Screen does not wake up from suspend ("No signal") on a Radeon RX 580

Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 1086 bytes --]

Hi!

I've deployed Guix on a desktop computer with pretty much the exact same
config as my laptop.

On my laptop I can run

--8<---------------cut here---------------start------------->8---
$ loginctl suspend
--8<---------------cut here---------------end--------------->8---

to hibernate.  Pressing a key will successfully wake it up.

On my desktop, the same command sends the computer to sleep, but while
pressing a key wakes it up (judging from the fans and the hard drive
spinning up) the screen desperately remains off.  Turning it off and on
again displays "No signal".

In short, my desktop uses:

- Linux-libre kernel, instead of a custom kernel for my laptop.
- BIOS Grub because somehow Grub fails in EFI mode...  Whatever.
- A Radeon RX 580 (which works perfectly so far) instead of an Intel for
my laptop.

My config.scm:

- https://gitlab.com/ambrevar/dotfiles/raw/master/.guix-packages/ambrevar/system/default.scm
- https://gitlab.com/ambrevar/dotfiles/raw/master/.guix-packages/ambrevar/system/desktop-bababa.scm

Any pointer?

--
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[swedebugia]

[-- Attachment #1: Type: text/plain, Size: 1460 bytes --]

Pierre Neidhardt <mail@ambrevar.xyz> skrev: (26 januari 2019 11:39:20 CET)
>Hi!
>
>I've deployed Guix on a desktop computer with pretty much the exact
>same
>config as my laptop.
>
>On my laptop I can run
>
>--8<---------------cut here---------------start------------->8---
>$ loginctl suspend
>--8<---------------cut here---------------end--------------->8---
>
>to hibernate.  Pressing a key will successfully wake it up.
>
>On my desktop, the same command sends the computer to sleep, but while
>pressing a key wakes it up (judging from the fans and the hard drive
>spinning up) the screen desperately remains off.  Turning it off and on
>again displays "No signal".
>
>In short, my desktop uses:
>
>- Linux-libre kernel, instead of a custom kernel for my laptop.
>- BIOS Grub because somehow Grub fails in EFI mode...  Whatever.
>- A Radeon RX 580 (which works perfectly so far) instead of an Intel
>for
>my laptop.
>
>My config.scm:
>
>-
>https://gitlab.com/ambrevar/dotfiles/raw/master/.guix-packages/ambrevar/system/default.scm
>-
>https://gitlab.com/ambrevar/dotfiles/raw/master/.guix-packages/ambrevar/system/desktop-bababa.scm
>
>Any pointer?
>
>--
>Pierre Neidhardt
>https://ambrevar.xyz/

Would it be possible for you to SSH into it to read the kernel messages during wakeup?

On my dev laptop suspend, via custom suspend.sh, and wake up, via power button works.
-- 
Sent from my k-9 mail for Android.

[-- Attachment #2: Type: text/html, Size: 2020 bytes --]

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Maxim Cournoyer]

Pierre Neidhardt <mail@ambrevar.xyz> writes:

> Hi!
>
> I've deployed Guix on a desktop computer with pretty much the exact same
> config as my laptop.
>
> On my laptop I can run
>
> $ loginctl suspend
>
> to hibernate.  Pressing a key will successfully wake it up.
>
> On my desktop, the same command sends the computer to sleep, but while
> pressing a key wakes it up (judging from the fans and the hard drive
> spinning up) the screen desperately remains off.  Turning it off and on
> again displays "No signal".
>
> In short, my desktop uses:
>
> - Linux-libre kernel, instead of a custom kernel for my laptop.
> - BIOS Grub because somehow Grub fails in EFI mode...  Whatever.
> - A Radeon RX 580 (which works perfectly so far) instead of an Intel for
> my laptop.
>
> My config.scm:
>
> - https://gitlab.com/ambrevar/dotfiles/raw/master/.guix-packages/ambrevar/system/default.scm
> - https://gitlab.com/ambrevar/dotfiles/raw/master/.guix-packages/ambrevar/system/desktop-bababa.scm
>
> Any pointer?

