The following kernel crash was noticed on arm64 Juno-r2 and Raspberry Pi 4
Model B on Linux next-20230321.
Reported-by: Linux Kernel Functional Testing <[email protected]>
log:
----
[ 3.071500] Unable to handle kernel paging request at virtual
address 0000000000001000
[ 3.079432] Mem abort info:
[ 3.082225] ESR = 0x0000000096000004
[ 3.085977] EC = 0x25: DABT (current EL), IL = 32 bits
[ 3.091295] SET = 0, FnV = 0
[ 3.094350] EA = 0, S1PTW = 0
[ 3.097491] FSC = 0x04: level 0 translation fault
[ 3.102373] Data abort info:
[ 3.105252] ISV = 0, ISS = 0x00000004
[ 3.109089] CM = 0, WnR = 0
[ 3.112055] [0000000000001000] user address but active_mm is swapper
[ 3.114230] usb 1-1: new high-speed USB device number 2 using ehci-platform
[ 3.118418] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
[ 3.118426] Modules linked in:
[ 3.134717] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
6.3.0-rc3-next-20230321 #1
[ 3.142126] Hardware name: ARM Juno development board (r2) (DT)
[ 3.148052] pstate: 000000c5 (nzcv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[ 3.155027] pc : string (lib/vsprintf.c:644 lib/vsprintf.c:726)
[ 3.158443] lr : vsnprintf (lib/vsprintf.c:2817)
[ 3.162196] sp : ffff80000b34b8f0
[ 3.165511] x29: ffff80000b34b8f0 x28: ffff800009d7b1ca x27: ffff80000b34bab0
[ 3.172666] x26: ffff800009d7b1ca x25: 0000000000000020 x24: 0000000000000008
[ 3.179820] x23: 00000000ffffffd8 x22: ffff8000099858a0 x21: ffff80000b34bc30
[ 3.186973] x20: ffff80000b34ba90 x19: ffff80000b34ba98 x18: 000000003c98bfdd
[ 3.194127] x17: 000000000000001c x16: 00000000eec48da2 x15: 00000000a9dbdd17
[ 3.201280] x14: ffff80000b0242e8 x13: 0000000057a049ef x12: 00000000cfa47237
[ 3.208433] x11: 0000000000000001 x10: 00000000bd5b8780 x9 : ffff80000812cbec
[ 3.215586] x8 : 00000000ffffffff x7 : 0000000000000002 x6 : ffff80000b34ba98
[ 3.222739] x5 : ffffffffffffffff x4 : 0000000000000000 x3 : ffff0a00ffffff04
[ 3.229891] x2 : 0000000000001000 x1 : 0000000000000000 x0 : ffff80000b34bab0
[ 3.237044] Call trace:
[ 3.239489] string (lib/vsprintf.c:644 lib/vsprintf.c:726)
[ 3.242551] vsnprintf (lib/vsprintf.c:2817)
[ 3.245954] vprintk_store (kernel/printk/printk.c:2200)
[ 3.249712] vprintk_emit (kernel/printk/printk.c:2297)
[ 3.253381] vprintk_default (kernel/printk/printk.c:2328)
[ 3.257137] vprintk (kernel/printk/printk_safe.c:50)
[ 3.260198] _printk (kernel/printk/printk.c:2341)
[ 3.263257] sysctl_err (fs/proc/proc_sysctl.c:1109)
[ 3.266577] __register_sysctl_table (fs/proc/proc_sysctl.c:1140
fs/proc/proc_sysctl.c:1383)
[ 3.271202] __register_sysctl_init (fs/proc/proc_sysctl.c:1462)
[ 3.275569] memory_failure_sysctl_init (mm/memory-failure.c:152)
[ 3.280281] do_one_initcall (init/main.c:1306)
[ 3.280453] hub 1-1:1.0: USB hub found
[ 3.284123] kernel_init_freeable (init/main.c:1378 init/main.c:1395
init/main.c:1414 init/main.c:1634)
[ 3.284133] kernel_init (init/main.c:1526)
[ 3.288246] hub 1-1:1.0: 4 ports detected
[ 3.292239] ret_from_fork (arch/arm64/kernel/entry.S:871)
[ 3.292250] Code: 91000400 110004e1 eb08009f 540000c0 (38646845)
All code
========
0:* 00 04 00 add %al,(%rax,%rax,1) <-- trapping instruction
3: 91 xchg %eax,%ecx
4: e1 04 loope 0xa
6: 00 11 add %dl,(%rcx)
8: 9f lahf
9: 00 08 add %cl,(%rax)
b: eb c0 jmp 0xffffffffffffffcd
d: 00 00 add %al,(%rax)
f: 54 push %rsp
10: 45 rex.RB
11: 68 .byte 0x68
12: 64 fs
13: 38 .byte 0x38
Code starting with the faulting instruction
===========================================
0: 45 rex.RB
1: 68 .byte 0x68
2: 64 fs
3: 38 .byte 0x38
[ 3.309416] ---[ end trace 0000000000000000 ]---
[ 3.314037] note: swapper/0[1] exited with irqs disabled
[ 3.319392] Kernel panic - not syncing: Attempted to kill init!
