amdgpu errors on 14.2 with kernel 5.6.18

[D Br]

Hello,

Long time lurker and slackware user (since 7.0). Just moved my hard drive over from an old intel i5 750 box using AMD R7 265 to a new X570 box with Ryzen 3600X and 5600XT (16GB in both boxes). I was running the 4.4.240 kernel with no problem, then I built 5.6.18 and get a blank screen on bootup. So I killed it and booted back into the 4.4.240, no problems. So I booted with 5.6.18 again and sshed from my raspberry pi. I see gcvm_protection_fault_status 0x00041f52 in my dmesg. Apparently A LOT of people have been running into this but it was supposedly fixed in 5.6 kernel.

Unfortunately though I've been a long time user, it's been just that; user and I've never run into major issues before. Any thoughts on getting around this?

Dean

[l***@manpage.com]

Are you using the <Huge> Boot option on startup? The drivers for your new system
may not be loading if you are just choosing <Default>

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard drive o=
ver from an old intel i5 750 box using AMD R7 265 to a new X570 box with Ry=
zen 3600X and 5600XT (16GB in both boxes). I was running the 4.4.240 kernel=
with no problem, then I built 5.6.18 and get a blank screen on bootup. So =
I killed it and booted back into the 4.4.240, no problems. So I booted with=
5.6.18 again and sshed from my raspberry pi. I see gcvm_protection_fault_s=
tatus 0x00041f52 in my dmesg. Apparently A LOT of people have been running =
into this but it was supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just that; user =
and I've never run into major issues before. Any thoughts on getting around=
this?
Dean

[D Br]

Post by l***@manpage.com
Are you using the <Huge> Boot option on startup? The drivers for your new system
may not be loading if you are just choosing <Default>

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard drive o=
ver from an old intel i5 750 box using AMD R7 265 to a new X570 box with Ry=
zen 3600X and 5600XT (16GB in both boxes). I was running the 4.4.240 kernel=
with no problem, then I built 5.6.18 and get a blank screen on bootup. So =
I killed it and booted back into the 4.4.240, no problems. So I booted with=
5.6.18 again and sshed from my raspberry pi. I see gcvm_protection_fault_s=
tatus 0x00041f52 in my dmesg. Apparently A LOT of people have been running =
into this but it was supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just that; user =
and I've never run into major issues before. Any thoughts on getting around=
this?
Dean

--
----------------------------------------- --- -- -
Posted with NewsLeecher v7.0 Final
------------------------------- ----- ---- -- -

I copied the .config file from /usr/src/linux to the 5.8.16 directory before doing 'make oldconfig'. My assumption is that the new kernel would be using the same options as the existing one. But I don't know if I'm missing anything at that point.

[Chris Elvidge]

Post by D Br

Post by l***@manpage.com
Are you using the <Huge> Boot option on startup? The drivers for your new system
may not be loading if you are just choosing <Default>

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard drive o=
ver from an old intel i5 750 box using AMD R7 265 to a new X570 box with Ry=
zen 3600X and 5600XT (16GB in both boxes). I was running the 4.4.240 kernel=
with no problem, then I built 5.6.18 and get a blank screen on bootup. So =
I killed it and booted back into the 4.4.240, no problems. So I booted with=
5.6.18 again and sshed from my raspberry pi. I see gcvm_protection_fault_s=
tatus 0x00041f52 in my dmesg. Apparently A LOT of people have been running =
into this but it was supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just that; user =
and I've never run into major issues before. Any thoughts on getting around=
this?
Dean

--
----------------------------------------- --- -- -
Posted with NewsLeecher v7.0 Final
------------------------------- ----- ---- -- -

I copied the .config file from /usr/src/linux to the 5.8.16 directory before doing 'make oldconfig'. My assumption is that the new kernel would be using the same options as the existing one. But I don't know if I'm missing anything at that point.

I prefer to use 'olddefconfig' to make sure I get any new stuff in too.
I'm currently on 5.9.2 in a VBOX.

