Re: radeon_fbdev_river_fbdev: failed to initialize framebuffer and setup emulation

On 8/7/25 9:10 AM, Christian Zigotzky wrote:

I bisected today.

git clone https://gitlab.freedesktop.org/drm/kernel a

git bisect start

git bisect good 2f29b5c231011b94007d2c8a6d793992f2275db1
(Good: video: screen_info: Relocate framebuffers behind PCI bridges)

git bisect bad f62408efc8669b82541295a4611494c8c8c52684
(Bad: drm/xe/vf: Disable CSC support on VF)

4d4c10f763d7808fbade28d83d237411603bca05 is the first bad commit
commit 4d4c10f763d7808fbade28d83d237411603bca05
Author: Mario Limonciello [off-list ref]
Date:   Wed Apr 23 23:31:32 2025 -0500

     PCI: Explicitly put devices into D0 when initializing

     AMD BIOS team has root caused an issue that NVMe storage failed to come
     back from suspend to a lack of a call to _REG when NVMe device was 
probed.

     112a7f9c8edbf ("PCI/ACPI: Call _REG when transitioning D-states") added
     support for calling _REG when transitioning D-states, but this only 
works
     if the device actually "transitions" D-states.

     967577b062417 ("PCI/PM: Keep runtime PM enabled for unbound PCI 
devices")
     added support for runtime PM on PCI devices, but never actually
     'explicitly' sets the device to D0.

     To make sure that devices are in D0 and that platform methods such as
     _REG are called, explicitly set all devices into D0 during 
initialization.

     Fixes: 967577b062417 ("PCI/PM: Keep runtime PM enabled for unbound 
PCI devices")
     Signed-off-by: Mario Limonciello [off-list ref]
     Signed-off-by: Bjorn Helgaas [off-list ref]
     Tested-by: Denis Benato [off-list ref]
     Tested-By: Yijun Shen [off-list ref]
     Tested-By: David Perry [off-list ref]
     Reviewed-by: Rafael J. Wysocki [off-list ref]
     Link: https://patch.msgid.link/20250424043232.1848107-1- 
superm1@kernel.org

  drivers/pci/pci-driver.c |  6 ------
  drivers/pci/pci.c        | 13 ++++++++++---
  drivers/pci/pci.h        |  1 +
  3 files changed, 11 insertions(+), 9 deletions(-)

I bisected with a ATi Radeon 7000 / VE (-device ati-vga,model=rv100) :-)

quoted

On 04 August 2025 at 06:26 pm, mad skateman [off-list ref] wrote:


I am using the R580 chip which is the Radeon X1950 XT.

The board i am using is a T1042 PowerPC based one with e5500 core.

I also lost the ability to start Xorg and therefore just a blank screen.

Dave





Op ma 4 aug 2025, 18:04 schreef Christian Zigotzky 
<chzigotzky@xenosoft.de <mailto:chzigotzky@xenosoft.de>>:


    On 04 August 2025 at 04:42 pm, Alex Deucher wrote:
    > On Sun, Aug 3, 2025 at 11:28 AM Christian Zigotzky
    > <chzigotzky@xenosoft.de <mailto:chzigotzky@xenosoft.de>> wrote:
    >>
    >> Hello,
    >>
    >> I have the same issue on another machine either. Blank screen
    during the
    >> boot. The Radeon graphics framebuffer device doesn't work anymore.
    >>
    >> Here is the modifed code from the DRM updates (drm-
    next-2025-07-30):
    >>
    >> -
    >> https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/
    linux.git/diff/drivers/gpu/drm/radeon/radeon_fbdev.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2 <https://
    web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
    diff/drivers/gpu/drm/radeon/radeon_fbdev.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2>
    >> -
    >> https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/
    linux.git/diff/drivers/gpu/drm/radeon/radeon_drv.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2 <https://
    web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
    diff/drivers/gpu/drm/radeon/radeon_drv.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2>
    >
    > This change just bumps the driver version it shouldn't affect
    the fbdev.
    >
    >> -
    >> https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/
    linux.git/diff/drivers/gpu/drm/radeon/radeon_display.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2 <https://
    web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
    diff/drivers/gpu/drm/radeon/radeon_display.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2>
    >> -
    >> https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/
    linux.git/diff/drivers/gpu/drm/radeon/evergreen_cs.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2 <https://
    web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/
    diff/drivers/gpu/drm/radeon/evergreen_cs.c?
    id=260f6f4fda93c8485c8037865c941b42b9cba5d2>
    >
    > This only affects acceleration, it should not affect fbdev.
    >
    > What chip are you using?  Can you attach your full dmesg
    output?  Can
    > you bisect?  This cultrit could be a core drm change.
    >
    > Alex
    >
    >>
    >> Do you have a patch for reverting these modifications? I would
    like to
    >> revert these modifications for testing.
    >>
    >> Thanks in advance,
    >> Christian
    >>
    >> On 03 August 2025 at 1:35 pm, Christian Zigotzky wrote:
    >>> Hello,
    >>>
    >>> Xorg doesn't work after the DRM updates drm-next-2025-07-30.
    >>>
    >>> Error messages:
    >>>
    >>> [dr:.radeon_fbdev_river_fbdev_proe] *ERROR* failed to initialize
    >>> framebuffer -22
    >>> radeon 0000:01:0.0: [drm] *ERRO* fbdev: Failedto setup
    emulaton (ret=-22)
    >>> brd module loadedm
    >>> lop: module loadd
    >>> mpt3ss version 52.10.00.00 loaded
    >>> nve nvme0: pci fuction 0002:04:0.0
    >>> fsl-sata ffe20000.sata: SataFSL Platform/CS Driver init
    >>> scs host0: sata_fs
    >>> ta1: SATA max DMA/133 irq 68 pm-pol 0
    >>> ;21;39mfsl-sat ffe221000.sata Sata FSL Platfrm/CSB Driver iit
    >>> ata1: Signatue Update detectd @ 0 msecs
    >>> scsihost1: sata_fsl[0m
    >>> ta2: SATA max UMA/133 irq 69 lm-pol 0
    >>> nvme nvm0: 4/0/0 defaul/read/poll queus
    >>> nvme0n1: p1 p p3 p4 p5 p6 p7[0m
    >>> drm:.radeon_fbdv_driver_fbdev_robe] *ERROR* filed to initialze
    >>> framebuffer 22
    >>> fsl_dpaa_mac fe4e6000.ethernt: of_get_mac_adress(/soc@ffe00000/
    >>> fman@40000/ethernet@e6000 failed
    >>> radeon 000:01:00.0: [dr] *ERROR* fbdev Failed to setu
    emulation (ret-22)
    >>> fsl_dpaa_ma ffe4e6000.ethenet: FMan MEMAC[0m
    >>> sl_dpaa_mac ffee6000.ethernet:Using random MA address: 0e
    >>>
    >>> Could you please check the latest DRM updates?
    >>>
    >>> Thanks,
    >>> Christian
    >>>
    >>

    Hello Alex,

    Thanks for your answer.

    I use a BARTS chip (Radeon HD6870). I would like to bisect but I
    don't
    have time for it currently. Sorry

    There are some other user reports:

    - https://forum.hyperion-entertainment.com/viewtopic.php?
    p=60606#p60606 <https://forum.hyperion-entertainment.com/
    viewtopic.php?p=60606#p60606>
    - https://forum.hyperion-entertainment.com/viewtopic.php?
    p=60595#p60595 <https://forum.hyperion-entertainment.com/
    viewtopic.php?p=60595#p60595>

    They use other Radeon graphics chips.

    Bug report: https://github.com/chzigotzky/kernels/issues/15
    <https://github.com/chzigotzky/kernels/issues/15>

    What do you think about the following commit?

    Link:
    https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/
    linux.git/commit/?id=0e7d5874fb6b80c44be3cfbcf1cf356e81d91232
    <https://web.git.kernel.org/pub/scm/linux/kernel/git/torvalds/
    linux.git/commit/?id=0e7d5874fb6b80c44be3cfbcf1cf356e81d91232>

    @Hypex
    Could you please attach your full dmesg output?

    Thanks,
    Christian

Does applying 
https://github.com/torvalds/linux/commit/907a7a2e5bf40c6a359b2f6cc53d6fdca04009e0 
help?