exitcode=0x0000000b
[ 3.327060] SMP: stopping secondary CPUs
[ 3.331189] Kernel Offset: disabled
[ 3.334678] CPU features: 0x400002,0c3c0400,0000421b
[ 3.339649] Memory Limit: none
[ 3.342708] ---[ end Kernel panic - not syncing: Attempted to kill
init! exitcode=0x0000000b ]---
links:
----
- https://qa-reports.linaro.org/lkft/linux-next-master-sanity/build/next-20230321/testrun/15765295/suite/log-parser-test/tests/
- https://qa-reports.linaro.org/lkft/linux-next-master-sanity/build/next-20230321/testrun/15765168/suite/log-parser-test/tests/
- https://qa-reports.linaro.org/lkft/linux-next-master-sanity/build/next-20230321/testrun/15765168/suite/log-parser-test/test/check-kernel-panic/details/
metadata:
git_ref: master
git_repo: https://gitlab.com/Linaro/lkft/mirrors/next/linux-next
git_sha: f3594f0204b756638267242e26d9de611435c3ba
git_describe: next-20230321
kernel_version: 6.3.0-rc3
kernel-config:
https://storage.tuxsuite.com/public/linaro/lkft/builds/2NJ83YzjXEfP00u1LgnjMjjZfxo/config
build-url: https://gitlab.com/Linaro/lkft/mirrors/next/linux-next/-/pipelines/812627412/
artifact-location:
https://storage.tuxsuite.com/public/linaro/lkft/builds/2NJ83YzjXEfP00u1LgnjMjjZfxo
toolchain: gcc-11
--
Linaro LKFT
https://lkft.linaro.org
Hi,
I add Luis into Cc.
On Tue 2023-03-21 11:42:56, Naresh Kamboju wrote:
> The following kernel crash was noticed on arm64 Juno-r2 and Raspberry Pi 4
> Model B on Linux next-20230321.
>
> Reported-by: Linux Kernel Functional Testing <[email protected]>
>
> log:
> ----
> [ 3.071500] Unable to handle kernel paging request at virtual
> address 0000000000001000
I guess that this is exactly PAGE_SIZE (4k).
> [ 3.079432] Mem abort info:
> [ 3.082225] ESR = 0x0000000096000004
> [ 3.085977] EC = 0x25: DABT (current EL), IL = 32 bits
> [ 3.091295] SET = 0, FnV = 0
> [ 3.094350] EA = 0, S1PTW = 0
> [ 3.097491] FSC = 0x04: level 0 translation fault
> [ 3.102373] Data abort info:
> [ 3.105252] ISV = 0, ISS = 0x00000004
> [ 3.109089] CM = 0, WnR = 0
> [ 3.112055] [0000000000001000] user address but active_mm is swapper
> [ 3.114230] usb 1-1: new high-speed USB device number 2 using ehci-platform
> [ 3.118418] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
> [ 3.118426] Modules linked in:
> [ 3.134717] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> 6.3.0-rc3-next-20230321 #1
> [ 3.142126] Hardware name: ARM Juno development board (r2) (DT)
> [ 3.148052] pstate: 000000c5 (nzcv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> [ 3.155027] pc : string (lib/vsprintf.c:644 lib/vsprintf.c:726)
> [ 3.158443] lr : vsnprintf (lib/vsprintf.c:2817)
> [ 3.162196] sp : ffff80000b34b8f0
> [ 3.165511] x29: ffff80000b34b8f0 x28: ffff800009d7b1ca x27: ffff80000b34bab0
> [ 3.172666] x26: ffff800009d7b1ca x25: 0000000000000020 x24: 0000000000000008
> [ 3.179820] x23: 00000000ffffffd8 x22: ffff8000099858a0 x21: ffff80000b34bc30