[King Beowulf]

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new X570
box with Ryzen 3600X and 5600XT (16GB in both boxes). I was running the
4.4.240 kernel with no problem, then I built 5.6.18 and get a blank
screen on bootup. So I killed it and booted back into the 4.4.240, no
problems. So I booted with 5.6.18 again and sshed from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg. Apparently A
LOT of people have been running into this but it was supposedly fixed in
5.6 kernel.

I'm running Slackware64-current (presently 5.4.73) with
Ryzen 7 3800X, X570, RX590, 16GB
Ryzen 9 3900X, X570, RX5700XT, 32GB

No issues. (did a quick search and see mostly ARCH, UBUNTU and that ilk
having issues. Not what I would call "representative".)

Also, updating the kernel in 14.2 is not enough IIRC. I suggest the
kernel-firmware as well, and Mesa. You may be running into issues on
switch from EFI VGA to amdgpudrmfb, and I'm not sure where all the
interrelationships are.

[King Beowulf]

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new X570
box with Ryzen 3600X and 5600XT (16GB in both boxes). I was running the
4.4.240 kernel with no problem, then I built 5.6.18 and get a blank
screen on bootup. So I killed it and booted back into the 4.4.240, no
problems. So I booted with 5.6.18 again and sshed from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg. Apparently A
LOT of people have been running into this but it was supposedly fixed
in 5.6 kernel.

I'm running Slackware64-current (presently 5.4.73) with Ryzen 7 3800X,
X570, RX590, 16GB Ryzen 9 3900X, X570, RX5700XT, 32GB
No issues.

...

Almost forgot: You can also try this,
/etc/profile.d/amdgpu.sh
-----------
#!/bin/sh

#export AMD_DEBUG="nongg,nodma"
export AMD_DEBUG="nongg"
-----------

Either one or the other, may stabilize the NAVI GPU. Here, "nongg" seems
sufficient for the RX5700XT.

[Thomas Gibson]

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard drive over from an old intel i5 750 box using AMD R7 265 to a new X570 box with Ryzen 3600X and 5600XT (16GB in both boxes). I was running the 4.4.240 kernel with no problem, then I built 5.6.18 and get a blank screen on bootup. So I killed it and booted back into the 4.4.240, no problems. So I booted with 5.6.18 again and sshed from my raspberry pi. I see gcvm_protection_fault_status 0x00041f52 in my dmesg. Apparently A LOT of people have been running into this but it was supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just that; user and I've never run into major issues before. Any thoughts on getting around this?
Dean

This may be irrelevant.
I sometimes use a Radeon R7 240/340 card in My Slackware 14.2 system.
Normally I use the internal Radeon RS780L.
I noticed recently that the R7 240/340 card will not run under the
recent xorg-server-1.18.3 versions ( -4 & -5) and will only work with
the original version -2.
If I use the new versions startx halts with 'no screens found'.

[King Beowulf]

Post by Thomas Gibson

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new X570
box with Ryzen 3600X and 5600XT (16GB in both boxes). I was running the
4.4.240 kernel with no problem, then I built 5.6.18 and get a blank
screen on bootup. So I killed it and booted back into the 4.4.240, no
problems. So I booted with 5.6.18 again and sshed from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg. Apparently A
LOT of people have been running into this but it was supposedly fixed
in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just that;
user and I've never run into major issues before. Any thoughts on
getting around this?
Dean

This may be irrelevant.

Nope.

Post by Thomas Gibson
I sometimes use a Radeon R7 240/340 card in My Slackware 14.2 system.
Normally I use the internal Radeon RS780L.
I noticed recently that the R7 240/340 card will not run under the
recent xorg-server-1.18.3 versions ( -4 & -5) and will only work with
the original version -2.
If I use the new versions startx halts with 'no screens found'.

