Glad to hear the install went well. Congrads!

Yes backups are very important.

You're Welcome.
Hope you have a good week ahead too.

-::-Don't forget to mark your thread SOLVED.-::-

Hi and thanks :-)
Hope you all had a great weekend/week too.

Came back and had a play with kernels 4.10.2 and 4.11-rc2
Looks like I could get there but only for these errors below:

Kernel 4.10.2
bash-4.4# dmesg | grep amd
[ 6.938614] [drm] amdgpu kernel modesetting enabled.
[ 6.938771] amdgpu 0000:01:00.0: enabling device (0000 -> 0003)
[ 7.061994] amdgpu 0000:01:00.0: VRAM: 2048M 0x0000000000000000 - 0x000000007FFFFFFF (2048M used)
[ 7.062001] amdgpu 0000:01:00.0: GTT: 2048M 0x0000000080000000 - 0x00000000FFFFFFFF
[ 7.062099] [drm] amdgpu: 2048M of VRAM memory ready
[ 7.062102] [drm] amdgpu: 2048M of GTT memory ready.
[ 7.063792] amdgpu 0000:01:00.0: PCIE GART of 2048M enabled (table at 0x0000000000040000).
[ 7.063850] amdgpu 0000:01:00.0: amdgpu: using MSI.
[ 7.063865] [drm] amdgpu: irq initialized.
[ 7.066224] [drm] amdgpu: dpm initialized
[ 7.096468] amdgpu 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000080000010, cpu addr 0xffff88041ae53010
[ 7.096508] amdgpu 0000:01:00.0: fence driver on ring 1 use gpu addr 0x0000000080000020, cpu addr 0xffff88041ae53020
[ 7.096536] amdgpu 0000:01:00.0: fence driver on ring 2 use gpu addr 0x0000000080000030, cpu addr 0xffff88041ae53030
[ 7.096576] amdgpu 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000080000040, cpu addr 0xffff88041ae53040
[ 7.096607] amdgpu 0000:01:00.0: fence driver on ring 4 use gpu addr 0x0000000080000050, cpu addr 0xffff88041ae53050
[ 7.658139] [drm] Initialized amdgpu 3.9.0 20150101 for 0000:01:00.0 on minor 1
[ 31.236275] amdgpu 0000:01:00.0: PCIE GART of 2048M enabled (table at 0x0000000000040000).
[ 31.582316] [drm:gfx_v6_0_ring_test_ring [amdgpu]] *ERROR* amdgpu: ring 0 test failed (scratch(0x2140)=0xCAFEDEAD)
[ 31.582328] [drm:amdgpu_resume [amdgpu]] *ERROR* resume of IP block <gfx_v6_0> failed -22
[ 31.582338] [drm:amdgpu_device_resume [amdgpu]] *ERROR* amdgpu_resume failed (-22).
[ 31.913664] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 32.712199] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 34.178561] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
bash-4.4#


Gets better with the 4.11-rc2 kernel as I can now see modules linked in.
But the error is still there.
Will have to google for it after work.