> [ 3.186973] x20: ffff80000b34ba90 x19: ffff80000b34ba98 x18: 000000003c98bfdd
> [ 3.194127] x17: 000000000000001c x16: 00000000eec48da2 x15: 00000000a9dbdd17
> [ 3.201280] x14: ffff80000b0242e8 x13: 0000000057a049ef x12: 00000000cfa47237
> [ 3.208433] x11: 0000000000000001 x10: 00000000bd5b8780 x9 : ffff80000812cbec
> [ 3.215586] x8 : 00000000ffffffff x7 : 0000000000000002 x6 : ffff80000b34ba98
> [ 3.222739] x5 : ffffffffffffffff x4 : 0000000000000000 x3 : ffff0a00ffffff04
> [ 3.229891] x2 : 0000000000001000 x1 : 0000000000000000 x0 : ffff80000b34bab0
> [ 3.237044] Call trace:
> [ 3.239489] string (lib/vsprintf.c:644 lib/vsprintf.c:726)
This is the line where vsprintf() reads a string to be printed:
static char *string_nocheck(char *buf, char *end, const char *s,
struct printf_spec spec)
{
int len = 0;
int lim = spec.precision;
while (lim--) {
----> char c = *s++;
> [ 3.242551] vsnprintf (lib/vsprintf.c:2817)
> [ 3.245954] vprintk_store (kernel/printk/printk.c:2200)
> [ 3.249712] vprintk_emit (kernel/printk/printk.c:2297)
> [ 3.253381] vprintk_default (kernel/printk/printk.c:2328)
> [ 3.257137] vprintk (kernel/printk/printk_safe.c:50)
> [ 3.260198] _printk (kernel/printk/printk.c:2341)
> [ 3.263257] sysctl_err (fs/proc/proc_sysctl.c:1109)
> [ 3.266577] __register_sysctl_table (fs/proc/proc_sysctl.c:1140
> fs/proc/proc_sysctl.c:1383)
> [ 3.271202] __register_sysctl_init (fs/proc/proc_sysctl.c:1462)
The has been done some reractoring/modification of this code
by the patchset ("[PATCH 00/11] sysctl: deprecate
register_sysctl_paths()"), see
https://lore.kernel.org/lkml/[email protected]/
Luis, does it trigger any bell, please?
Do you have an idea where this code could pass a pointer PAGE_SIZE
as string to printk("%s")?
Best Regards,
Petr
> [ 3.275569] memory_failure_sysctl_init (mm/memory-failure.c:152)
> [ 3.280281] do_one_initcall (init/main.c:1306)
> [ 3.280453] hub 1-1:1.0: USB hub found
> [ 3.284123] kernel_init_freeable (init/main.c:1378 init/main.c:1395
> init/main.c:1414 init/main.c:1634)
> [ 3.284133] kernel_init (init/main.c:1526)
> [ 3.288246] hub 1-1:1.0: 4 ports detected
> [ 3.292239] ret_from_fork (arch/arm64/kernel/entry.S:871)
> [ 3.292250] Code: 91000400 110004e1 eb08009f 540000c0 (38646845)
> All code
> ========
> 0:* 00 04 00 add %al,(%rax,%rax,1) <-- trapping instruction
> 3: 91 xchg %eax,%ecx
> 4: e1 04 loope 0xa
> 6: 00 11 add %dl,(%rcx)
> 8: 9f lahf
> 9: 00 08 add %cl,(%rax)
> b: eb c0 jmp 0xffffffffffffffcd
> d: 00 00 add %al,(%rax)
> f: 54 push %rsp
> 10: 45 rex.RB
> 11: 68 .byte 0x68
> 12: 64 fs
> 13: 38 .byte 0x38
>
> Code starting with the faulting instruction
> ===========================================
> 0: 45 rex.RB
> 1: 68 .byte 0x68
> 2: 64 fs
> 3: 38 .byte 0x38
> [ 3.309416] ---[ end trace 0000000000000000 ]---
> [ 3.314037] note: swapper/0[1] exited with irqs disabled
> [ 3.319392] Kernel panic - not syncing: Attempted to kill init!