Your GPUs are antiques that do not use the amdgpu kernel model. many of
the old radeon GPUs are being dropped as unsupported. OP says he has a
RX 5600 XT which is a NAVI GPU, and is highly, if not fully supported.
Slackware64-Current works fine on my RX 5700 XT. Kernel 5.6.x may have
some regressions, so for Polaris, Vega and Navi GPUs, either upgrade to
current, or use the current kernel, mesa, xorg, firmware packages.

[D Br]

Post by King Beowulf

Post by Thomas Gibson

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new X570
box with Ryzen 3600X and 5600XT (16GB in both boxes). I was running the
4.4.240 kernel with no problem, then I built 5.6.18 and get a blank
screen on bootup. So I killed it and booted back into the 4.4.240, no
problems. So I booted with 5.6.18 again and sshed from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg. Apparently A
LOT of people have been running into this but it was supposedly fixed
in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just that;
user and I've never run into major issues before. Any thoughts on
getting around this?
Dean

This may be irrelevant.

Nope.

Post by Thomas Gibson
I sometimes use a Radeon R7 240/340 card in My Slackware 14.2 system.
Normally I use the internal Radeon RS780L.
I noticed recently that the R7 240/340 card will not run under the
recent xorg-server-1.18.3 versions ( -4 & -5) and will only work with
the original version -2.
If I use the new versions startx halts with 'no screens found'.

Your GPUs are antiques that do not use the amdgpu kernel model. many of
the old radeon GPUs are being dropped as unsupported. OP says he has a
RX 5600 XT which is a NAVI GPU, and is highly, if not fully supported.
Slackware64-Current works fine on my RX 5700 XT. Kernel 5.6.x may have
some regressions, so for Polaris, Vega and Navi GPUs, either upgrade to
current, or use the current kernel, mesa, xorg, firmware packages.

So just realized I gave the wrong kernel (dyslexia?). It's 5.8.16 not 5.6.18. The amdgpu.sh made no difference.

But one thing I've missed is that I still have modesetting enabled and I may be deprecated at this point.

This weekend got away from me so I haven't gone any further. This morning I added 'nomodeset' to my boot parameters and I got my login screen
when I set my default to to 3. However KDE(Don't laugh, my kids like it) is a nogo.

So my initial is solved. Gotta goto work. Has anyone tried KDE with the amdgpu driver?

BTW thanks for the input sofar.

Dean

[King Beowulf]

Post by D Br

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new
X570 box with Ryzen 3600X and 5600XT (16GB in both boxes). I was
running the 4.4.240 kernel with no problem, then I built 5.6.18 and
get a blank screen on bootup. So I killed it and booted back into
the 4.4.240, no problems. So I booted with 5.6.18 again and sshed
from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg.
Apparently A LOT of people have been running into this but it was
supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just
that; user and I've never run into major issues before. Any thoughts
on getting around this?

...snip...

Post by D Br
So just realized I gave the wrong kernel (dyslexia?). It's 5.8.16 not
5.6.18. The amdgpu.sh made no difference.
But one thing I've missed is that I still have modesetting enabled and I
may be deprecated at this point.
This weekend got away from me so I haven't gone any further. This
morning I added 'nomodeset' to my boot parameters and I got my login
screen when I set my default to to 3. However KDE(Don't laugh, my kids
like it) is a nogo.
So my initial is solved. Gotta goto work. Has anyone tried KDE with the amdgpu driver?

I'm running KDE4 with Slackware64-current on XFX AMD Radeon RX590 Fatboy
OC+ (Polaris) with amdgpu. Although my main box uses Xfce, KDE4 runs
fine on the XFX AMD Radeon RX 5700 XT THICC Ultra III (Navi) with amdgpu.
My eldest son also has a RX5700XT with slackware64-current and KDE plasma
5 (alienbob's Ktown repo) using amdgou with no issues.

