos5  a3693671-f526-4814-80ab-90d8d5ff633e
		else
		  search --no-floppy --fs-uuid --set=root a3693671-f526-4814-80ab-90d8d5ff633e
		fi
		echo	'Завантаження Linux linux…'
		linux	/boot/vmlinuz-linux root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw  quiet
		echo	'Завантаження початкового диска у оперативній пам’яті…'
		initrd	/boot/intel-ucode.img /boot/initramfs-linux.img
	}
	menuentry 'Arch Linux, with Linux linux (fallback initramfs)' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'gnulinux-linux-fallback-a3693671-f526-4814-80ab-90d8d5ff633e' {
		load_video
		set gfxpayload=keep
		insmod gzio
		insmod part_msdos 
		insmod reiserfs
		set root='hd0,msdos5'
		if [ x$feature_platform_search_hint = xy ]; then
		  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos5 --hint-efi=hd0,msdos5 --hint-baremetal=ahci0,msdos5  a3693671-f526-4814-80ab-90d8d5ff633e
		else
		  search --no-floppy --fs-uuid --set=root a3693671-f526-4814-80ab-90d8d5ff633e
		fi
		echo	'Завантаження Linux linux…'
		linux	/boot/vmlinuz-linux root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw  quiet
		echo	'Завантаження початкового диска у оперативній пам’яті…'
		initrd	/boot/intel-ucode.img /boot/initramfs-linux-fallback.img
	}
}

### END /etc/grub.d/10_linux ###

### BEGIN /etc/grub.d/20_linux_xen ###
### END /etc/grub.d/20_linux_xen ###

### BEGIN /etc/grub.d/30_os-prober ###
menuentry 'Windows Vista (loader) (на /dev/sda1)' --class windows --class os $menuentry_id_option 'osprober-chain-8A74DE2874DE16B3' {
	insmod part_msdos 
	insmod ntfs
	set root='hd0,msdos1'
	if [ x$feature_platform_search_hint = xy ]; then
	  search --no-floppy --fs-uuid --set=root --hint-bios=hd0,msdos1 --hint-efi=hd0,msdos1 --hint-baremetal=ahci0,msdos1  8A74DE2874DE16B3
	else
	  search --no-floppy --fs-uuid --set=root 8A74DE2874DE16B3
	fi
	chainloader +1
}
### END /etc/grub.d/30_os-prober ###

### BEGIN /etc/grub.d/40_custom ###
# This file provides an easy way to add custom menu entries.  Simply type the
# menu entries you want to add after this comment.  Be careful not to change
# the 'exec tail' line above.
### END /etc/grub.d/40_custom ###

### BEGIN /etc/grub.d/41_custom ###
if [ -f  ${config_directory}/custom.cfg ]; then
  source ${config_directory}/custom.cfg
elif [ -z "${config_directory}" -a -f  $prefix/custom.cfg ]; then
  source $prefix/custom.cfg;
fi
### END /etc/grub.d/41_custom ###

### BEGIN /etc/grub.d/60_memtest86+ ###
### END /etc/grub.d/60_memtest86+ ###

DMESG



full journalctl -b for problematic loading:  (for the case of 4.1.4-pf, but the same story for stock).

Offline

#9 2015-09-07 16:07:34

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Xorglog

Offline

#10 2015-09-07 16:15:22

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

R00KIE wrote:

I'm using the same card here and generally I have no problems(1). Like Lone_Wolf says, post the full dmesg, and xorg log as it may provide more clues.

(1)
Are you using a compositing window manager? I'm asking because I can make things hang under certain conditions. I can manage to make the card hang if I use a compositing manager (in my case XFCE's compositor) and if I'm using glamor either by specifying it on a xorg .conf file or if using the modesetting driver (which uses glamor).

Sometimes things go a bit wonky when watching videos with mplayer using vdpau but I have not seen the machine/card hang when this happens.

I am using xfce too with built in compositor and vdpau and on 3.19.2 linux pf everything fine.
On any kernel from 4.1.x - I have a problem during kernel boot with drm initializing. Why? and what happens with 4.1.x and above kernels  - still a question.