bash-4.4# dmesg | grep amd
[ 7.172116] [drm] amdgpu kernel modesetting enabled.
[ 7.175185] amdgpu 0000:01:00.0: enabling device (0000 -> 0003)
[ 7.229914] amdgpu 0000:01:00.0: VRAM: 2048M 0x0000000000000000 - 0x000000007FFFFFFF (2048M used)
[ 7.230824] amdgpu 0000:01:00.0: GTT: 2048M 0x0000000080000000 - 0x00000000FFFFFFFF
[ 7.235680] [drm] amdgpu: 2048M of VRAM memory ready
[ 7.236319] [drm] amdgpu: 2048M of GTT memory ready.
[ 7.239698] amdgpu 0000:01:00.0: PCIE GART of 2048M enabled (table at 0x0000000000040000).
[ 7.241670] amdgpu 0000:01:00.0: amdgpu: using MSI.
[ 7.242291] [drm] amdgpu: irq initialized.
[ 7.269024] [drm] amdgpu: dpm initialized
[ 7.397673] amdgpu 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000080000010, cpu addr 0xffff88041c45a010
[ 7.397731] amdgpu 0000:01:00.0: fence driver on ring 1 use gpu addr 0x0000000080000020, cpu addr 0xffff88041c45a020
[ 7.397772] amdgpu 0000:01:00.0: fence driver on ring 2 use gpu addr 0x0000000080000030, cpu addr 0xffff88041c45a030
[ 7.397833] amdgpu 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000080000040, cpu addr 0xffff88041c45a040
[ 7.397875] amdgpu 0000:01:00.0: fence driver on ring 4 use gpu addr 0x0000000080000050, cpu addr 0xffff88041c45a050
[ 7.964721] [drm] Initialized amdgpu 3.10.0 20150101 for 0000:01:00.0 on minor 0
[ 8.152750] Modules linked in: snd_hda_codec_hdmi btusb hid_generic usbhid btrtl hid coretemp intel_rapl btbcm btintel bluetooth x86_pkg_temp_thermal intel_powerclamp kvm_intel kvm irqbypass i915 snd_hda_codec_realtek snd_hda_codec_generic rtsx_pci_ms memstick snd_hda_intel rtsx_pci_sdmmc snd_hda_codec snd_hwdep snd_hda_core snd_pcm mmc_core crc32_pclmul toshiba_wmi amdgpu hwmon ttm i2c_dev drm_kms_helper snd_timer sparse_keymap crc32c_intel rtsx_pci intel_gtt syscopyarea sysfillrect sysimgblt fb_sys_fops drm iwlmvm mac80211 iwlwifi mei_me mei agpgart snd r8169 cfg80211 ghash_clmulni_intel mii psmouse soundcore i2c_i801 i2c_algo_bit evdev i2c_core lpc_ich xhci_pci xhci_hcd ehci_pci ehci_hcd shpchp serio_raw efivars mfd_core thermal toshiba_haps video toshiba_bluetooth rfkill wmi ac battery button
[ 31.781060] amdgpu 0000:01:00.0: PCIE GART of 2048M enabled (table at 0x0000000000040000).
[ 32.128398] [drm:gfx_v6_0_ring_test_ring [amdgpu]] *ERROR* amdgpu: ring 0 test failed (scratch(0x2140)=0xCAFEDEAD)
[ 32.128409] [drm:amdgpu_resume [amdgpu]] *ERROR* resume of IP block <gfx_v6_0> failed -22
[ 32.128418] [drm:amdgpu_device_resume [amdgpu]] *ERROR* amdgpu_resume failed (-22).
[ 32.453450] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 33.042246] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 33.488176] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 34.741386] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 48.702603] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
bash-4.4#


Perhaps I should update the amdgpu driver as it is the latest from the current repo?

bash-4.4# slackpkg info xf86-video-amdgpu-1.2.0-x86_64-1

PACKAGE NAME: xf86-video-amdgpu-1.2.0-x86_64-1.txz
PACKAGE LOCATION: ./slackware64/x
PACKAGE SIZE (compressed): 68 K
PACKAGE SIZE (uncompressed): 220 K
PACKAGE DESCRIPTION:
xf86-video-amdgpu: xf86-video-amdgpu (AMD GPU driver for the Xorg X server)
xf86-video-amdgpu:
xf86-video-amdgpu: xf86-video-amdgpu is part of X11.
xf86-video-amdgpu:
xf86-video-amdgpu: For more information about the X.Org Foundation (the providers of the
xf86-video-amdgpu: X.Org implementation of the X Window System), see their website:
xf86-video-amdgpu:
xf86-video-amdgpu: http://www.x.org
xf86-video-amdgpu:

Thanks in advance.

They've had a decent amount of commits to the codebase, but no new release yet.

https://cgit.freedesktop.org/xorg/dr...eo-amdgpu/log/

There was an announcement today for xf86-video-amdgpu 1.3.0
https://lists.x.org/archives/xorg-an...ch/002788.html

1 members found this post helpful.