> exitcode=0x0000000b
> [ 3.327060] SMP: stopping secondary CPUs
> [ 3.331189] Kernel Offset: disabled
> [ 3.334678] CPU features: 0x400002,0c3c0400,0000421b
> [ 3.339649] Memory Limit: none
> [ 3.342708] ---[ end Kernel panic - not syncing: Attempted to kill
> init! exitcode=0x0000000b ]---
>
> links:
> ----
> - https://qa-reports.linaro.org/lkft/linux-next-master-sanity/build/next-20230321/testrun/15765295/suite/log-parser-test/tests/
> - https://qa-reports.linaro.org/lkft/linux-next-master-sanity/build/next-20230321/testrun/15765168/suite/log-parser-test/tests/
> - https://qa-reports.linaro.org/lkft/linux-next-master-sanity/build/next-20230321/testrun/15765168/suite/log-parser-test/test/check-kernel-panic/details/
>
> metadata:
> git_ref: master
> git_repo: https://gitlab.com/Linaro/lkft/mirrors/next/linux-next
> git_sha: f3594f0204b756638267242e26d9de611435c3ba
> git_describe: next-20230321
> kernel_version: 6.3.0-rc3
> kernel-config:
> https://storage.tuxsuite.com/public/linaro/lkft/builds/2NJ83YzjXEfP00u1LgnjMjjZfxo/config
> build-url: https://gitlab.com/Linaro/lkft/mirrors/next/linux-next/-/pipelines/812627412/
> artifact-location:
> https://storage.tuxsuite.com/public/linaro/lkft/builds/2NJ83YzjXEfP00u1LgnjMjjZfxo
> toolchain: gcc-11
>
>
> --
> Linaro LKFT
> https://lkft.linaro.org
On Tue, Mar 21, 2023 at 10:27:02AM +0100, Petr Mladek wrote:
> Hi,
>
> I add Luis into Cc.
>
> On Tue 2023-03-21 11:42:56, Naresh Kamboju wrote:
> > The following kernel crash was noticed on arm64 Juno-r2 and Raspberry Pi 4
> > Model B on Linux next-20230321.
> >
> > Reported-by: Linux Kernel Functional Testing <[email protected]>
> >
> > log:
> > ----
> > [ 3.071500] Unable to handle kernel paging request at virtual
> > address 0000000000001000
>
> I guess that this is exactly PAGE_SIZE (4k).
>
> > [ 3.079432] Mem abort info:
> > [ 3.082225] ESR = 0x0000000096000004
> > [ 3.085977] EC = 0x25: DABT (current EL), IL = 32 bits
> > [ 3.091295] SET = 0, FnV = 0
> > [ 3.094350] EA = 0, S1PTW = 0
> > [ 3.097491] FSC = 0x04: level 0 translation fault
> > [ 3.102373] Data abort info:
> > [ 3.105252] ISV = 0, ISS = 0x00000004
> > [ 3.109089] CM = 0, WnR = 0
> > [ 3.112055] [0000000000001000] user address but active_mm is swapper
> > [ 3.114230] usb 1-1: new high-speed USB device number 2 using ehci-platform
> > [ 3.118418] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
> > [ 3.118426] Modules linked in:
> > [ 3.134717] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> > 6.3.0-rc3-next-20230321 #1
> > [ 3.142126] Hardware name: ARM Juno development board (r2) (DT)
> > [ 3.148052] pstate: 000000c5 (nzcv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> > [ 3.155027] pc : string (lib/vsprintf.c:644 lib/vsprintf.c:726)
> > [ 3.158443] lr : vsnprintf (lib/vsprintf.c:2817)
> > [ 3.162196] sp : ffff80000b34b8f0
> > [ 3.165511] x29: ffff80000b34b8f0 x28: ffff800009d7b1ca x27: ffff80000b34bab0
> > [ 3.172666] x26: ffff800009d7b1ca x25: 0000000000000020 x24: 0000000000000008
> > [ 3.179820] x23: 00000000ffffffd8 x22: ffff8000099858a0 x21: ffff80000b34bc30
> > [ 3.186973] x20: ffff80000b34ba90 x19: ffff80000b34ba98 x18: 000000003c98bfdd
> > [ 3.194127] x17: 000000000000001c x16: 00000000eec48da2 x15: 00000000a9dbdd17