I don't know for sure, but that Radeon RX 580 card could be the
problem. I've set up a machine with Debian 9 and a Radeon R9 285 in the
past, and it was a pain to make it work -- it would not work at all
without the firmware blobs, and even then, I had to use a custom kernel
from AMD to get it working correctly...

If you have a 2nd (old nVidia would be the best) card laying around, I'd
suggest trying that -- if it works, you'll know the problem is with the
amggpu driver.

Good luck,

Maxim

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 6696 bytes --]

I forgot to mention: the driver in use is ATI, not Radeon.

> I don't know for sure, but that Radeon RX 580 card could be the
> problem. I've set up a machine with Debian 9 and a Radeon R9 285 in the
> past, and it was a pain to make it work -- it would not work at all
> without the firmware blobs, and even then, I had to use a custom kernel
> from AMD to get it working correctly...

In my case the RX580 work very well Linux libre 4.20.4.

Maybe you Debian 9 was too old (kernel + mesa).

I'm still confused about the state of FOSS-ness with AMD.  Is any blob required
to run them at all?  Are is it only specific to AMDGPU-PRO?

> If you have a 2nd (old nVidia would be the best) card laying around, I'd
> suggest trying that -- if it works, you'll know the problem is with the
> amggpu driver.

Well, everything seems to work, I can control it my computer over SSH, so it's
most probably the RX 580 to blame.

I have also tried connecting my screen both over HDMI and DisplayPort, should that
make a difference.  It doesn't.

> Would it be possible for you to SSH into it to read the kernel messages during wakeup?

I did, and here is the relevant part of the dmesg output:

