search

Bumblebee-Project / bbswitch

Disable discrete graphics (currently nvidia only)
GNU General Public License v2.0
486 stars 78 forks source link

Dell XPS 15 9560: System crashes when turning GPU on/off repeatedly #148

Open skoehler opened 7 years ago

skoehler commented 7 years ago

On my brand new Dell XPS 9560, the bbswitch kernel module loads and repeatedly turning the GPU on and off lets the system hang. I don't know why and I can't see any kernel output, unfortunately. I'm using kernel 4.9.7.

My Distro is Gentoo. When booting with systemd, my system would actually hang before I could reach the login prompt. (Probably because the bumblebee daemon tries to disable the GPU). When booting with OpenRC (very serial booting process), I could reach the graphical login primpt, but turning GPU on and off repeatedly would result in a system freeze and crash. I just an echo ON/OFF > /proc/acpi/bbswitch.

It might be that the new 10 series GPU come with new ACPI tables for turning on/off the GPU. On a Dell XPS 15 9550 (960M GPU) everything was working fine.

XVilka commented 6 years ago

For everyone who had the issues with ACPI that required acpi_rev_override - important fixes were merged into the mainline kernel git, so starting from 4.17-rc2 kernel you can try without ACPI override option. I have tried with Dell XPS 9560 and it really works.

chenxiaolong commented 6 years ago

@XVilka That's great news! Do you happen to know which exact patches from 4.17-rc2 fix the issue?

XVilka commented 6 years ago

@chenxiaolong I don't know exact commit - the MLC fixes landed in multiple commits, but it is mostly updated drivers/acpi/acpica from https://github.com/acpica/acpica/

PumbaPe commented 6 years ago

@XVilka I just upgraded from 4.16 to 4.17.2 (Fedora 28), but unfortunately it still doesn't work without acpi_rev_override. The system still hangs before the login screen.

XVilka commented 6 years ago

Very strange, can you record dmesg for exact ACPI error?

On Mon, Jun 25, 2018, 10:12 PM PumbaPe notifications@github.com wrote:

@XVilka https://github.com/XVilka I just upgraded from 4.16 to 4.17.2 (Fedora 28), but unfortunately it still doesn't work without acpi_rev_override. The system still hangs before the login screen.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Bumblebee-Project/bbswitch/issues/148#issuecomment-399965979, or mute the thread https://github.com/notifications/unsubscribe-auth/AAMZ_Sp6LZHa-9JOQ3jjFOKtKdqF6Nl9ks5uAO-3gaJpZM4L3MIz .

PumbaPe commented 6 years ago

@XVilka Actually there aren't any errors. The log shows exactly the same messages regarding ACPI (except the boot parameter of course). The last couple of messages from a failed boot:

$ journalctl -k -b -1 | grep -i acpi
...
Jun 28 01:13:13 kernel: ACPI: AC Adapter [AC] (on-line)
Jun 28 01:13:13 kernel: ACPI: Lid Switch [LID0]
Jun 28 01:13:13 kernel: ACPI: Power Button [PBTN]
Jun 28 01:13:13 kernel: ACPI: Sleep Button [SBTN]
Jun 28 01:13:13 kernel: ACPI: Power Button [PWRF]
Jun 28 01:13:13 kernel: ACPI: Thermal Zone [THM] (25 C)
Jun 28 01:13:13 kernel: ACPI: Battery Slot [BAT0] (battery present)
Jun 28 01:13:14 kernel: ACPI: Video Device [GFX0] (multi-head: yes  rom: no  post: no) 
Jun 28 01:13:14 kernel: ACPI: Video Device [PEGP] (multi-head: no  rom: yes  post: no)

And from a successful boot:

$ journalctl -k -b -2 | grep -i acpi
...
Jun 28 01:14:50 kernel: ACPI: AC Adapter [AC] (on-line)
Jun 28 01:14:50 kernel: ACPI: Lid Switch [LID0]
Jun 28 01:14:50 kernel: ACPI: Power Button [PBTN]
Jun 28 01:14:50 kernel: ACPI: Sleep Button [SBTN]
Jun 28 01:14:50 kernel: ACPI: Power Button [PWRF]
Jun 28 01:14:50 kernel: ACPI: Thermal Zone [THM] (25 C)
Jun 28 01:14:50 kernel: ACPI: Battery Slot [BAT0] (battery present)
Jun 28 01:14:51 kernel: ACPI: Video Device [GFX0] (multi-head: yes  rom: no  post: no) 
Jun 28 01:14:51 kernel: ACPI: Video Device [PEGP] (multi-head: no  rom: yes  post: no) 
Jun 27 23:14:53 kernel: ACPI Warning: \_SB.IETM._TRT: Return Package has no elements (empty) (20180313/nsprepkg-96)
Jun 27 23:14:56 kernel: ACPI Warning: \_SB.PCI0.PEG0.PEGP._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20180313/nsarguments-66)

The only error in the whole log is

$ journalctl -b -1 -p 3
Jun 28 01:13:13 kernel: rtc_cmos 00:02: nvmem registration failed

It's the same for a successful boot though.

mxncr commented 6 years ago

@XVilka Same as @PumbaPe, I have upgrade to kernel 4.17.3 and I still need acpi_rev_override=1 to boot (Manjaro, Dell XPS15 9560).

demurgos commented 6 years ago

Using Arch Linux, updated to the latest versions, I still had to use acpi_rev_override=1 (or bbswitch would cause a black screen after rebooting). I had some trouble finding this info, so I sent the PR #174 to make this information easier to find.