linux-kernel - Fwd: Hybrid graphics regression in 5.4.49+ [bisected]

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <CAErSpo5sTeK_my1dEhWp7aHD0xOp87+oHYWkTjbL7ALgDbXo-Q@mail.gmail.com>
Date:   Wed, 15 Jul 2020 10:37:53 -0500
From:   Bjorn Helgaas <bhelgaas@...gle.com>
To:     Linux PCI <linux-pci@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Cc:     Bjorn Helgaas <bjorn@...gaas.com>,
        Patrick Volkerding <volkerdi@...il.com>,
        Mika Westerberg <mika.westerberg@...ux.intel.com>
Subject: Fwd: Hybrid graphics regression in 5.4.49+ [bisected]

---------- Forwarded message ---------
From: Patrick Volkerding <volkerdi@...il.com>
Date: Tue, Jul 14, 2020 at 2:22 PM
Subject: Hybrid graphics regression in 5.4.49+ [bisected]
To: Bjorn Helgaas <bhelgaas@...gle.com>, Mika Westerberg
<mika.westerberg@...ux.intel.com>
Cc: Patrick Volkerding <volkerdi@...il.com>


Hello,

I ran into a problem recently with a laptop with hybrid graphics. Here
are the specs of the machine:

OS: Slackware 14.2 x86_64 (post 14.2 -current) x86_64
Host: 20QV000GUS ThinkPad X1 Extreme 2nd
CPU: Intel i7-9850H (12) @ 4.600GHz
GPU: NVIDIA GeForce GTX 1650 Mobile / Max-Q
GPU: Intel UHD Graphics 630

The graphics are set to hybrid in the BIOS, and no proprietary drivers
are in use.

With 5.4.48, starting X from a console works as expected and the Intel
UHD Graphics 630 is used.

Starting with 5.4.49, attempting to start X from a console causes the
machine to become unresponsive, and thousands of these messages are
found in the syslog:

Jul  1 14:39:26 z-mp kernel: [ 1657.318941] nouveau 0000:01:00.0: fifo:
PBDMA0: 01000000 [] ch 0 [00ff992000 DRM] subc 0 mthd 0008 data 00000000

A git bisect between 5.4.48 and 5.4.49 pointed to the commit
[828b192c57e8f4fee77f7a34bd19c1b58b049dad] PCI/PM: Assume ports without
DLL Link Active train links in 100 ms.

The commit reverts cleanly from 5.4.51, and then that kernel works as
expected. Additionally I should note that an unmodified kernel 5.7.7
also works as expected.

Happy to help test any proposed fix.

Here's the git bisect log:

git bisect start
# good: [67cb016870e2fa9ffc8d34cf20db5331e6f2cf4d] Linux 5.4.48
git bisect good 67cb016870e2fa9ffc8d34cf20db5331e6f2cf4d
# good: [67cb016870e2fa9ffc8d34cf20db5331e6f2cf4d] Linux 5.4.48
git bisect good 67cb016870e2fa9ffc8d34cf20db5331e6f2cf4d
# bad: [99705220b22ca116457edeae51ae817d056a6622] net: core:
device_rename: Use rwsem instead of a seqcount
git bisect bad 99705220b22ca116457edeae51ae817d056a6622
# bad: [7454c171a88c188583bbfe17db1f9e6e75723b5f] tty: n_gsm: Fix bogus
i++ in gsm_data_kick
git bisect bad 7454c171a88c188583bbfe17db1f9e6e75723b5f
# good: [390f1688fb4648474b29c8ea6c2c710b689cd669] f2fs: handle readonly
filesystem in f2fs_ioc_shutdown()
git bisect good 390f1688fb4648474b29c8ea6c2c710b689cd669
# good: [9eb54d0e8962d2f5d7a430f3a245037c990e3b42] PCI: v3-semi: Fix a
memory leak in v3_pci_probe() error handling paths
git bisect good 9eb54d0e8962d2f5d7a430f3a245037c990e3b42
# good: [4536dbe64bd17b7d16cb0da56a15616dda247940] USB: ohci-sm501: fix
error return code in ohci_hcd_sm501_drv_probe()
git bisect good 4536dbe64bd17b7d16cb0da56a15616dda247940
# bad: [b1bc8753eefc82b79e285977c6177c67bc17344e] ipmi: use vzalloc
instead of kmalloc for user creation
git bisect bad b1bc8753eefc82b79e285977c6177c67bc17344e
# good: [860b8717f7247384e500ba15c89e9b5327d00423] tty: n_gsm: Fix
waking up upper tty layer when room available
git bisect good 860b8717f7247384e500ba15c89e9b5327d00423
# good: [c6737f3a1b01120ce0991bb0b6b4c8da2b086f1e] HID: Add quirks for
Trust Panora Graphic Tablet
git bisect good c6737f3a1b01120ce0991bb0b6b4c8da2b086f1e
# bad: [5c2207ba2394ee6c2dd7383890818aca89ff4b9b] habanalabs: increase
timeout during reset
git bisect bad 5c2207ba2394ee6c2dd7383890818aca89ff4b9b
# bad: [828b192c57e8f4fee77f7a34bd19c1b58b049dad] PCI/PM: Assume ports
without DLL Link Active train links in 100 ms
git bisect bad 828b192c57e8f4fee77f7a34bd19c1b58b049dad
# first bad commit: [828b192c57e8f4fee77f7a34bd19c1b58b049dad] PCI/PM:
Assume ports without DLL Link Active train links in 100 ms

Best regards,

Pat