Hi,
On Thu, Sep 01, 2022 at 11:15:09AM +0100, Richard Henderson wrote:
> The value previously chosen overlaps GUSA_MASK.
>
> Rename all DELAY_SLOT_* and GUSA_* defines to emphasize
> that they are included in TB_FLAGs. Add aliases for the
> FPSCR and SR bits that are included in TB_FLAGS, so that
> we don't accidentally reassign those bits.
>
> Fixes: 4da06fb3062 ("target/sh4: Implement prctl_unalign_sigbus")
> Resolves: https://gitlab.com/qemu-project/qemu/-/issues/856
> Signed-off-by: Richard Henderson <richard.hender...@linaro.org>
I noticed that my sh4 emulations crash randomly with qemu v7.2-rc4.
This happens with all Linux kernel versions. Testing shows that this
patch is responsible. Reverting it fixes the problem.
Some of the symptoms are attached below.
Guenter
---
Symptoms:
- Random crashes, such as
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
CPU: 0 PID: 1 Comm: init Not tainted 5.10.158 #1
Stack: (0x8c821e60 to 0x8c822000)
1e60: 8c436726 00000000 8c5db1fc 8c011a64 8ca7aa80 8c821e9c ab2577ac 8c021fca
1e80: 8c011a64 8c81dde0 00020000 8c81dda0 00000000 0000000b 8c81f8e0 0000000b
1ea0: 8c81f8e0 00000001 00000000 8c81fb9c 00000000 8c821eb0 8c821f5c 8c821fa4
1ec0: 8c81fa5c 8c81fc1c 000000cd 00000000 00000000 00000000 ab2577ac 8c022af8
1ee0: 8c81dda0 8c81dde0 00020000 8c821f5c 8c81dde0 8c81dda0 0000000b 8c02b1e8
1f00: 8c821f5c 400004d8 8c821f48 8c011a64 0000000a 0000000a 8c81ca60 8c012db4
1f20: 29558c9c 00000406 295f9294 8c821fe4 8c57702c 8c821fa4 09000002 8c821f68
1f40: 8c011a64 295f9294 8c02b0d2 29558c9c 00000406 8c57702c 0000000b 0000000b
1f60: 00000000 00000001 00000008 00000000 00000000 00000000 00000000 00000000
1f80: ab2577ac 8c0150f8 29558c9c 00000406 295f9294 00000000 40008000 8c0150ec
1fa0: 8c820000 7bfcfadc ffffffff 00000040 000080f0 cfffffff 00000000 00000000
1fc0: 8c820000 295fae80 0d39ad3d 295fae80 295630ee 295f9294 00000406 29558c9c
1fe0: 7bfcfadc 295af5ac 295af6ea 00008100 295fafbc 00000000 0d39acf0 ffffffff
Call trace:
[<8c436d88>] printk+0x0/0x48
[<8c011a64>] arch_local_irq_restore+0x0/0x24
[<8c021fca>] do_exit+0x8a6/0x8f0
[<8c011a64>] arch_local_irq_restore+0x0/0x24
[<8c022af8>] do_group_exit+0x34/0x90
[<8c02b1e8>] get_signal+0xd8/0x5f8
[<8c011a64>] arch_local_irq_restore+0x0/0x24
[<8c012db4>] do_notify_resume+0x6c/0x54c
[<8c011a64>] arch_local_irq_restore+0x0/0x24
[<8c02b0d2>] force_sig_fault_to_task+0x3a/0x6c
[<8c0150f8>] resume_userspace+0x0/0x10
[<8c0150ec>] ret_from_exception+0x0/0xc
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b^M
^M
CPU: 0 PID: 1 Comm: init Not tainted 4.14.301 #1^M
Stack: (0x8fc19e08 to 0x8fc1a000)^M
...
- Alleged FPU use
BUG: FPU is used in kernel mode.
------------[ cut here ]------------
kernel BUG at arch/sh/kernel/cpu/fpu.c:60!
Kernel BUG: 003e [#1]
Modules linked in:
CPU: 0 PID: 1166 Comm: sh Not tainted 4.14.301-rc2-00084-gdd6fc0ede260 #1
task: 8ff38800 task.stack: 8f40e000
PC is at fpu_state_restore+0x60/0x88
PR is at fpu_state_restore+0x60/0x88
PC : 8c01969c SP : 8fc2be6c SR : 400080f1
TEA : 004382e8
R0 : 00000020 R1 : 8c4f21a4 R2 : 8c4f21a4 R3 : 8c011be8
R4 : 000000f0 R5 : 00000000 R6 : 00000023 R7 : 8c1b97e0
R8 : 8fc2bec0 R9 : 8ff38800 R10 : 8c0196c4 R11 : 00000000
R12 : 8c011be0 R13 : 8ff38800 R14 : 8f40fe24
MACH: 000003de MACL: 00000184 GBR : 295fafbc PR : 8c01969c
Call trace:
[<8c0196d0>] fpu_state_restore_trap_handler+0xc/0x18
[<8c0196c4>] fpu_state_restore_trap_handler+0x0/0x18
[<8c0150ec>] ret_from_exception+0x0/0xc
[<8c0150ec>] ret_from_exception+0x0/0xc
[<8c3cb1dc>] __schedule+0x1bc/0x50c
[<8c011be0>] arch_local_save_flags+0x0/0x8
[<8c017016>] save_fpu+0x16/0x80
[<8c011fd6>] __switch_to+0x5a/0x8c
[<8c3cb1dc>] __schedule+0x1bc/0x50c
[<8c011be0>] arch_local_save_flags+0x0/0x8
...
- Alleged unhandled unaligned access errors in different locations
(varies per run)
Fixing up unaligned userspace access in "S40network" pid=1111 pc=0x0043761e
ins=0x112d
Fixing up unaligned userspace access in "S40network" pid=1111 pc=0x0043761e
ins=0x112d
Sending SIGBUS to "S40network" due to unaligned access (PC 43761e PR 295b6796)
Bus error
Fixing up unaligned userspace access in "sh" pid=1122 pc=0x295b1714 ins=0x1123
Fixing up unaligned userspace access in "sh" pid=1122 pc=0x295b1714 ins=0x1123
Sending SIGBUS to "sh" due to unaligned access (PC 295b1714 PR 295b170c)
Fixing up unaligned userspace access in "klogd" pid=1084 pc=0x295ac464
ins=0x2922
Fixing up unaligned userspace access in "klogd" pid=1084 pc=0x295ac464
ins=0x2922
Sending SIGBUS to "klogd" due to unaligned access (PC 295ac464 PR 295ac45c)