Have some guessing that early KMS in initramfs is not working, for the case I am publishing also initramfs for 4.2.03, in mkinitcpio.conf in MODULES radeon is added! (actually marked as explicit)

lsinitcpio -a /boot/initramfs-linux.img 
==> Image: /boot/initramfs-linux.img 
==> Created with mkinitcpio 18
==> Kernel: 4.2.0-3-ARCH
==> Size: 6.16 MiB
==> Compressed with: gzip
  -> Uncompressed size: 16.48 MiB (.373 ratio)
  -> Estimated extraction time: 0.242s

==> Included modules:
  ahci			  hid			  scsi_mod
  ata_generic		  i2c-algo-bit		  sd_mod
  ata_piix		  i8042			  sdhci
  atkbd			  led-class		  sdhci-pci
  cdrom			  libahci		  serio
  crc-itu-t		  libata		  sr_mod
  drm			  libps2		  ttm
  drm_kms_helper	  mmc_block		  uas
  ehci-hcd		  mmc_core		  uhci-hcd
  ehci-pci		  nls_cp437		  usb-common
  fat			  nls_iso8859-1		  usb-storage
  firewire-core		  pata_acpi		  usbcore
  firewire-ohci		  radeon [explicit]	  usbhid
  firewire-sbp2		  reiserfs [explicit]	  vfat [explicit]

==> Included binaries:
  blkid			  fsck.reiserfs		  switch_root
  busybox		  kmod			  systemd-tmpfiles
  fsck			  mount			  udevadm

==> Early hook run order:
  udev

==> Hook run order:
  consolefont
  keymap
  udev

==> Cleanup hook run order:
  udev

Last edited by niam (2015-09-07 16:50:05)

Offline

#11 2015-09-07 17:14:48

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Another experiment:

After booting with stock 4.2.03 with error in console (display) radeon 0000:01:00.0 Fatal error during GPU init

I did this trick modprobe -r radeon && modprobe radeon and here is the result sad

cat ~/rader/radeonmodprobe 
[   26.609072] sha256_ssse3: Using SSSE3 optimized SHA-256 implementation
[   31.768193] systemd-journald[158]: Failed to read ACL on /var/log/journal/b9df940b2d024c5cb1fff57289715b95/user-1000.journal, ignoring: Operation not supported
[   48.886745] [drm] Module unloaded
[   65.339589] [drm] Initialized drm 1.1.0 20060810
[   65.427890] [drm] radeon kernel modesetting enabled.
[   65.469252] AMD IOMMUv2 driver by Joerg Roedel <jroedel@suse.de>
[   65.469255] AMD IOMMUv2 functionality not available on this system
[   65.510389] CRAT table not found
[   65.510393] Finished initializing topology ret=0
[   65.510437] kfd kfd: Initialized module
[   65.510772] checking generic (c0000000 3d0000) vs hw (0 0)
[   65.511064] [drm] initializing kernel modesetting (RV610 0x1002:0x94C9 0x1043:0x15B2).
[   65.511084] [drm] register mmio base: 0xFDEF0000
[   65.511086] [drm] register mmio size: 65536
[   65.511182] ATOM BIOS: F7Sr
[   65.511210] radeon 0000:01:00.0: VRAM: 256M 0x0000000000000000 - 0x000000000FFFFFFF (256M used)
[   65.511213] radeon 0000:01:00.0: GTT: 512M 0x0000000010000000 - 0x000000002FFFFFFF
[   65.511215] [drm] Detected VRAM RAM=256M, BAR=0M
[   65.511217] [drm] RAM width 64bits DDR
[   65.511269] [TTM] Zone  kernel: Available graphics memory: 2025924 kiB
[   65.511270] [TTM] Initializing pool allocator
[   65.511277] [TTM] Initializing DMA pool allocator
[   65.511303] radeon 0000:01:00.0: Fatal error during GPU init
[   65.511886] [drm] radeon: finishing device.
[   65.518917] [TTM] Trying to take down uninitialized memory manager type 1
[   65.519537] [TTM] Finalizing pool allocator
[   65.519543] [TTM] Finalizing DMA pool allocator
[   65.519606] [TTM] Zone  kernel: Used memory at exit: 0 kiB
[   65.519609] [drm] radeon: ttm finalized
[   65.519926] radeon: probe of 0000:01:00.0 failed with error -12

radeon: probe of 0000:01:00.0 failed with error -12     - what does this mean?
I have switched from linux-firmware 20150722.e10097a-1 to  linux-firmware-20150904.6ebf5d5-1 with rebuilding all initramfs - nothing helps.

