Tracking issue for kexec on LoongArch

[heiher]

Currently, the vmlinux built from Clang cannot be loaded via kexec.

kexec -l --reuse-cmdline vmlinux
kexec -e

[   21.434033] kexec_core: Starting new kernel
[   21.434285] EFI boot flag 0x1
[   21.434317] Command line at 0x9000000000108000
[   21.434362] System table at 0xf730018
[   21.434400] We will call new kernel at 0x90000000013e0000
[   21.434453] Bye ...
[   21.434908] CPU 0 Unable to handle kernel paging request at virtual address 0000000000000031, era == 9000000000100058, ra == 9000000002780a20
[   21.435183] Oops[#1]:
[   21.435300] CPU: 0 PID: 242 Comm: kexec Not tainted 6.6.2+ #27
[   21.435417] Hardware name: QEMU QEMU Virtual Machine, BIOS 0.0.0 02/06/2015
[   21.435546] pc 9000000000100058 ra 9000000002780a20 tp 900000010598c000 sp 900000010598fd00
[   21.435658] a0 0000000000000001 a1 9000000000108000 a2 000000000f730018 a3 90000000013e0000
[   21.435770] a4 900000010581c000 a5 9000000000100000 a6 0000000000000001 a7 0000000000000001
[   21.435883] t0 fffffffffffffff7 t1 0000000000000000 t2 0000000000000000 t3 00000000000000ff
[   21.435996] t4 000000000000ffff t5 0000000000ffff0a t6 00000000ffff0a10 t7 0000000000000016
[   21.436110] t8 9000000003916a24 u0 90000000042e46a0 s9 0000000000000001 s0 900000010581c210
[   21.436224] s1 90000001109f8060 s2 0000000000000008 s3 9000000000100060 s4 001400a52880c2c5
[   21.436337] s5 00000000000007f4 s6 0000555558cfc000 s7 000055557ffcc4c0 s8 0000000000000000
[   21.436453]    ra: 9000000002780a20 relocate_new_kernel+0x0/0x8
[   21.436737]   ERA: 9000000000100058 0x9000000000100058
[   21.436874]  CRMD: 000000b0 (PLV0 -IE -DA +PG DACF=CC DACM=CC -WE)
[   21.437038]  PRMD: 00000000 (PPLV0 -PIE -PWE)
[   21.437136]  EUEN: 00000000 (-FPE -SXE -ASXE -BTE)
[   21.437246]  ECFG: 00071c1c (LIE=2-4,10-12 VS=7)
[   21.437359] ESTAT: 00010000 [PIL] (IS= ECode=1 EsubCode=0)
[   21.437455]  BADV: 0000000000000031
[   21.437510]  PRID: 0014c010 (Loongson-64bit, Loongson-3A5000)
[   21.437611] Modules linked in:
[   21.437685] Process kexec (pid: 242, threadinfo=000000004c51a381, task=00000000b87a44d4)
[   21.437860] Stack : fffffffffee1dead 9000000003e4c748 fffffffffee1dead 9000000003e62c30
[   21.438023]         0000000045584543 900000000432ccb0 900000000432cc98 9000000002862740
[   21.438145]         00007ffff241fd58 00000000ffffffea 00007ffff241fd58 90000000027bf354
[   21.438270]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.438391]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.438520]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.438640]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.438757]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.438876]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.438999]         0000000000000000 0000000000000000 0000000000000000 0000000000000000
[   21.439125]         ...
[   21.439181] Call Trace:
[   21.439349] [<9000000002862740>] kernel_kexec+0xc0/0xe0
[   21.439477] [<90000000027bf354>] sys_reboot+0x1f4/0x280
[   21.439565] [<900000000392f23c>] do_syscall+0x7c/0xa0
[   21.439641] [<90000000027712bc>] handle_syscall+0xbc/0x158
[   21.439738] 
[   21.439776] Code: 29c02060  559653ff  28c282c4 <2880c2c5> 001400a5  02c0235a  02c02318  02ffff9c  43ff9f9f 
[   21.439980] 
[   21.440085] ---[ end trace 0000000000000000 ]---

[heiher]

Patch: https://lore.kernel.org/kexec/20231124085410.107080-1-wangrui@loongson.cn/T/#u

[nathanchance]

If I understand correctly, this should be resolved by https://git.kernel.org/pub/scm/utils/kernel/kexec/kexec-tools.git/commit/?id=eaa717e912eef22706a3ccfb579a544ac1ee0904, so I am closing this because it lives outside of the kernel tree. Thanks a lot for testing and fixing this!