--8<---------------cut here---------------start------------->8---
[ 2976.265296] PM: suspend entry (deep)
[ 2976.265299] PM: Syncing filesystems ... done.
[ 2976.624270] Freezing user space processes ... (elapsed 0.001 seconds) done.
[ 2976.625607] OOM killer disabled.
[ 2976.625656] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[ 2976.626804] printk: Suspending console(s) (use no_console_suspend to debug)
[ 2976.627367] serial 00:04: disabled
[ 2976.627803] r8169 0000:07:00.0 enp7s0: Link is Down
[ 2976.657161] sd 5:0:0:0: [sdb] Synchronizing SCSI cache
[ 2976.657192] sd 4:0:0:0: [sda] Synchronizing SCSI cache
[ 2976.657629] sd 4:0:0:0: [sda] Stopping disk
[ 2976.657724] sd 5:0:0:0: [sdb] Stopping disk
[ 2977.617490] ACPI: Preparing to enter system sleep state S3
[ 2977.941844] PM: Saving platform NVS memory
[ 2977.941882] Disabling non-boot CPUs ...
[ 2977.958842] smpboot: CPU 1 is now offline
[ 2977.974825] smpboot: CPU 2 is now offline
[ 2977.990581] smpboot: CPU 3 is now offline
[ 2978.014817] smpboot: CPU 4 is now offline
[ 2978.026540] smpboot: CPU 5 is now offline
[ 2978.042508] smpboot: CPU 6 is now offline
[ 2978.058498] smpboot: CPU 7 is now offline
[ 2978.074474] smpboot: CPU 8 is now offline
[ 2978.090460] smpboot: CPU 9 is now offline
[ 2978.106384] smpboot: CPU 10 is now offline
[ 2978.122308] smpboot: CPU 11 is now offline
[ 2978.122935] ACPI: Low-level resume complete
[ 2978.122976] PM: Restoring platform NVS memory
[ 2978.123526] Enabling non-boot CPUs ...
[ 2978.123607] x86: Booting SMP configuration:
[ 2978.123608] smpboot: Booting Node 0 Processor 1 APIC 0x1
[ 2978.125878]  cache: parent cpu1 should not be sleeping
[ 2978.125952] microcode: CPU1: patch_level=0x08008206
[ 2978.126084] CPU1 is up
[ 2978.126110] smpboot: Booting Node 0 Processor 2 APIC 0x2
[ 2978.128338]  cache: parent cpu2 should not be sleeping
[ 2978.128411] microcode: CPU2: patch_level=0x08008206
[ 2978.128491] CPU2 is up
[ 2978.128500] smpboot: Booting Node 0 Processor 3 APIC 0x3
[ 2978.130727]  cache: parent cpu3 should not be sleeping
[ 2978.130792] microcode: CPU3: patch_level=0x08008206
[ 2978.130880] CPU3 is up
[ 2978.130889] smpboot: Booting Node 0 Processor 4 APIC 0x4
[ 2978.133112]  cache: parent cpu4 should not be sleeping
[ 2978.133179] microcode: CPU4: patch_level=0x08008206
[ 2978.133269] CPU4 is up
[ 2978.133279] smpboot: Booting Node 0 Processor 5 APIC 0x5
[ 2978.135516]  cache: parent cpu5 should not be sleeping
[ 2978.135579] microcode: CPU5: patch_level=0x08008206
[ 2978.135678] CPU5 is up
[ 2978.135687] smpboot: Booting Node 0 Processor 6 APIC 0x8
[ 2978.137993]  cache: parent cpu6 should not be sleeping
[ 2978.138064] microcode: CPU6: patch_level=0x08008206
[ 2978.138219] CPU6 is up
[ 2978.138230] smpboot: Booting Node 0 Processor 7 APIC 0x9
[ 2978.140518]  cache: parent cpu7 should not be sleeping
[ 2978.140589] microcode: CPU7: patch_level=0x08008206
[ 2978.140734] CPU7 is up
[ 2978.140744] smpboot: Booting Node 0 Processor 8 APIC 0xa
[ 2978.143023]  cache: parent cpu8 should not be sleeping
[ 2978.143094] microcode: CPU8: patch_level=0x08008206
[ 2978.143251] CPU8 is up
[ 2978.143262] smpboot: Booting Node 0 Processor 9 APIC 0xb
[ 2978.145540]  cache: parent cpu9 should not be sleeping
[ 2978.145610] microcode: CPU9: patch_level=0x08008206
[ 2978.145769] CPU9 is up
[ 2978.145779] smpboot: Booting Node 0 Processor 10 APIC 0xc
[ 2978.148060]  cache: parent cpu10 should not be sleeping
[ 2978.148132] microcode: CPU10: patch_level=0x08008206
[ 2978.148312] CPU10 is up
[ 2978.148324] smpboot: Booting Node 0 Processor 11 APIC 0xd
[ 2978.150602]  cache: parent cpu11 should not be sleeping
[ 2978.150673] microcode: CPU11: patch_level=0x08008206
[ 2978.150872] CPU11 is up
[ 2978.152067] ACPI: Waking up from system sleep state S3
[ 2978.230420] usb usb1: root hub lost power or was reset
[ 2978.230422] usb usb2: root hub lost power or was reset
[ 2978.230633] sd 4:0:0:0: [sda] Starting disk
[ 2978.230636] sd 5:0:0:0: [sdb] Starting disk
[ 2978.231175] serial 00:04: activated
[ 2978.544675] ata2: SATA link down (SStatus 0 SControl 330)
[ 2978.545981] ata9: SATA link down (SStatus 0 SControl 300)
[ 2978.545984] ata1: SATA link down (SStatus 0 SControl 330)
[ 2978.638465] usb 1-9: reset full-speed USB device number 3 using xhci_hcd
[ 2978.705962] ata6: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 2978.706736] ata6.00: configured for UDMA/133
[ 2978.895739] r8169 0000:07:00.0 enp7s0: Link is Down
[ 2979.134479] usb 1-8: reset low-speed USB device number 2 using xhci_hcd
[ 2979.555278] OOM killer enabled.
[ 2979.555329] Restarting tasks ... done.
[ 2979.557153] PM: suspend exit
[ 2979.773654] IPv6: ADDRCONF(NETDEV_UP): enp7s0: link is not ready
[ 2979.773926] Generic PHY r8169-700:00: attached PHY driver [Generic PHY] (mii_bus:phy_addr=r8169-700:00, irq=IGNORE)
[ 2979.876671] IPv6: ADDRCONF(NETDEV_UP): enp7s0: link is not ready
[ 2981.488341] r8169 0000:07:00.0 enp7s0: Link is Up - 100Mbps/Full - flow control rx/tx
[ 2981.488353] IPv6: ADDRCONF(NETDEV_CHANGE): enp7s0: link becomes ready
[ 2984.946005] ata5: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
[ 2984.947770] ata5.00: configured for UDMA/100
--8<---------------cut here---------------end--------------->8---

No error, but no trace of ATI/RADEON either.