Haha, what timing!

Yeah, that was good timing :-)

And I had another go at it...
I cloned these two links locally, then created/installed xf86-video-amdgpu-1.3.0-x86_64-1 and libdrm-2.4.75-x86_64-1-18-03-2017 (they have newer patches)
git://anongit.freedesktop.org/xorg/driver/xf86-video-amdgpu
git://anongit.freedesktop.org/mesa/drm

Then followed the instructions from these good people:
1. To pass/try some boot parameters to a test 4.11.0-rc2 kernel
http://www.mikejonesey.co.uk/linux/a...ver-parameters
2. To include the firmware blobs in the test kernel
https://wiki.gentoo.org/wiki/AMDGPU

And the same error with drm:gfx_v6_0_ring_test_ring
I'll resume the search for the meaning of this error later this week.

bash-4.4# dmesg | grep drm
[ 7.964061] [drm] Memory usable by graphics device = 2048M
[ 7.964731] fb: switching to inteldrmfb from simple
[ 7.965464] [drm] Replacing VGA console driver
[ 7.971520] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 7.971527] [drm] Driver supports precise vblank timestamp query.
[ 7.997779] [drm] Initialized i915 1.6.0 20170123 for 0000:00:02.0 on minor 0
[ 7.998442] [drm] Initialized i915 1.6.0 20170123 for 0000:00:02.0 on minor 0
[ 8.006553] fbcon: inteldrmfb (fb0) is primary device
[ 8.108644] [drm] amdgpu kernel modesetting enabled.
[ 9.215208] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device
[ 9.215811] [drm] initializing kernel modesetting (VERDE 0x1002:0x6823 0x1179:0xF945 0x00).
[ 9.215903] [drm] register mmio base: 0xF7B00000
[ 9.215941] [drm] register mmio size: 262144
[ 9.247666] [drm] GPU posting now...
[ 9.293326] [drm] Detected VRAM RAM=2048M, BAR=256M
[ 9.293344] [drm] RAM width 128bits UNKNOWN
[ 9.297684] [drm] amdgpu: 2048M of VRAM memory ready
[ 9.298462] [drm] amdgpu: 2048M of GTT memory ready.
[ 9.299252] [drm] GART: num cpu pages 524288, num gpu pages 524288
[ 9.302461] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 9.303238] [drm] Driver supports precise vblank timestamp query.
[ 9.304847] [drm] amdgpu: irq initialized.
[ 9.319621] [drm] probing gen 2 caps for device 8086:c01 = 261ac83/e
[ 9.319637] [drm] Internal thermal controller without fan control
[ 9.319645] [drm] amdgpu: dpm initialized
[ 9.319662] [drm] AMDGPU Display Connectors
[ 9.327948] [drm] probing gen 2 caps for device 8086:c01 = 261ac83/e
[ 9.327951] [drm] PCIE gen 3 link speeds already enabled
[ 9.363352] Modules linked in: snd_hda_codec_hdmi hid_generic btusb btrtl btbcm usbhid btintel bluetooth amdgpu(+) hid coretemp hwmon ttm intel_rapl i915 drm_kms_helper x86_pkg_temp_thermal snd_hda_codec_realtek intel_powerclamp syscopyarea snd_hda_codec_generic iwlmvm sysfillrect mac80211 sysimgblt fb_sys_fops drm rtsx_pci_sdmmc snd_hda_intel kvm_intel intel_gtt snd_hda_codec agpgart rtsx_pci_ms memstick kvm i2c_dev mmc_core toshiba_wmi sparse_keymap iwlwifi snd_hwdep cfg80211 snd_hda_core evdev i2c_algo_bit mei_me r8169 efivars snd_pcm irqbypass mii rtsx_pci crc32_pclmul thermal psmouse crc32c_intel shpchp xhci_pci xhci_hcd ghash_clmulni_intel i2c_i801 video serio_raw mei lpc_ich ehci_pci snd_timer ehci_hcd i2c_core snd toshiba_haps mfd_core soundcore toshiba_bluetooth rfkill battery wmi button
[ 9.519800] [drm] ring test on 0 succeeded in 8 usecs
[ 9.522589] [drm] ring test on 1 succeeded in 1 usecs
[ 9.523370] [drm] ring test on 2 succeeded in 1 usecs
[ 9.524135] [drm] ring test on 3 succeeded in 3 usecs
[ 9.524905] [drm] ring test on 4 succeeded in 4 usecs
[ 9.525847] [drm] ib test on ring 0 succeeded
[ 9.526614] [drm] ib test on ring 1 succeeded
[ 9.527365] [drm] ib test on ring 2 succeeded
[ 9.528100] [drm] ib test on ring 3 succeeded
[ 9.528828] [drm] ib test on ring 4 succeeded
[ 9.886209] [drm] Initialized amdgpu 3.10.0 20150101 for 0000:01:00.0 on minor 1
[ 33.806514] [drm] probing gen 2 caps for device 8086:c01 = 261ac83/e
[ 33.806517] [drm] PCIE gen 3 link speeds already enabled
[ 34.153284] [drm:gfx_v6_0_ring_test_ring [amdgpu]] *ERROR* amdgpu: ring 0 test failed (scratch(0x2140)=0xCAFEDEAD)
[ 34.153295] [drm:amdgpu_resume [amdgpu]] *ERROR* resume of IP block <gfx_v6_0> failed -22
[ 34.153304] [drm:amdgpu_device_resume [amdgpu]] *ERROR* amdgpu_resume failed (-22).
[ 34.461657] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 34.741508] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
[ 35.722204] [drm:amdgpu_fill_buffer [amdgpu]] *ERROR* Trying to clear memory with ring turned off.
bash-4.4#