As I mentioned earlier (or hinted), Slackware 14.2 is too old for
reliable amdpgu support. You need to AT LEAST upgrade, () is in current
and it works:

kernel -> 5.4.x or newer (5.4.74)
mesa -> 20.1 or newer (20.2.1)
xorg-server -> 1.20.x or newer (1.20.9)
xf86-video-amdgpu -> 19.1 or newer (19.1.0)

Just upgrading the kernel by itself is not enough - that's just the
kernel interface portion. You need a compatible X.org server, Xorg
driver, and the newer amdpgu support libs in Mesa.

[D Br]

Post by King Beowulf

Post by D Br

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new
X570 box with Ryzen 3600X and 5600XT (16GB in both boxes). I was
running the 4.4.240 kernel with no problem, then I built 5.6.18 and
get a blank screen on bootup. So I killed it and booted back into
the 4.4.240, no problems. So I booted with 5.6.18 again and sshed
from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg.
Apparently A LOT of people have been running into this but it was
supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just
that; user and I've never run into major issues before. Any thoughts
on getting around this?

...snip...

Post by D Br
So just realized I gave the wrong kernel (dyslexia?). It's 5.8.16 not
5.6.18. The amdgpu.sh made no difference.
But one thing I've missed is that I still have modesetting enabled and I
may be deprecated at this point.
This weekend got away from me so I haven't gone any further. This
morning I added 'nomodeset' to my boot parameters and I got my login
screen when I set my default to to 3. However KDE(Don't laugh, my kids
like it) is a nogo.
So my initial is solved. Gotta goto work. Has anyone tried KDE with the amdgpu driver?

I'm running KDE4 with Slackware64-current on XFX AMD Radeon RX590 Fatboy
OC+ (Polaris) with amdgpu. Although my main box uses Xfce, KDE4 runs
fine on the XFX AMD Radeon RX 5700 XT THICC Ultra III (Navi) with amdgpu.
My eldest son also has a RX5700XT with slackware64-current and KDE plasma
5 (alienbob's Ktown repo) using amdgou with no issues.
As I mentioned earlier (or hinted), Slackware 14.2 is too old for
reliable amdpgu support. You need to AT LEAST upgrade, () is in current
kernel -> 5.4.x or newer (5.4.74)
mesa -> 20.1 or newer (20.2.1)
xorg-server -> 1.20.x or newer (1.20.9)
xf86-video-amdgpu -> 19.1 or newer (19.1.0)
Just upgrading the kernel by itself is not enough - that's just the
kernel interface portion. You need a compatible X.org server, Xorg
driver, and the newer amdpgu support libs in Mesa.

Will do. That will have to wait till the weekend. Thanks for the info
my Navi is the Gigabyte Windforce. My guess is if it works for the 5700 it should
work for the 5600 but I'll update the group once I'm done.

[D Br]

Post by D Br

Post by King Beowulf

Post by D Br

Post by D Br
Hello,
Long time lurker and slackware user (since 7.0). Just moved my hard
drive over from an old intel i5 750 box using AMD R7 265 to a new
X570 box with Ryzen 3600X and 5600XT (16GB in both boxes). I was
running the 4.4.240 kernel with no problem, then I built 5.6.18 and
get a blank screen on bootup. So I killed it and booted back into
the 4.4.240, no problems. So I booted with 5.6.18 again and sshed
from my raspberry pi.
I see gcvm_protection_fault_status 0x00041f52 in my dmesg.
Apparently A LOT of people have been running into this but it was
supposedly fixed in 5.6 kernel.
Unfortunately though I've been a long time user, it's been just
that; user and I've never run into major issues before. Any thoughts
on getting around this?

...snip...

Post by D Br
So just realized I gave the wrong kernel (dyslexia?). It's 5.8.16 not
5.6.18. The amdgpu.sh made no difference.
But one thing I've missed is that I still have modesetting enabled and I
may be deprecated at this point.
This weekend got away from me so I haven't gone any further. This
morning I added 'nomodeset' to my boot parameters and I got my login
screen when I set my default to to 3. However KDE(Don't laugh, my kids
like it) is a nogo.
So my initial is solved. Gotta goto work. Has anyone tried KDE with the
amdgpu driver?

