Le 20/12/2025 à 12:17, Christian Zigotzky a écrit :
On 19/12/25 13:23, Christophe Leroy (CS GROUP) wrote:
Commit 2997876c4a1a ("powerpc/32: Restore clearing of MSR[RI] at
interrupt/syscall exit") delayed clearing of MSR[RI], but missed that
both MSR[RI] and MSR[EE] are cleared at the same time, so the commit
also delayed the disabling of interrupts, leading to unexpected
behaviour.

To fix that, mostly revert the blamed commit and restore the clearing
of MSR[RI] in interrupt_exit_kernel_prepare() instead. For 8xx it
implies adding a synchronising instruction after the mtspr in order to
make sure no instruction counter interrupt (used for perf events) will
fire just after clearing MSR[RI].

Reported-by: Christian Zigotzky <[email protected]>
Closes: https://eur01.safelinks.protection.outlook.com/? url=https%3A%2F%2Flore.kernel.org%2Fall%2F4d0bd05d-6158-1323-3509-744d3fbe8fc7%40xenosoft.de%2F&data=05%7C02%7Cchristophe.leroy2%40cs-soprasteria.com%7C04147486078f4487431908de3fb987ff%7C8b87af7d86474dc78df45f69a2011bb5%7C0%7C0%7C639018263263978487%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=oNkrme1a69HcSK3wzCMc%2BJhc6GnkS2VHoGfQFM242%2BA%3D&reserved=0
Reported-by: Guenter Roeck <[email protected]>
Closes: https://eur01.safelinks.protection.outlook.com/? url=https%3A%2F%2Flore.kernel.org%2Fall%2F6b05eb1c- fdef-44e0-91a7-8286825e68f1%40roeck- us.net%2F&data=05%7C02%7Cchristophe.leroy2%40cs- soprasteria.com%7C04147486078f4487431908de3fb987ff%7C8b87af7d86474dc78df45f69a2011bb5%7C0%7C0%7C639018263263994543%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=q8cMycJtK%2B2nZLB2nx4Ho8SuDD7ZjEbGNb67IULrOCQ%3D&reserved=0 Fixes: 2997876c4a1a ("powerpc/32: Restore clearing of MSR[RI] at interrupt/syscall exit")
Signed-off-by: Christophe Leroy (CS GROUP) <[email protected]>
---
  arch/powerpc/include/asm/hw_irq.h |  2 +-
  arch/powerpc/include/asm/reg.h    |  1 +
  arch/powerpc/kernel/entry_32.S    | 15 ---------------
  arch/powerpc/kernel/interrupt.c   |  5 ++++-
  4 files changed, 6 insertions(+), 17 deletions(-)

diff --git a/arch/powerpc/include/asm/hw_irq.h b/arch/powerpc/include/ asm/hw_irq.h
index 1078ba88efaf..9cd945f2acaf 100644
--- a/arch/powerpc/include/asm/hw_irq.h
+++ b/arch/powerpc/include/asm/hw_irq.h
@@ -90,7 +90,7 @@ static inline void __hard_EE_RI_disable(void)
      if (IS_ENABLED(CONFIG_BOOKE))
          wrtee(0);
      else if (IS_ENABLED(CONFIG_PPC_8xx))
-        wrtspr(SPRN_NRI);
+        wrtspr_sync(SPRN_NRI);
      else if (IS_ENABLED(CONFIG_PPC_BOOK3S_64))
          __mtmsrd(0, 1);
      else
diff --git a/arch/powerpc/include/asm/reg.h b/arch/powerpc/include/ asm/reg.h
index 3fe186635432..3449dd2b577d 100644
--- a/arch/powerpc/include/asm/reg.h
+++ b/arch/powerpc/include/asm/reg.h
@@ -1400,6 +1400,7 @@ static inline void mtmsr_isync(unsigned long val)
                       : "r" ((unsigned long)(v)) \
                       : "memory")
  #define wrtspr(rn)    asm volatile("mtspr " __stringify(rn) ",2" : : : "memory") +#define wrtspr_sync(rn)    asm volatile("mtspr " __stringify(rn) ",2; sync" : : : "memory")
  static inline void wrtee(unsigned long val)
  {
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/ entry_32.S
index 16f8ee6cb2cd..d8426251b1cd 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -101,17 +101,6 @@ SYM_FUNC_END(__kuep_unlock)
  .endm
  #endif
-.macro    clr_ri trash
-#ifndef CONFIG_BOOKE
-#ifdef CONFIG_PPC_8xx
-    mtspr   SPRN_NRI, \trash
-#else
-    li    \trash, MSR_KERNEL & ~MSR_RI
-    mtmsr    \trash
-#endif
-#endif
-.endm
-
      .globl    transfer_to_syscall
  transfer_to_syscall:
      stw    r3, ORIG_GPR3(r1)
@@ -160,7 +149,6 @@ ret_from_syscall:
      cmpwi    r3,0
      REST_GPR(3, r1)
  syscall_exit_finish:
-    clr_ri    r4
      mtspr    SPRN_SRR0,r7
      mtspr    SPRN_SRR1,r8
@@ -237,7 +225,6 @@ fast_exception_return:
      /* Clear the exception marker on the stack to avoid confusing stacktrace */
      li    r10, 0
      stw    r10, 8(r11)
-    clr_ri    r10
      mtspr    SPRN_SRR1,r9
      mtspr    SPRN_SRR0,r12
      REST_GPR(9, r11)
@@ -270,7 +257,6 @@ interrupt_return:
  .Lfast_user_interrupt_return:
      lwz    r11,_NIP(r1)
      lwz    r12,_MSR(r1)
-    clr_ri    r4
      mtspr    SPRN_SRR0,r11
      mtspr    SPRN_SRR1,r12
@@ -313,7 +299,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_NEED_PAIRED_STWCX)
      cmpwi    cr1,r3,0
      lwz    r11,_NIP(r1)
      lwz    r12,_MSR(r1)
-    clr_ri    r4
      mtspr    SPRN_SRR0,r11
      mtspr    SPRN_SRR1,r12
diff --git a/arch/powerpc/kernel/interrupt.c b/arch/powerpc/kernel/ interrupt.c
index aea6f7e8e9c6..e63bfde13e03 100644
--- a/arch/powerpc/kernel/interrupt.c
+++ b/arch/powerpc/kernel/interrupt.c
@@ -38,7 +38,7 @@ static inline bool exit_must_hard_disable(void)
  #else
  static inline bool exit_must_hard_disable(void)
  {
-    return false;
+    return true;
  }
  #endif
@@ -443,6 +443,9 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs)
          if (unlikely(stack_store))
              __hard_EE_RI_disable();
+#else
+    } else {
+        __hard_EE_RI_disable();
  #endif /* CONFIG_PPC64 */
      }

The patched kernel 6.19.0-rc1 boots without any problems. Thank you.

Tested-by Christian Zigotzky <[email protected]>


Thanks

b4/patchwork missed your tested-by which lacks the colon.

Tested-by: Christian Zigotzky <[email protected]>

Reply via email to