> > [ 3.201280] x14: ffff80000b0242e8 x13: 0000000057a049ef x12: 00000000cfa47237
> > [ 3.208433] x11: 0000000000000001 x10: 00000000bd5b8780 x9 : ffff80000812cbec
> > [ 3.215586] x8 : 00000000ffffffff x7 : 0000000000000002 x6 : ffff80000b34ba98
> > [ 3.222739] x5 : ffffffffffffffff x4 : 0000000000000000 x3 : ffff0a00ffffff04
> > [ 3.229891] x2 : 0000000000001000 x1 : 0000000000000000 x0 : ffff80000b34bab0
> > [ 3.237044] Call trace:
> > [ 3.239489] string (lib/vsprintf.c:644 lib/vsprintf.c:726)
>
> This is the line where vsprintf() reads a string to be printed:
>
> static char *string_nocheck(char *buf, char *end, const char *s,
> struct printf_spec spec)
> {
> int len = 0;
> int lim = spec.precision;
>
> while (lim--) {
> ----> char c = *s++;
>
>
> > [ 3.242551] vsnprintf (lib/vsprintf.c:2817)
> > [ 3.245954] vprintk_store (kernel/printk/printk.c:2200)
> > [ 3.249712] vprintk_emit (kernel/printk/printk.c:2297)
> > [ 3.253381] vprintk_default (kernel/printk/printk.c:2328)
> > [ 3.257137] vprintk (kernel/printk/printk_safe.c:50)
> > [ 3.260198] _printk (kernel/printk/printk.c:2341)
> > [ 3.263257] sysctl_err (fs/proc/proc_sysctl.c:1109)
> > [ 3.266577] __register_sysctl_table (fs/proc/proc_sysctl.c:1140
> > fs/proc/proc_sysctl.c:1383)
> > [ 3.271202] __register_sysctl_init (fs/proc/proc_sysctl.c:1462)
>
> The has been done some reractoring/modification of this code
> by the patchset ("[PATCH 00/11] sysctl: deprecate
> register_sysctl_paths()"), see
> https://lore.kernel.org/lkml/[email protected]/
>
> Luis, does it trigger any bell, please?
> Do you have an idea where this code could pass a pointer PAGE_SIZE
> as string to printk("%s")?
My bisect actually landed on commit cfe7e6ea5ee2 ("mm: memory-failure:
Move memory failure sysctls to its own file"). The new sysctl table is
missing a sentinel. The following diff resolves it for me.
diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 6367714af61d..b2377f12f062 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -143,6 +143,7 @@ static struct ctl_table memory_failure_table[] = {
.extra1 = SYSCTL_ZERO,
.extra2 = SYSCTL_ONE,
},
+ { }
};
static int __init memory_failure_sysctl_init(void)
Cheers,
Nathan
# bad: [f3594f0204b756638267242e26d9de611435c3ba] Add linux-next specific files for 20230321
# good: [7d31677bb7b1944ac89e9155110dc1b9acbb3895] gpu: host1x: fix uninitialized variable use
git bisect start 'f3594f0204b756638267242e26d9de611435c3ba' '7d31677bb7b1944ac89e9155110dc1b9acbb3895'
# good: [454faa8515d7c7b90b3e291bc7cc1a84697a23d9] Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/cryptodev-2.6.git
git bisect good 454faa8515d7c7b90b3e291bc7cc1a84697a23d9
# good: [9bfa51aa7b4f87fe6215a200cca280ebd1c111b4] Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git
git bisect good 9bfa51aa7b4f87fe6215a200cca280ebd1c111b4
# good: [c637c9b5d4540e07ed95698a67770c5fd2241d7e] Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/phy/linux-phy.git
git bisect good c637c9b5d4540e07ed95698a67770c5fd2241d7e
# good: [f1fb10a55d81bc43b7c16e96122c8611e0131ca2] Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/andy/linux-gpio-intel.git
git bisect good f1fb10a55d81bc43b7c16e96122c8611e0131ca2
