Re: kexec reports "Cannot get kernel _text symbol address" on arm64 platform

From: bhe@xxxxxxxxxx
Date: Mon Aug 14 2023 - 08:40:18 EST


On 08/11/23 at 01:27pm, Pandey, Radhey Shyam wrote:
> > -----Original Message-----
> > From: bhe@xxxxxxxxxx <bhe@xxxxxxxxxx>
> > Sent: Wednesday, August 9, 2023 7:42 AM
> > To: Pandey, Radhey Shyam <radhey.shyam.pandey@xxxxxxx>;
> > piliu@xxxxxxxxxx
> > Cc: kexec@xxxxxxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> > Subject: Re: kexec reports "Cannot get kernel _text symbol address" on
> > arm64 platform
> >
> > On 08/08/23 at 07:17pm, Pandey, Radhey Shyam wrote:
> > > Hi,
> > >
> > > I am trying to bring up kdump on arm64 platform[1]. But I get "Cannot get
> > kernel _text symbol address".
> > >
> > > Is there some Dump-capture kernel config options that I am missing?
> > >
> > > FYI, copied below complete kexec debug log.
> > >
> > > [1]: https://www.xilinx.com/products/boards-and-kits/vck190.html
> >
> > Your description isn't clear. You saw the printing, then your kdump kernel
> > loading succeeded or not?
> >
> > If no, have you tried applying Pingfan's patchset and still saw the issue?
> >
> > [PATCHv7 0/5] arm64: zboot support
> > https://lore.kernel.org/all/20230803024152.11663-1-piliu@xxxxxxxxxx/T/#u
>
> I was able to proceed further with loading with crash kernel on triggering system crash.
> echo c > /proc/sysrq-trigger
>
> But when I copy /proc/vmcore it throws memory abort. Also I see size of /proc/vmcore really huge (18446603353488633856).
> Any possible guess on what could be wrong?

I didn't reproduce this issue on a arm64 baremetal system with the
latest kernel. From the log, It could be the iov_iter convertion
patch which caused this. Can you revert below patch to see if it works?