Offline

#12 2015-09-07 19:16:34

R00KIE
Forum Fellow
From: Between a computer and a chair
Registered: 2008-09-14
Posts: 4,734

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

From what I can see your card is not the same as mine although the name is the same. Mine (as far as I can tell) is integrated in the northbridge, while yours may be a discrete card. The card on my laptop has only 128MB of dedicated ram while yours has 256MB. My system is all AMD (cpu+gpu) while yours is intel cpu + amd gpu.

From the output of dmesg I'd say your problem most probably starts even before the loading of the radeon driver with:

[    0.262888] vesafb: cannot reserve video memory at 0xc0000000
...
[    0.262916] resource sanity check: requesting [mem 0xc0000000-0xc03cffff], which spans more than PCI Bus 0000:01 [mem 0xc0000000-0xc01fffff 64bit pref]
[    0.262917] ------------[ cut here ]------------
[    0.262921] WARNING: CPU: 0 PID: 1 at arch/x86/mm/ioremap.c:202 __ioremap_caller+0x2c2/0x3b0()
[    0.262922] Info: mapping multiple BARs. Your kernel is fine.
[    0.262923] Modules linked in:
 
[    0.262928] CPU: 0 PID: 1 Comm: BFS/0 Not tainted 4.1.4-pf #1
[    0.262929] Hardware name: ASUSTeK Computer Inc.         F7SR                /F7SR      , BIOS 203     03/11/2008
[    0.262931]  0000000000000000 000000005df14d43 ffff88013ab37b58 ffffffff8158a0e1
[    0.262935]  0000000000000000 ffff88013ab37bb0 ffff88013ab37b98 ffffffff81079c4a
[    0.262937]  ffff88013ab37bc8 ffffc90000800000 00000000c03d0000 00000000003d0000
[    0.262940] Call Trace:
[    0.262946]  [<ffffffff8158a0e1>] dump_stack+0x4c/0x6e
[    0.262949]  [<ffffffff81079c4a>] warn_slowpath_common+0x8a/0xc0
[    0.262951]  [<ffffffff81079cd5>] warn_slowpath_fmt+0x55/0x70
[    0.262954]  [<ffffffff81068322>] __ioremap_caller+0x2c2/0x3b0
[    0.262956]  [<ffffffff81068427>] ioremap_nocache+0x17/0x20
[    0.262960]  [<ffffffff813513cd>] vesafb_probe+0xa0d/0xa60
[    0.262963]  [<ffffffff813faa44>] platform_drv_probe+0x34/0xa0
[    0.262967]  [<ffffffff813f86c9>] driver_probe_device+0x209/0x4c0
[    0.262969]  [<ffffffff813f8a5b>] __driver_attach+0x9b/0xa0
[    0.262972]  [<ffffffff813f89c0>] ? __device_attach+0x40/0x40
[    0.262974]  [<ffffffff813f6423>] bus_for_each_dev+0x73/0xc0
[    0.262977]  [<ffffffff813f7f8e>] driver_attach+0x1e/0x20
[    0.262979]  [<ffffffff813f7b50>] bus_add_driver+0x180/0x250
[    0.262982]  [<ffffffff81935c0f>] ? fbmem_init+0x98/0x98
[    0.262984]  [<ffffffff813f9304>] driver_register+0x64/0xf0
[    0.262987]  [<ffffffff813fa98a>] __platform_driver_register+0x4a/0x50
[    0.262989]  [<ffffffff81935c21>] vesafb_driver_init+0x12/0x14
[    0.262991]  [<ffffffff81002128>] do_one_initcall+0xb8/0x200
[    0.262995]  [<ffffffff818f91b6>] kernel_init_freeable+0x174/0x215
[    0.262999]  [<ffffffff81580810>] ? rest_init+0x90/0x90
[    0.263001]  [<ffffffff8158081e>] kernel_init+0xe/0x100
[    0.263018]  [<ffffffff815905e2>] ret_from_fork+0x42/0x70
[    0.263020]  [<ffffffff81580810>] ? rest_init+0x90/0x90
[    0.263024] ---[ end trace f6f812f7af199aad ]---