# good: [d3294413e4518ef8b31ee46535c873ef0b8a1d1a] Merge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/livepatching/livepatching
git bisect good d3294413e4518ef8b31ee46535c873ef0b8a1d1a
# good: [009e0171be8db0a77624899623fadcb59fc40a12] Merge branch 'main' of git://git.infradead.org/users/willy/xarray.git
git bisect good 009e0171be8db0a77624899623fadcb59fc40a12
# bad: [cfe7e6ea5ee2149cf2dfc69d3f8a220ab00cb65d] mm: memory-failure: Move memory failure sysctls to its own file
git bisect bad cfe7e6ea5ee2149cf2dfc69d3f8a220ab00cb65d
# good: [2e88a8303d02ff107fd6aa5041f9def177483956] proc_sysctl: enhance documentation
git bisect good 2e88a8303d02ff107fd6aa5041f9def177483956
# good: [8869a82dcbbf6fbe584ffff16b620b68281b3180] ntfs: simplfy one-level sysctl registration for ntfs_sysctls
git bisect good 8869a82dcbbf6fbe584ffff16b620b68281b3180
# good: [7385b7cb60e030f0041c8c83d17d3c63e598c77d] ppc: simplify one-level sysctl registration for nmi_wd_lpm_factor_ctl_table
git bisect good 7385b7cb60e030f0041c8c83d17d3c63e598c77d
# good: [36657db1c77c1539812cc1303119cf4ad2e8f34a] x86: simplify one-level sysctl registration for itmt_kern_table
git bisect good 36657db1c77c1539812cc1303119cf4ad2e8f34a
# good: [0234a6faf178985c74806e33bde8a9e3052d3555] arm: simplify two-level sysctl registration for ctl_isa_vars
git bisect good 0234a6faf178985c74806e33bde8a9e3052d3555
# first bad commit: [cfe7e6ea5ee2149cf2dfc69d3f8a220ab00cb65d] mm: memory-failure: Move memory failure sysctls to its own file
On Tue, Mar 21, 2023 at 08:56:26AM -0700, Nathan Chancellor wrote:
> On Tue, Mar 21, 2023 at 10:27:02AM +0100, Petr Mladek wrote:
> > Hi,
> >
> > I add Luis into Cc.
> >
> > On Tue 2023-03-21 11:42:56, Naresh Kamboju wrote:
> > > The following kernel crash was noticed on arm64 Juno-r2 and Raspberry Pi 4
> > > Model B on Linux next-20230321.
> > >
> > > Reported-by: Linux Kernel Functional Testing <[email protected]>
> > >
> > > log:
> > > ----
> > > [ 3.071500] Unable to handle kernel paging request at virtual
> > > address 0000000000001000
> >
> > I guess that this is exactly PAGE_SIZE (4k).
> >
> > > [ 3.079432] Mem abort info:
> > > [ 3.082225] ESR = 0x0000000096000004
> > > [ 3.085977] EC = 0x25: DABT (current EL), IL = 32 bits
> > > [ 3.091295] SET = 0, FnV = 0
> > > [ 3.094350] EA = 0, S1PTW = 0
> > > [ 3.097491] FSC = 0x04: level 0 translation fault
> > > [ 3.102373] Data abort info:
> > > [ 3.105252] ISV = 0, ISS = 0x00000004
> > > [ 3.109089] CM = 0, WnR = 0
> > > [ 3.112055] [0000000000001000] user address but active_mm is swapper
> > > [ 3.114230] usb 1-1: new high-speed USB device number 2 using ehci-platform
> > > [ 3.118418] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
> > > [ 3.118426] Modules linked in:
> > > [ 3.134717] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
> > > 6.3.0-rc3-next-20230321 #1
> > > [ 3.142126] Hardware name: ARM Juno development board (r2) (DT)
> > > [ 3.148052] pstate: 000000c5 (nzcv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
> > > [ 3.155027] pc : string (lib/vsprintf.c:644 lib/vsprintf.c:726)
> > > [ 3.158443] lr : vsnprintf (lib/vsprintf.c:2817)
> > > [ 3.162196] sp : ffff80000b34b8f0
> > > [ 3.165511] x29: ffff80000b34b8f0 x28: ffff800009d7b1ca x27: ffff80000b34bab0