I'm running KDE4 with Slackware64-current on XFX AMD Radeon RX590 Fatboy
OC+ (Polaris) with amdgpu. Although my main box uses Xfce, KDE4 runs
fine on the XFX AMD Radeon RX 5700 XT THICC Ultra III (Navi) with amdgpu.
My eldest son also has a RX5700XT with slackware64-current and KDE plasma
5 (alienbob's Ktown repo) using amdgou with no issues.
As I mentioned earlier (or hinted), Slackware 14.2 is too old for
reliable amdpgu support. You need to AT LEAST upgrade, () is in current
kernel -> 5.4.x or newer (5.4.74)
mesa -> 20.1 or newer (20.2.1)
xorg-server -> 1.20.x or newer (1.20.9)
xf86-video-amdgpu -> 19.1 or newer (19.1.0)
Just upgrading the kernel by itself is not enough - that's just the
kernel interface portion. You need a compatible X.org server, Xorg
driver, and the newer amdpgu support libs in Mesa.

Will do. That will have to wait till the weekend. Thanks for the info
my Navi is the Gigabyte Windforce. My guess is if it works for the 5700 it should
work for the 5600 but I'll update the group once I'm done.

Hello all,
I did the upgrade to -current with the same result: Screen goes blank during the bootup. I changed to different vga
setting within lilo with no success. However, in my dmesg output there is this curious section:

[ 7.078470] amdgpu 0000:2f:00.0: remove_conflicting_pci_framebuffers: bar 0: 0xe0000000 -> 0xefffffff
[ 7.078554] amdgpu 0000:2f:00.0: remove_conflicting_pci_framebuffers: bar 2: 0xf0000000 -> 0xf01fffff
[ 7.078639] amdgpu 0000:2f:00.0: remove_conflicting_pci_framebuffers: bar 5: 0xfcc00000 -> 0xfcc7ffff
[ 7.078724] amdgpu 0000:2f:00.0: vgaarb: deactivate vga console
[ 7.079422] Console: switching to colour dummy device 80x25
[ 7.079546] [drm] initializing kernel modesetting (NAVI10 0x1002:0x731F 0x1458:0x231D 0xCA).

Which I think is the beginning of a trail to follow: Output not being sent to 0000:2f:00.0.

What does vgaarb do this?
So here's my lspci output:

sudo /sbin/lspci -vs 2f:00.0
Password:
2f:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Navi 10 [Radeon RX 5600 OEM/5600 XT / 5700/5700 XT] (rev ca) (prog-if 00 [VGA controller])
Subsystem: Gigabyte Technology Co., Ltd Navi 10 [Radeon RX 5700 / 5700 XT]
Flags: bus master, fast devsel, latency 0, IRQ 73
Memory at e0000000 (64-bit, prefetchable) [size=256M]
Memory at f0000000 (64-bit, prefetchable) [size=2M]
I/O ports at f000 [size=256]
Memory at fcc00000 (32-bit, non-prefetchable) [size=512K]
Expansion ROM at 000c0000 [disabled] [size=128K]
Capabilities: [48] Vendor Specific Information: Len=08 <?>
Capabilities: [50] Power Management version 3
Capabilities: [64] 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 <?>
Capabilities: [150] Advanced Error Reporting
Capabilities: [200] Physical Resizable BAR
Capabilities: [240] Power Budgeting <?>
Capabilities: [270] Secondary PCI Express
Capabilities: [2a0] Access Control Services
Capabilities: [2b0] Address Translation Service (ATS)
Capabilities: [2c0] Page Request Interface (PRI)
Capabilities: [2d0] Process Address Space ID (PASID)
Capabilities: [320] Latency Tolerance Reporting
Capabilities: [400] Data Link Feature <?>
Capabilities: [410] Physical Layer 16.0 GT/s <?>
Capabilities: [440] Lane Margining at the Receiver <?>
Kernel driver in use: amdgpu
Kernel modules: amdgpu

Thoughts?

Dean