[PATCH] powerpc/8xx: Fix Oops with STRICT_KERNEL_RWX without DEBUG_RODATA_TEST

[PATCH] powerpc/8xx: Fix Oops with STRICT_KERNEL_RWX without DEBUG_RODATA_TEST

[Christophe Leroy]

Until now, all tests involving CONFIG_STRICT_KERNEL_RWX were done with
DEBUG_RODATA_TEST to check the result. But now that
CONFIG_STRICT_KERNEL_RWX is selected by default, it came without
CONFIG_DEBUG_RODATA_TEST and led to the following Oops

[    6.830908] Freeing unused kernel image (initmem) memory: 352K
[    6.840077] BUG: Unable to handle kernel data access on write at 0xc1285200
[    6.846836] Faulting instruction address: 0xc0004b6c
[    6.851745] Oops: Kernel access of bad area, sig: 11 [#1]
[    6.857075] BE PAGE_SIZE=16K PREEMPT CMPC885
[    6.861348] SAF3000 DIE NOTIFICATION
[    6.864830] CPU: 0 PID: 1 Comm: swapper Not tainted 5.15.0-rc5-s3k-dev-02255-g2747d7b7916f #451
[    6.873429] NIP:  c0004b6c LR: c0004b60 CTR: 00000000
[    6.878419] REGS: c902be60 TRAP: 0300   Not tainted  (5.15.0-rc5-s3k-dev-02255-g2747d7b7916f)
[    6.886852] MSR:  00009032 <EE,ME,IR,DR,RI>  CR: 53000335  XER: 8000ff40
[    6.893564] DAR: c1285200 DSISR: 82000000
[    6.893564] GPR00: 0c000000 c902bf20 c20f4000 08000000 00000001 04001f00 c1800000 00000035
[    6.893564] GPR08: ff0001ff c1280000 00000002 c0004b60 00001000 00000000 c0004b1c 00000000
[    6.893564] GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[    6.893564] GPR24: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 c1060000
[    6.932034] NIP [c0004b6c] kernel_init+0x50/0x138
[    6.936682] LR [c0004b60] kernel_init+0x44/0x138
[    6.941245] Call Trace:
[    6.943653] [c902bf20] [c0004b60] kernel_init+0x44/0x138 (unreliable)
[    6.950022] [c902bf30] [c001122c] ret_from_kernel_thread+0x5c/0x64
[    6.956135] Instruction dump:
[    6.959060] 48ffc521 48045469 4800d8cd 3d20c086 89295fa0 2c090000 41820058 480796c9
[    6.966890] 4800e48d 3d20c128 39400002 3fe0c106 <91495200> 3bff8000 4806fa1d 481f7d75
[    6.974902] ---[ end trace 1e397bacba4aa610 ]---

0xc1285200 corresponds to 'system_state' global var that the kernel is trying to set to
SYSTEM_RUNNING. This var is above the RO/RW limit so it shouldn't Oops.

It oopses because the dirty bit is missing.

Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
---
 arch/powerpc/kernel/head_8xx.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/powerpc/kernel/head_8xx.S b/arch/powerpc/kernel/head_8xx.S
index 9bdb95f5694f..2d596881b70e 100644
--- a/arch/powerpc/kernel/head_8xx.S
+++ b/arch/powerpc/kernel/head_8xx.S
@@ -755,7 +755,7 @@ _GLOBAL(mmu_pin_tlb)
 	cmplw	r6, r9
 	bdnzt	lt, 2b
 
-4:	LOAD_REG_IMMEDIATE(r8, 0xf0 | _PAGE_SPS | _PAGE_SH | _PAGE_PRESENT)
+4:	LOAD_REG_IMMEDIATE(r8, 0xf0 | _PAGE_DIRTY | _PAGE_SPS | _PAGE_SH | _PAGE_PRESENT)
 2:	ori	r0, r6, MD_EVALID
 	mtspr	SPRN_MD_CTR, r5
 	mtspr	SPRN_MD_EPN, r0
-- 
2.31.1

Re: [PATCH] powerpc/8xx: Fix Oops with STRICT_KERNEL_RWX without DEBUG_RODATA_TEST

[Michael Ellerman]

On Fri, 29 Oct 2021 17:10:45 +0200, Christophe Leroy wrote:
> Until now, all tests involving CONFIG_STRICT_KERNEL_RWX were done with
> DEBUG_RODATA_TEST to check the result. But now that
> CONFIG_STRICT_KERNEL_RWX is selected by default, it came without
> CONFIG_DEBUG_RODATA_TEST and led to the following Oops
> 
> [    6.830908] Freeing unused kernel image (initmem) memory: 352K
> [    6.840077] BUG: Unable to handle kernel data access on write at 0xc1285200
> [    6.846836] Faulting instruction address: 0xc0004b6c
> [    6.851745] Oops: Kernel access of bad area, sig: 11 [#1]
> [    6.857075] BE PAGE_SIZE=16K PREEMPT CMPC885
> [    6.861348] SAF3000 DIE NOTIFICATION
> [    6.864830] CPU: 0 PID: 1 Comm: swapper Not tainted 5.15.0-rc5-s3k-dev-02255-g2747d7b7916f #451
> [    6.873429] NIP:  c0004b6c LR: c0004b60 CTR: 00000000
> [    6.878419] REGS: c902be60 TRAP: 0300   Not tainted  (5.15.0-rc5-s3k-dev-02255-g2747d7b7916f)
> [    6.886852] MSR:  00009032 <EE,ME,IR,DR,RI>  CR: 53000335  XER: 8000ff40
> [    6.893564] DAR: c1285200 DSISR: 82000000
> [    6.893564] GPR00: 0c000000 c902bf20 c20f4000 08000000 00000001 04001f00 c1800000 00000035
> [    6.893564] GPR08: ff0001ff c1280000 00000002 c0004b60 00001000 00000000 c0004b1c 00000000
> [    6.893564] GPR16: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
> [    6.893564] GPR24: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 c1060000
> [    6.932034] NIP [c0004b6c] kernel_init+0x50/0x138
> [    6.936682] LR [c0004b60] kernel_init+0x44/0x138
> [    6.941245] Call Trace:
> [    6.943653] [c902bf20] [c0004b60] kernel_init+0x44/0x138 (unreliable)
> [    6.950022] [c902bf30] [c001122c] ret_from_kernel_thread+0x5c/0x64
> [    6.956135] Instruction dump:
> [    6.959060] 48ffc521 48045469 4800d8cd 3d20c086 89295fa0 2c090000 41820058 480796c9
> [    6.966890] 4800e48d 3d20c128 39400002 3fe0c106 <91495200> 3bff8000 4806fa1d 481f7d75
> [    6.974902] ---[ end trace 1e397bacba4aa610 ]---
> 
> [...]

Applied to powerpc/next.

[1/1] powerpc/8xx: Fix Oops with STRICT_KERNEL_RWX without DEBUG_RODATA_TEST
      https://git.kernel.org/powerpc/c/c12ab8dbc492b992e1ea717db933cee568780c47

cheers