> > > [ 3.172666] x26: ffff800009d7b1ca x25: 0000000000000020 x24: 0000000000000008
> > > [ 3.179820] x23: 00000000ffffffd8 x22: ffff8000099858a0 x21: ffff80000b34bc30
> > > [ 3.186973] x20: ffff80000b34ba90 x19: ffff80000b34ba98 x18: 000000003c98bfdd
> > > [ 3.194127] x17: 000000000000001c x16: 00000000eec48da2 x15: 00000000a9dbdd17
> > > [ 3.201280] x14: ffff80000b0242e8 x13: 0000000057a049ef x12: 00000000cfa47237
> > > [ 3.208433] x11: 0000000000000001 x10: 00000000bd5b8780 x9 : ffff80000812cbec
> > > [ 3.215586] x8 : 00000000ffffffff x7 : 0000000000000002 x6 : ffff80000b34ba98
> > > [ 3.222739] x5 : ffffffffffffffff x4 : 0000000000000000 x3 : ffff0a00ffffff04
> > > [ 3.229891] x2 : 0000000000001000 x1 : 0000000000000000 x0 : ffff80000b34bab0
> > > [ 3.237044] Call trace:
> > > [ 3.239489] string (lib/vsprintf.c:644 lib/vsprintf.c:726)
> >
> > This is the line where vsprintf() reads a string to be printed:
> >
> > static char *string_nocheck(char *buf, char *end, const char *s,
> > struct printf_spec spec)
> > {
> > int len = 0;
> > int lim = spec.precision;
> >
> > while (lim--) {
> > ----> char c = *s++;
> >
> >
> > > [ 3.242551] vsnprintf (lib/vsprintf.c:2817)
> > > [ 3.245954] vprintk_store (kernel/printk/printk.c:2200)
> > > [ 3.249712] vprintk_emit (kernel/printk/printk.c:2297)
> > > [ 3.253381] vprintk_default (kernel/printk/printk.c:2328)
> > > [ 3.257137] vprintk (kernel/printk/printk_safe.c:50)
> > > [ 3.260198] _printk (kernel/printk/printk.c:2341)
> > > [ 3.263257] sysctl_err (fs/proc/proc_sysctl.c:1109)
> > > [ 3.266577] __register_sysctl_table (fs/proc/proc_sysctl.c:1140
> > > fs/proc/proc_sysctl.c:1383)
> > > [ 3.271202] __register_sysctl_init (fs/proc/proc_sysctl.c:1462)
> >
> > The has been done some reractoring/modification of this code
> > by the patchset ("[PATCH 00/11] sysctl: deprecate
> > register_sysctl_paths()"), see
> > https://lore.kernel.org/lkml/[email protected]/
> >
> > Luis, does it trigger any bell, please?
> > Do you have an idea where this code could pass a pointer PAGE_SIZE
> > as string to printk("%s")?
>
> My bisect actually landed on commit cfe7e6ea5ee2 ("mm: memory-failure:
> Move memory failure sysctls to its own file"). The new sysctl table is
> missing a sentinel. The following diff resolves it for me.
>
> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
> index 6367714af61d..b2377f12f062 100644
> --- a/mm/memory-failure.c
> +++ b/mm/memory-failure.c
> @@ -143,6 +143,7 @@ static struct ctl_table memory_failure_table[] = {
> .extra1 = SYSCTL_ZERO,
> .extra2 = SYSCTL_ONE,
> },
> + { }
> };
Thanks pushed this fix in to sysctl-next.
Luis
On 2023/3/22 0:59, Luis Chamberlain wrote:
> On Tue, Mar 21, 2023 at 08:56:26AM -0700, Nathan Chancellor wrote:
>> On Tue, Mar 21, 2023 at 10:27:02AM +0100, Petr Mladek wrote:
>>> Hi,
>>>
>>> I add Luis into Cc.
>>>
>>> On Tue 2023-03-21 11:42:56, Naresh Kamboju wrote:
>>>> The following kernel crash was noticed on arm64 Juno-r2 and Raspberry Pi 4
>>>> Model B on Linux next-20230321.
>>>>
>>>> Reported-by: Linux Kernel Functional Testing <[email protected]>
>>>>
>>>> log:
>>>> ----
>>>> [ 3.071500] Unable to handle kernel paging request at virtual
>>>> address 0000000000001000
>>>
>>> I guess that this is exactly PAGE_SIZE (4k).
