[pve-devel] applied: [PATCH pve-kernel] revert 2 changes in thermal driver causing an early kernel Oops.

[Thomas Lamprecht <t.lamprecht@proxmox.com>]

Am 05/04/2024 um 11:27 schrieb Stoiko Ivanov:
> The second patch, that is reverted (first):
> `thermal: trip: Drop lockdep assertion from thermal_zone_trip_id()`
> only touches code introduced by the first patch.
> The first patch causes the following Oops (reproduced on an old
> HP DL380 G8):
> ```
> [    2.960519] ACPI: button: Power Button [PWRF]
> [    2.963126] BUG: kernel NULL pointer dereference, address: 000000000000000c
> [    2.965667] #PF: supervisor read access in kernel mode
> [    2.966954] #PF: error_code(0x0000) - not-present page
> [    2.966954] PGD 0 P4D 0
> [    2.966954] Oops: 0000 [#1] PREEMPT SMP PTI
> [    2.966954] CPU: 0 PID: 1 Comm: swapper/0 Tainted: G          I        6.5.13-4-pve #1
> [    2.966954] Hardware name: HP ProLiant DL380p Gen8, BIOS P70 05/24/2019
> [    2.966954] RIP: 0010:step_wise_throttle+0x48/0x360
> [    2.966954] Code: 04 25 28 00 00 00 48 89 45 d0 31 c0 48 63 c6 48 8d 14 40 48 8b 87 50 03 00 00 4c 8d 24 90 e8 cf d0 ff ff c6 45 bf 00 89 45 b4 <41> 8b 04 24 41 39 85 78 03 00 00 0f 8d a9 02 00 00 0f 1f 44 00 00
> [    2.966954] RSP: 0000:ffff9e2b8014bae8 EFLAGS: 00010246
> [    2.966954] RAX: 0000000000000002 RBX: 0000000000000001 RCX: 0000000000000000
> [    2.966954] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
> [    2.966954] RBP: ffff9e2b8014bb40 R08: 0000000000000000 R09: 0000000000000000
> [    2.966954] R10: 0000000000000000 R11: 0000000000000000 R12: 000000000000000c
> [    2.966954] R13: ffff8c7ac421d000 R14: 0000000000000001 R15: 0000000000000000
> [    2.966954] FS:  0000000000000000(0000) GS:ffff8c7def600000(0000) knlGS:0000000000000000
> [    2.966954] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [    2.966954] CR2: 000000000000000c CR3: 0000000513a34001 CR4: 00000000000606f0
> [    2.966954] Call Trace:
> [    2.966954]  <TASK>
> ```
> 
> the relevant mainline kernels (6.6.15), corresponding to the
> Ubuntu-patchset (which mixes changes from 6.6.15, with ones from
> 6.1.76) [0] - also boot happily - so I strongly assume that the
> changes depend on one of the many commits introduced in linux-upstream
> between v6.5.1 and v6.6.1.
> As it looks like a refactoring (upon which later changes are based),
> and not a bug-fix in itself - simply dropping it seems sensible.
> 
> Signed-off-by: Stoiko Ivanov <s.ivanov@proxmox.com>
> ---
>  ...rip-Drop-lockdep-assertion-from-ther.patch |  24 ++
>  ...ore-Store-trip-pointer-in-struct-the.patch | 343 ++++++++++++++++++
>  2 files changed, 367 insertions(+)
>  create mode 100644 patches/kernel/0014-Revert-thermal-trip-Drop-lockdep-assertion-from-ther.patch
>  create mode 100644 patches/kernel/0015-Revert-thermal-core-Store-trip-pointer-in-struct-the.patch
> 
>

applied, thanks!