Linux >= 5.9 PV dom0 crash when starting a HVM - AMD Ryzen 4750U, Xen 4.14.1

Linux >= 5.9 PV dom0 crash when starting a HVM - AMD Ryzen 4750U, Xen 4.14.1

[Marek Marczykowski-Górecki]

[-- Attachment #1: Type: text/plain, Size: 5116 bytes --]

Hi,

I've got a report about a Linux regression on AMD hardware, this time
with a backtrace! Reportedly, the last functional kernel was 5.8.x (.16,
but that's because we moved to 5.9.x and skipped .17 and .18). This is
consistent with other reports that anything >= 5.9 is unstable on some
AMD machines.

Details of this config:
 - Xen 4.14.1 close to vanilla hypervisor-wise
 - PV dom0
 - Dom0 kernel also very close to vanilla
 - HVM domains in question both have PCI passthrough and qemu in
   stubdomain (but that's probably not relevant here)
 - AMD Ryzen 4750U

The backtrace:
[   43.497440] kernel BUG at arch/x86/xen/p2m.c:656!
[   43.497923] invalid opcode: 0000 [#1] SMP NOPTI
[   43.498404] CPU: 0 PID: 2347 Comm: xenstored Tainted: G        W         5.10.16-1.fc32.qubes.x86_64 #1
[   43.498874] Hardware name: HP HP EliteBook 835 G7 Notebook PC/8760, BIOS S77 Ver. 01.02.03 11/27/2020
[   43.499362] RIP: e030:___set_phys_to_machine+0x71/0x190
[   43.499847] Code: 85 c0 75 28 b8 01 00 00 00 48 8b 4c 24 08 65 48 2b 0c 25 28 00 00 00 0f 85 0a 01 00 00 00 48 83 c4 10 5b 5d c3 48 83 fe ff 74 da <0f> 0b 48 03 3d ce
[   43.500368] RSP: e02b:ffffc90001ff7cc8 EFLAGS: 00010293
[   43.500895] RAX: 0000000000000000 RBX: 0000000000000001 RCX: ffff888139c18a88
[   43.501426] RDX: 00000000ffffffd8 RSI: 0000000000a51101 RDI: 0000000000147fff
[   43.501958] RBP: 0000000000a51101 B08: 0000000000000000 B09: ffffc90001ff7b30
[   43.502407] R10: ffffc90001ff7b28 R11: ffffffff82944488 R12: ffff88810714c6a8
[   43.503020] R13: 0000000000a51101 R14: 0000000000147fff R15: 0000000000000000
[   43.503557] FS: 00007813cd907d00(0000) GS:ffff888139c00000(0000) kn1GS:0000000000000000
[   43.504101] CS: e030 DS: 0000 ES: 0000 CR0: 0000000000050033
[   43.504651] CR2: 000000000000d8c038 CR3: 0000000102ae4000 CR4: 0000000000050660
[   43.505203] Call Trace:
[   43.505744]  set_foreign_p2m_mapping.part.0+0x164/0x320
[   43.506302]  gntdev_map_grant_pages+0x5d/0x350 (xen_gntdev)
[   43.506852]  gntdev_mmap+0x199/0x2f9 (xen_gntdev)
[   43.507408]  mmap_region+0x47e/0x720
[   43.507957]  do_mmap+0x438/0x540
[   43.508514]  ? security_mmap_file+0x81/0xd0
[   43.509054]  vm_mmap_pgoff+0xdf/0x130
[   43.509606]  ksys_mmap_pgoff+Ox1d6/0x240
[   43.510170]  do_syscall_64+0x33/0x40
[   43.510714]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[   43.511263] RIP: 0033:0x7813cdbfa2e6
[   43.511835] Code: 01 00 66 90 f3 0f 1e fa 41 f7 c1 ff 0f 00 00 75 2b 55 48 89 fd 53 89 cb 48 85 ff 74 37 41 89 da 48 89 ef b8 09 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77
[   43.512442] RSP: 002b:00007ffeb83042f8 EGLAGS 00000246 ORIG_RAX: 0000000000000009
[   43.513045] RAX: ffffffffffffffda RBX: 0000000000000001 RCX: 00007813cdbfa2c6
[   43.513654] RDX: 0000000000000003 RSI: 0000000000001000 RDI: 0000000000000000
[   43.514276] RBP: 0000000000000000 R08: 000000000000000f R09: 0000000000000000
[   43.514688] R10: 0000000000000001 R11: 0000000000000246 R12: 00007ffcb8304310
[   43.515498] R13: 0000000000000003 Ri4: 000000000000000f R15: 0000000000000001
[   43.516188] Modules linked in: nf_tables nfnetlink vfat fat hp_smi sparse_keymap wmi_bmof uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_v412 videobuf2_common 
da_codec_generic k10temp i2c_piix4 snd_pci_acp3x lcdtrig_audio mc snd_hda_codec_hdmi snd_hda_intel btusb snd_intel_dspcfg joydev btrtl snd_hda_codec btbcm btintel snd_hda 
pcm snd_timer iwlwifi snd soundcore cfg80211 ucsi_acpi typec_ucsi rfkill typec wmi cm32181 video industrialio hp_wireless fuse xenfs ip_tables dm_thin_pool dm_persistent
titouch amdgpu crct10duf_pclmul crc32_pclmul crc32c_intel nvme iommu_v2 gpu_sched ttm i2c_algo_bit drm_kms_helper cec drm ghash_cimuini_intel xhci_pci xhci_pci_renesas 
back xen_blkback xen_gntalloc xen_gntdev
[   43.516186] xen_evtchn uinput
[   43.528491] ---[ end trace 0f195b254c08766c ]---
[   43.689504] RIP: e030:__set_phys_to_machine+0x71/0x190
[   43.618326] Code: 05 c0 75 28 b8 01 00 00 00 48 0b 4c 24 O8 65 48 2b 0c 25 28 00 00 00 Of 05 0a 01 00 00 48 83 c4 10 5b 5d c3 48 83 fe ff 74 da <Of> 0b 48 03 34 ce 31
[   43.611124] RSP: e02b:ffffc90001ff7cc8 EFLAGS: 00010293
[   43.611933] RAX: 0000000000000000 RBX: 0000000000000001 RCX: ffff888139c18a88
[   43.612736] RDX: 00000000ffffffd8 RSI: 0000000000a51101 RDI: 0000000000147fff
[   43.613534] RBP: 9999999999a51101 R08: 0000000000000000 R09: ffffc90001ff7b38
[   43.614320] R10: ffffc90001ff7b20 R11: fffffffff8294488 R12: ffff88818714c6a8
[   43.615119] R13: 0000000000a51101 R14: 0000000000147fff R15: 0000000000000000
[   43.615925] FS:  00007813cd987d00(0000) GS:ffff888139c00000(0000) kn1GS:0000000000000000
[   43.616714] CS:  e030 DS: 0000 ES: 0000 CRO: 0000000000050033
[   43.617519] CR2: 0000000000d8c038 CR3: 0000000102ae4000 CR4: 0000000000058660
[   43.618313] Kernel panic - not syncing: Fatal exception
[   43.619137] Kernel Offset: disabled

The original report:
https://github.com/QubesOS/qubes-issues/issues/6397#issuecomment-788879968

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]