I read up on this thread and i am a bit curious to why you want this?
I also have an southern island dgpu but i never felt any need for amdgpu.
And why do you want to use the AMD gpu all the time? since it got no display connectors you will always pass everything thru the intel chip anyway.
I'm just trying to understand why, but sometimes it just researching ones curiosity to see if something can be done.

Hello Nille,
https://cdn.meme.am/instances/51907583.jpg

Meanwhile check this post:
varunpriolkar.com/2016/12/how-to-use-amdgpu-driver-for-southern-islands-and-sea-islands-card-on-ubuntu-linux/
Should provide better performance for gaming, gpu intensive apps, etc...

Cheers :-)

While eventually this should be the case, for the most part, it's about the same.

Things have steadily been improving, but amdgpu isn't a clear winner over radeon on all cards and all software just yet.

I've worked out the errors appear when I start xorg.
I think I'm going to stick with radeon for now and revisit amdgpu at a much later date this year.
Thanks again for the assistance, was really appreciated.

My hope is that by the time the next LTS kernel is released (maybe 4.15 or 4.16) that things will have stabilized and they'll have resolved most of the regressions. I imagine once they enable amdgpu support for Southern Islands by default, it's a good indication things have improved immensely.

Will look forward to it :-)
Meanwhile adding a new "xorg_prime.sh" file in the profile.d directory just works.
bash-4.4# cat /etc/profile.d/xorg_prime.sh

#!/bin/bash
export DRI_PRIME=1
export VDPAU_DRIVER=radeonsi
bash-4.4#

vlc just works and turns on the discrete card without having to pass DRI Pime to it.

And I use opera browser which surprisingly also turned on the discrete card

bash-4.4# cat /sys/kernel/debug/vgaswitcheroo/switch
0:IGD:+:Pwr:0000:00:02.0
1IS: ynPwr:0000:01:00.0
bash-4.4#

And it runs really sloooow.
I may just have to get rid of opera after many many years and try the new kid on the block "vivaldi" and see.

Thanks and have a gret weekend.