Right now I can't think of anything that might help you and this definitely looks like a problem on the kernel side.
You might be able to get some help in the kernel bug tracker by submitting a bug with your findings. Your chances of getting help and a fix more quickly increase if you can do a kernel bisect to try to find the commit that breaks things or at least narrow it down to a small range of commits and test any proposed fixes.

R00KIE
Tm90aGluZyB0byBzZWUgaGVyZSwgbW92ZSBhbG9uZy4K

Offline

#13 2015-09-08 09:41:49

mich41
Member
Registered: 2012-06-22
Posts: 796

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Yep, the kernel apparently fails to allocate CPU address space for the memory-mapped part of GPU VRAM which causes the driver to fail (errno 12 is ENOMEM, BTW).

I don't fully understand this boot log at the first glance so I'm not sure if it's a kernel bug, BIOS bug or maybe PCI specification bug smile Anyway, I'd try this pci=realloc option, as suggested by the kernel. Maybe some other pci= options as well, see kernel-parameters.txt.

BTW, if you decide to report this bug, a dmesg from working kernel would probably be useful.

Last edited by mich41 (2015-09-08 11:54:20)

Offline

#14 2015-09-08 11:18:57

Lone_Wolf
Forum Moderator
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,922

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

niam,

These lines in post #11

[   65.469252] AMD IOMMUv2 driver by Joerg Roedel <jroedel@suse.de>
[   65.469255] AMD IOMMUv2 functionality not available on this system

suggest a wrong bios setting may be causing your problems.

Please
- remove the quiet parameter from the kernel line, that way you will get more info on screen during boot
- check your bios settings for IOMMU and enable it
(if there are no settings called IOMMU, look for settings that have to do with virtualization)

Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#15 2015-09-08 12:18:34

R00KIE
Forum Fellow
From: Between a computer and a chair
Registered: 2008-09-14
Posts: 4,734

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

The were quite a few changes regarding the iommu for intel systems between 3.19 and 4.1 so I'd say it was one of those changes that broke things as it was working fine before.

Regarding the AMD IOMMUv2 messages, I'm not sure that applies given that it is an Intel system, but I may be wrong. It's still worth checking the bios for any options that may affect the mmu settings or virtualization like Lone_Wolf says.

Also check if you have any options for memory hole or something like that and try to change it, the value 0xc0000000 may serve as a reference to compare against any values you may find in the bios.

R00KIE
Tm90aGluZyB0byBzZWUgaGVyZSwgbW92ZSBhbG9uZy4K

Offline

#16 2015-09-08 12:32:35

Lone_Wolf
Forum Moderator
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,922

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

I think AMD was the first to design/introduce IOMMU, but Intel has published a IOMMU specification.

All intel procs that support "Virtualization Technology for Directed I/O" or "VT-d" use Intel IOMMU .

In these 2 messagelines  i think it's the radeon kernelmodule or driver that wrote them and that's why it says AMD IOMMU .

Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

#17 2015-09-08 13:48:23

R00KIE
Forum Fellow
From: Between a computer and a chair
Registered: 2008-09-14
Posts: 4,734

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

I'm still not sure those come from the radeon driver, usually messages from the radeon driver have radeon somewhere, also the radeon driver does not (at least explicitly) depend on anything related to mmu. On top of that there is an amd_iommu_v2 driver, which for some reason is being loaded, however it is not uncommon for drivers to be loaded and probe if supported hardware is present and bail out if they can't find anything they support.