>>>
>>>> [ 3.079432] Mem abort info:
>>>> [ 3.082225] ESR = 0x0000000096000004
>>>> [ 3.085977] EC = 0x25: DABT (current EL), IL = 32 bits
>>>> [ 3.091295] SET = 0, FnV = 0
>>>> [ 3.094350] EA = 0, S1PTW = 0
>>>> [ 3.097491] FSC = 0x04: level 0 translation fault
>>>> [ 3.102373] Data abort info:
>>>> [ 3.105252] ISV = 0, ISS = 0x00000004
>>>> [ 3.109089] CM = 0, WnR = 0
>>>> [ 3.112055] [0000000000001000] user address but active_mm is swapper
>>>> [ 3.114230] usb 1-1: new high-speed USB device number 2 using ehci-platform
>>>> [ 3.118418] Internal error: Oops: 0000000096000004 [#1] PREEMPT SMP
>>>> [ 3.118426] Modules linked in:
>>>> [ 3.134717] CPU: 1 PID: 1 Comm: swapper/0 Not tainted
>>>> 6.3.0-rc3-next-20230321 #1
>>>> [ 3.142126] Hardware name: ARM Juno development board (r2) (DT)
>>>> [ 3.148052] pstate: 000000c5 (nzcv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
>>>> [ 3.155027] pc : string (lib/vsprintf.c:644 lib/vsprintf.c:726)
>>>> [ 3.158443] lr : vsnprintf (lib/vsprintf.c:2817)
>>>> [ 3.162196] sp : ffff80000b34b8f0
>>>> [ 3.165511] x29: ffff80000b34b8f0 x28: ffff800009d7b1ca x27: ffff80000b34bab0
>>>> [ 3.172666] x26: ffff800009d7b1ca x25: 0000000000000020 x24: 0000000000000008
>>>> [ 3.179820] x23: 00000000ffffffd8 x22: ffff8000099858a0 x21: ffff80000b34bc30
>>>> [ 3.186973] x20: ffff80000b34ba90 x19: ffff80000b34ba98 x18: 000000003c98bfdd
>>>> [ 3.194127] x17: 000000000000001c x16: 00000000eec48da2 x15: 00000000a9dbdd17
>>>> [ 3.201280] x14: ffff80000b0242e8 x13: 0000000057a049ef x12: 00000000cfa47237
>>>> [ 3.208433] x11: 0000000000000001 x10: 00000000bd5b8780 x9 : ffff80000812cbec
>>>> [ 3.215586] x8 : 00000000ffffffff x7 : 0000000000000002 x6 : ffff80000b34ba98
>>>> [ 3.222739] x5 : ffffffffffffffff x4 : 0000000000000000 x3 : ffff0a00ffffff04
>>>> [ 3.229891] x2 : 0000000000001000 x1 : 0000000000000000 x0 : ffff80000b34bab0
>>>> [ 3.237044] Call trace:
>>>> [ 3.239489] string (lib/vsprintf.c:644 lib/vsprintf.c:726)
>>>
>>> This is the line where vsprintf() reads a string to be printed:
>>>
>>> static char *string_nocheck(char *buf, char *end, const char *s,
>>> struct printf_spec spec)
>>> {
>>> int len = 0;
>>> int lim = spec.precision;
>>>
>>> while (lim--) {
>>> ----> char c = *s++;
>>>
>>>
>>>> [ 3.242551] vsnprintf (lib/vsprintf.c:2817)
>>>> [ 3.245954] vprintk_store (kernel/printk/printk.c:2200)
>>>> [ 3.249712] vprintk_emit (kernel/printk/printk.c:2297)
>>>> [ 3.253381] vprintk_default (kernel/printk/printk.c:2328)
>>>> [ 3.257137] vprintk (kernel/printk/printk_safe.c:50)
>>>> [ 3.260198] _printk (kernel/printk/printk.c:2341)
>>>> [ 3.263257] sysctl_err (fs/proc/proc_sysctl.c:1109)
>>>> [ 3.266577] __register_sysctl_table (fs/proc/proc_sysctl.c:1140
>>>> fs/proc/proc_sysctl.c:1383)
>>>> [ 3.271202] __register_sysctl_init (fs/proc/proc_sysctl.c:1462)
>>>
>>> The has been done some reractoring/modification of this code
>>> by the patchset ("[PATCH 00/11] sysctl: deprecate
>>> register_sysctl_paths()"), see
>>> https://lore.kernel.org/lkml/[email protected]/
>>>
>>> Luis, does it trigger any bell, please?
>>> Do you have an idea where this code could pass a pointer PAGE_SIZE
>>> as string to printk("%s")?
>>
>> My bisect actually landed on commit cfe7e6ea5ee2 ("mm: memory-failure:
>> Move memory failure sysctls to its own file"). The new sysctl table is
>> missing a sentinel. The following diff resolves it for me.
>>
>> diff --git a/mm/memory-failure.c b/mm/memory-failure.c
>> index 6367714af61d..b2377f12f062 100644
>> --- a/mm/memory-failure.c
>> +++ b/mm/memory-failure.c
>> @@ -143,6 +143,7 @@ static struct ctl_table memory_failure_table[] = {
>> .extra1 = SYSCTL_ZERO,
>> .extra2 = SYSCTL_ONE,
>> },
>> + { }
>> };
>
Sorry, I don't know why this is removed after v1[1] when resend :(
thanks for your fix.
[1] https://lkml.org/lkml/2023/3/8/1644
> Thanks pushed this fix in to sysctl-next.
>
> Luis