-- 
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Boris A. Dekshteyn]

Hello,

Pierre Neidhardt <mail@ambrevar.xyz> writes:

> I forgot to mention: the driver in use is ATI, not Radeon.

> No error, but no trace of ATI/RADEON either.

For the RX5xx series, it should be an "amdgpu" kernel module.

---
 WBR, Boris Dekshteyn

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 518 bytes --]


> For the RX5xx series, it should be an "amdgpu" kernel module.

Indeed.  So I tried to `modprobe amdgpu` and `lspci -k` showed that it was the
driver in use.

That said, we don't have a xf86-video-amdgpu
(https://www.archlinux.org/packages/extra/x86_64/xf86-video-amdgpu/), we
probably need it.

Any idea why amdgpu is not loaded on startup for me?
Maybe I can try a ‘modprobe.blacklist=ati,radeon’.
Still, I wonder why amdgpu is not chosen directly.

-- 
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Boris A. Dekshteyn]

Hello,

Pierre Neidhardt <mail@ambrevar.xyz> writes:

> That said, we don't have a xf86-video-amdgpu
> (https://www.archlinux.org/packages/extra/x86_64/xf86-video-amdgpu/), we
> probably need it.

It would be nice :). In its absence, xf86-video-modesetting should be
used. But without firmware you don't have KMS, so ... 

> Any idea why amdgpu is not loaded on startup for me?
> Maybe I can try a ‘modprobe.blacklist=ati,radeon’.
> Still, I wonder why amdgpu is not chosen directly.

What kernel version is used? What does dmesg say about amdgpu?
Anyway i'm suspect that without a firmware (deblobed in linux-libre) KMS
and power management (and hibernation too) will not work.

---
 WBR, Boris Dekshteyn

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 595 bytes --]


> It would be nice :). In its absence, xf86-video-modesetting should be
> used. But without firmware you don't have KMS, so ... 

Done.

> What kernel version is used?

4.20.4-gnu

> What does dmesg say about amdgpu?

Absolutely nothing.

> Anyway i'm suspect that without a firmware (deblobed in linux-libre) KMS
> and power management (and hibernation too) will not work.

Any idea where I could get more info about that?

Also what good it the amdgpu without the firmware?  What's the role of the
firmware in the picture?

-- 
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Boris A. Dekshteyn]

Pierre Neidhardt <mail@ambrevar.xyz> writes:

>> It would be nice :). In its absence, xf86-video-modesetting should be
>> used. But without firmware you don't have KMS, so ... 
>
> Done.
>
>> What kernel version is used?
>
> 4.20.4-gnu
>
>> What does dmesg say about amdgpu?
>
> Absolutely nothing.
>
>> Anyway i'm suspect that without a firmware (deblobed in linux-libre) KMS
>> and power management (and hibernation too) will not work.
>
> Any idea where I could get more info about that?


From arch-wiki:
--8<---------------cut here---------------start------------->8---
The amdgpu kernel module should load fine automatically on system boot.

If it does not happen, then:

Make sure to #Enable Southern Islands (SI) and Sea Islands (CIK) support
when needed.
Make sure you have the latest linux-firmware package installed. This
driver requires the latest firmware for each model to successfully boot.
Make sure you do not have nomodeset or vga= as a kernel parameter, since
amdgpu requires KMS.
Check that you have not disabled amdgpu by using any kernel module blacklisting.
--8<---------------cut here---------------end--------------->8---

> Also what good it the amdgpu without the firmware?  What's the role of the
> firmware in the picture?