5d8de293c224 vmcore: convert copy_oldmem_page() to take an iov_iter

>
>
> [ 80.733523] Starting crashdump kernel...
> [ 80.737435] Bye!
> [ 0.000000] Booting Linux on physical CPU 0x0000000001 [0x410fd083]
> [ 0.000000] Linux version 6.5.0-rc4-ge28001fb4e07 (radheys@xhdradheys41) (aarch64-xilinx-linux-gcc.real (GCC) 12.2.0, GNU ld (GNU Binutils) 2.39.0.20220819) #23 SMP Fri Aug 11 16:25:34 IST 2023
> <snip>
>
>
>
> xilinx-vck190-20232:/run/media/mmcblk0p1# cat /proc/meminfo | head
> MemTotal: 2092876 kB
> MemFree: 1219928 kB
> MemAvailable: 1166004 kB
> Buffers: 32 kB
> Cached: 756952 kB
> SwapCached: 0 kB
> Active: 1480 kB
> Inactive: 24164 kB
> Active(anon): 1452 kB
> Inactive(anon): 24160 kB
> xilinx-vck190-20232:/run/media/mmcblk0p1# cp /proc/vmcore dump
> [ 975.284865] Unable to handle kernel level 3 address size fault at virtual address ffff80008d7cf000
> [ 975.293871] Mem abort info:
> [ 975.296669] ESR = 0x0000000096000003
> [ 975.300425] EC = 0x25: DABT (current EL), IL = 32 bits
> [ 975.305738] SET = 0, FnV = 0
> [ 975.308788] EA = 0, S1PTW = 0
> [ 975.311925] FSC = 0x03: level 3 address size fault
> [ 975.316888] Data abort info:
> [ 975.319763] ISV = 0, ISS = 0x00000003, ISS2 = 0x00000000
> [ 975.325245] CM = 0, WnR = 0, TnD = 0, TagAccess = 0
> [ 975.330292] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0
> [ 975.335599] swapper pgtable: 4k pages, 48-bit VAs, pgdp=000005016ef6b000
> [ 975.342297] [ffff80008d7cf000] pgd=10000501eddfe003, p4d=10000501eddfe003, pud=10000501eddfd003, pmd=100005017b695003, pte=00687fff84000703
> [ 975.354827] Internal error: Oops: 0000000096000003 [#4] SMP
> [ 975.360392] Modules linked in:
> 3 975.
> 63440] CBPrUo:a d0c aPID: 664 Comm: cp Tainted: G D 6.5.0-rc4-ge28001fb4e07 #23
> [ 975.372822] Hardware name: Xilinx Versal vck190 Eval board revA (DT)
> [ 975.379165] pstate: a0000005 (NzCv daif -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> [ 975.386119] pc : __memcpy+0x110/0x230
> [ 975.389783] lr : _copy_to_iter+0x3d8/0x4d0
> [ 975.393874] sp : ffff80008dc939a0
> [ 975.397178] x29: ffff80008dc939a0 x28: ffff05013c1bea30 x27: 0000000000001000
> [ 975.404309] x26: 0000000000001000 x25: 0000000000001000 x24: ffff80008d7cf000
> [ 975.411440] x23: 0000040000000000 x22: ffff80008dc93ba0 x21: 0000000000001000
> [ 975.418570] x20: ffff000000000000 x19: 0000000000001000 x18: 0000000000000000
> [ 975.425699] x17: 0000000000000000 x16: 0000000000000000 x15: 0140000000000000
> [ 975.432829] x14: ffff8500a9919000 x13: 0040000000000001 x12: 0000fffef6831000
> [ 975.439958] x11: ffff80008d9cf000 x10: 0000000000000000 x9 : 0000000000000000
> [ 975.447088] x8 : ffff80008d7d0000 x7 : ffff0501addfd358 x6 : 0400000000000001
> [ 975.454217] x5 : ffff0501370e9000 x4 : ffff80008d7d0000 x3 : 0000000000000000
> [ 975.461346] x2 : 0000000000001000 x1 : ffff80008d7cf000 x0 : ffff0501370e8000
> [ 975.468476] Call trace:
> [ 975.470912] __memcpy+0x110/0x230
> [ 975.474221] copy_oldmem_page+0x70/0xac
> [ 975.478050] read_from_oldmem.part.0+0x120/0x188
> [ 975.482663] read_vmcore+0x14c/0x238
> [ 975.486231] proc_reg_read_iter+0x84/0xd8
> [ 975.490233] copy_splice_read+0x160/0x288
> [ 975.494236] vfs_splice_read+0xac/0x10c
> [ 975.498063] splice_direct_to_actor+0xa4/0x26c
> [ 975.502498] do_splice_direct+0x90/0xdc
> [ 975.506325] do_sendfile+0x344/0x454
> [ 975.509892] __arm64_sys_sendfile64+0x134/0x140
> [ 975.514415] invoke_syscall+0x54/0x124
> [ 975.518157] el0_svc_common.constprop.0+0xc4/0xe4
> [ 975.522854] do_el0_svc+0x38/0x98
> [ 975.526162] el0_svc+0x2c/0x84
> [ 975.529211] el0t_64_sync_handler+0x100/0x12c
> [ 975.533562] el0t_64_sync+0x190/0x194
> [ 975.537218] Code: cb01000e b4fffc2e eb0201df 540004a3 (a940342c)
> [ 975.543302] ---[ end trace 0000000000000000 ]---
> t message from systemd-journald@xilinx-vck190-20232 (Tue 2022-11-08 14:16:20 UTC):
>
> kernel[539]: [ 975.354827] Internal error: Oops: 0000000096000003 [#4] SMP
>
>
> Broadcast message from systemd-journald@xilinx-vck190-20232 (Tue 2022-11-08 14:16:20 UTC):
>
> kernel[539]: [ 975.537218] Code: cb01000e b4fffc2e eb0201df 540004a3 (a940342c)
>
> Segmentation fault
> xilinx-vck190-20232:/run/media/mmcblk0p1# ls -lrth /proc/vmcore
> -r-------- 1 root root 16.0E Nov 8 14:05 /proc/vmcore
> xilinx-vck190-20232:/run/media/mmcblk0p1# ls -lh /proc/vmcore
> -r-------- 1 root root 16.0E Nov 8 14:05 /proc/vmcore
> xilinx-vck190-20232:/run/media/mmcblk0p1# ls -l /proc/vmcore
> -r-------- 1 root root 18446603353488633856 Nov 8 14:05 /proc/vmcore
>
> Thanks,
> Radhey
>