R00KIE
Tm90aGluZyB0byBzZWUgaGVyZSwgbW92ZSBhbG9uZy4K

Offline

#18 2015-09-08 21:59:16

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

mich41 wrote:

Yep, the kernel apparently fails to allocate CPU address space for the memory-mapped part of GPU VRAM which causes the driver to fail (errno 12 is ENOMEM, BTW).

I don't fully understand this boot log at the first glance so I'm not sure if it's a kernel bug, BIOS bug or maybe PCI specification bug smile Anyway, I'd try this pci=realloc option, as suggested by the kernel. Maybe some other pci= options as well, see kernel-parameters.txt.

BTW, if you decide to report this bug, a dmesg from working kernel would probably be useful.

Have tried 

linux	/boot/vmlinuz-linux root=UUID=a3693671-f526-4814-80ab-90d8d5ff633e rw  pci=realloc pci=noacpi#quiet

the same story.
As it is clear from provided by me previously logs = I have this story on ASUS F7SR laptop, actually BIOS is lastest but is dated 2005 (AMIBIOS 203 version).

Regarding reporting bug  - to which list to report Arch or LKML?
I have not checked yet version from Gentoo for the newest cards (R8xx, Radeon HD5xxx and above), kernel-2.6.36 or above is necessary. 
May be I catched this one?

Offline

#19 2015-09-08 22:03:22

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Lone_Wolf wrote:

niam,

These lines in post #11

[   65.469252] AMD IOMMUv2 driver by Joerg Roedel <jroedel@suse.de>
[   65.469255] AMD IOMMUv2 functionality not available on this system

suggest a wrong bios setting may be causing your problems.

Please
- remove the quiet parameter from the kernel line, that way you will get more info on screen during boot
- check your bios settings for IOMMU and enable it
(if there are no settings called IOMMU, look for settings that have to do with virtualization)

As it is mentioned before  - On ASUS F7SR there is AMIBIOS 203 version dated 2005 - there are no any mentioned by you options like IOMMU and Virtualization or any other advanced modern things in it - it is very primitive.

Regarding wrong bios settings - I did not change nothing in bios - 3.19.2-pf (actually I have from the old kernels only this one) - everything is OK while from 4.1.x (pf or not pf) have mentioned above -12 ERROR.

Offline

#20 2015-09-08 22:06:54

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

On core/4.1.6 - the same story.
With pci=realloc pci=noacpi - no boot, stuck on initramfs.
with pci=realloc - the same error -12.

Last edited by niam (2015-09-08 22:29:59)

Offline

#21 2015-09-08 22:35:06

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

looks like since 4.1 there are significant changes to drm
Actually I tend report a bug to LKML. But before would like to make my custom 4.1.x kernel to ensure that problem still exists.

Offline

#22 2015-09-09 14:30:03

mich41
Member
Registered: 2012-06-22
Posts: 796

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

I think linux-pci would be a good place. Not really drm, because the BAR allocation seems to be screwed long before drm starts to run.

If you are going to test your own kernel, note that this issue may be config-dependent before concluding it's an Arch bug smile

Last edited by mich41 (2015-09-09 14:30:47)

Offline

#23 2015-09-20 12:43:36

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Have found may be my problem looks that in 3.19 (which is working) this was fixed but looks like from 4.1 it was again broken.
close problem, probably the same

Offline

#24 2015-10-03 14:30:47

niam
Member
Registered: 2015-09-06
Posts: 13

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

RESOLVED, please see patch that was tested by myself on stock ARCH 4.2.2-1

Arch Team, please fix this in the kernel build!

Offline

#25 2015-10-04 12:51:03

Lone_Wolf
Forum Moderator
From: Netherlands, Europe
Registered: 2005-10-04
Posts: 11,922

Re: Linux kernel 4.1.x and Radeon Rv610 = fatal error during GPU init

Niam,

Glad to see you found a solution.
As you are probably the most knowledgable archlinux user wrt this problem, maybe you could open a bug or feature report in Arch bugtracker to get the patch included ?

Disliking systemd intensely, but not satisfied with alternatives so focusing on taming systemd.


(A works at time B)  && (time C > time B ) ≠  (A works at time C)

Offline

Pages: 1

Board footer

Powered by FluxBB