Of course, I could be wrong, but most likely, for hardware acceleration,
power management etc., proprietary firmware required. Without this, only
a very limited functionality of the video card will work.

I think some explanations can be found in this thread:
https://www.reddit.com/r/Amd/comments/950g8r/will_we_be_able_to_game_anytime_soon_using_amd/

---
 WBR, Boris Dekshteyn

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 1138 bytes --]


> --8<---------------cut here---------------start------------->8---
> The amdgpu kernel module should load fine automatically on system boot.
> 
> If it does not happen, then:
> 
> Make sure to #Enable Southern Islands (SI) and Sea Islands (CIK) support
> when needed.
> Make sure you have the latest linux-firmware package installed. This
> driver requires the latest firmware for each model to successfully boot.
> Make sure you do not have nomodeset or vga= as a kernel parameter, since
> amdgpu requires KMS.
> Check that you have not disabled amdgpu by using any kernel module blacklisting.
> --8<---------------cut here---------------end--------------->8---

I read that, and my only guess is that I need the firmware then.
I wonder why I can modprobe "amdgpu" then, and what's the difference between
boot-time and run-time module loading.

> I think some explanations can be found in this thread:
> https://www.reddit.com/r/Amd/comments/950g8r/will_we_be_able_to_game_anytime_soon_using_amd/

Thanks for the links, that sheds some light on the matter indeed!

-- 
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Boris A. Dekshteyn]

Hello,

Pierre Neidhardt <mail@ambrevar.xyz> writes:

> I read that, and my only guess is that I need the firmware then.
> I wonder why I can modprobe "amdgpu" then, and what's the difference between
> boot-time and run-time module loading.

If the module is loaded, it does not mean that it is used.
Try checking out what dmesg says about vesa, for example.

---
 WBR, Boris Dekshteyn

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[ison]

I also have an RX 580 and experience the same issue. Although I was under the
impression that the cause may be the motherboard or CPU (amd ryzen 7 2700x), as
it seems to have other known issues like soft lock-ups.
However, I can confirm that I had the same issue with both the radeon and amdgpu
driver.

Although dmesg would report loading amdgpu if I ran:
lspci -k | grep driver
I would see it report radeon instead of amdgpu. The only way I could get it to
use the amdgpu driver was to add this package to my configs load path:
https://paste.debian.net/1065215/
and also add this to my services:
https://paste.debian.net/1065216/
(appologies if the code is ugly, but it works)

Now lspci reports amdgpu, but I still experience the same sleeping issue.
Personally I wouldn't suspect this is a Guix specific problem, but rather a
driver, kernel, or hardware issue.

Re: Screen does not wake up from suspend ("No signal") on a Radeon RX 580

[Pierre Neidhardt]

[-- Attachment #1: Type: text/plain, Size: 1421 bytes --]


> impression that the cause may be the motherboard or CPU (amd ryzen 7 2700x), as
> it seems to have other known issues like soft lock-ups.

That would be surprising since I can SSH to my machine and everything works perfectly
beside the screen not receiving any signal.  But who knows! :)

> https://paste.debian.net/1065215/

I've added xf86-video-amdgpu to Guix 1-2 weeks ago, so you don't need this
anymore.

> and also add this to my services:
> https://paste.debian.net/1065216/

Why do you need this?  I have managed to get everything to work without custom
services (see below).

> Now lspci reports amdgpu, but I still experience the same sleeping issue.
> Personally I wouldn't suspect this is a Guix specific problem, but rather a
> driver, kernel, or hardware issue.

I've tried adding the amdgpu firmware and now everything works.  Conclusion:
it's not a Guix issue, it's just that AMDGPU won't work without the proprietary
blob.  What a shame...

The ATI driver works reasonably well with the RX580, but it is lacking in terms
of X support.  I noticed that beside failing to wake up from sleep, it also
fails with many RandR queries (for instance `xrandr` would only list 2 modes out
of 14 with my screen, brightness queries would fail with redshift, etc.).

Does anyone know if ATI intends to improve support with the RX580?

-- 
Pierre Neidhardt
https://ambrevar.xyz/

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]