powerpc/pseries: hwpoison the pages upon hitting UE

Add support to hwpoison the pages upon hitting machine check
exception.

This patch queues the address where UE is hit to percpu array
and schedules work to plumb it into memory poison infrastructure.

Reviewed-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: Ganesh Goudar <ganeshgr@linux.ibm.com>
[mpe: Combine #ifdefs, drop PPC_BIT8(), and empty inline stub]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

3 files changed, 85 insertions, 1 deletions

diff --git a/arch/powerpc/include/asm/mce.h b/arch/powerpc/include/asm/mce.h
index 17996bc9382b..ad47fa865324 100644
--- a/arch/powerpc/include/asm/mce.h
+++ b/arch/powerpc/include/asm/mce.h
@@ -210,6 +210,7 @@ extern void release_mce_event(void);
 extern void machine_check_queue_event(void);
 extern void machine_check_print_event_info(struct machine_check_event *evt,
                                            bool user_mode, bool in_guest);
+unsigned long addr_to_pfn(struct pt_regs *regs, unsigned long addr);
 #ifdef CONFIG_PPC_BOOK3S_64
 void flush_and_reload_slb(void);
 #endif /* CONFIG_PPC_BOOK3S_64 */
diff --git a/arch/powerpc/kernel/mce_power.c b/arch/powerpc/kernel/mce_power.c
index 6b800eec31f2..367fbfa2e835 100644
--- a/arch/powerpc/kernel/mce_power.c
+++ b/arch/powerpc/kernel/mce_power.c
@@ -36,7 +36,7 @@
  * Convert an address related to an mm to a PFN. NOTE: we are in real
  * mode, we could potentially race with page table updates.
  */
-static unsigned long addr_to_pfn(struct pt_regs *regs, unsigned long addr)
+unsigned long addr_to_pfn(struct pt_regs *regs, unsigned long addr)
 {
         pte_t *ptep;
         unsigned long flags;
diff --git a/arch/powerpc/platforms/pseries/ras.c b/arch/powerpc/platforms/pseries/ras.c
index a25c2ac0c9c0..c97d15352f9f 100644
--- a/arch/powerpc/platforms/pseries/ras.c
+++ b/arch/powerpc/platforms/pseries/ras.c
@@ -707,6 +707,87 @@ out:
         return disposition;
 }
 
+#ifdef CONFIG_MEMORY_FAILURE
+
+static DEFINE_PER_CPU(int, rtas_ue_count);
+static DEFINE_PER_CPU(unsigned long, rtas_ue_paddr[MAX_MC_EVT]);
+
+#define UE_EFFECTIVE_ADDR_PROVIDED      0x40
+#define UE_LOGICAL_ADDR_PROVIDED        0x20
+
+
+static void pseries_hwpoison_work_fn(struct work_struct *work)
+{
+        unsigned long paddr;
+        int index;
+
+        while (__this_cpu_read(rtas_ue_count) > 0) {
+                index = __this_cpu_read(rtas_ue_count) - 1;
+                paddr = __this_cpu_read(rtas_ue_paddr[index]);
+                memory_failure(paddr >> PAGE_SHIFT, 0);
+                __this_cpu_dec(rtas_ue_count);
+        }
+}
+
+static DECLARE_WORK(hwpoison_work, pseries_hwpoison_work_fn);
+
+static void queue_ue_paddr(unsigned long paddr)
+{
+        int index;
+
+        index = __this_cpu_inc_return(rtas_ue_count) - 1;
+        if (index >= MAX_MC_EVT) {
+                __this_cpu_dec(rtas_ue_count);
+                return;
+        }
+        this_cpu_write(rtas_ue_paddr[index], paddr);
+        schedule_work(&hwpoison_work);
+}
+
+static void pseries_do_memory_failure(struct pt_regs *regs,
+                                      struct pseries_mc_errorlog *mce_log)
+{
+        unsigned long paddr;
+
+        if (mce_log->sub_err_type & UE_LOGICAL_ADDR_PROVIDED) {
+                paddr = be64_to_cpu(mce_log->logical_address);
+        } else if (mce_log->sub_err_type & UE_EFFECTIVE_ADDR_PROVIDED) {
+                unsigned long pfn;
+
+                pfn = addr_to_pfn(regs,
+                                  be64_to_cpu(mce_log->effective_address));
+                if (pfn == ULONG_MAX)
+                        return;
+                paddr = pfn << PAGE_SHIFT;
+        } else {
+                return;
+        }
+        queue_ue_paddr(paddr);
+}
+
+static void pseries_process_ue(struct pt_regs *regs,
+                               struct rtas_error_log *errp)
+{
+        struct pseries_errorlog *pseries_log;
+        struct pseries_mc_errorlog *mce_log;
+
+        if (!rtas_error_extended(errp))
+                return;
+
+        pseries_log = get_pseries_errorlog(errp, PSERIES_ELOG_SECT_ID_MCE);
+        if (!pseries_log)
+                return;
+
+        mce_log = (struct pseries_mc_errorlog *)pseries_log->data;
+
+        if (mce_log->error_type == MC_ERROR_TYPE_UE)
+                pseries_do_memory_failure(regs, mce_log);
+}
+#else
+static inline void pseries_process_ue(struct pt_regs *regs,
+                                      struct rtas_error_log *errp) { }
+#endif /*CONFIG_MEMORY_FAILURE */
+
 /*
  * Process MCE rtas errlog event.
  */
@@ -765,6 +846,8 @@ static int recover_mce(struct pt_regs *regs, struct rtas_error_log *err)
                 recovered = 1;
         }
 
+        pseries_process_ue(regs, err);
+
         /* Queue irq work to log this rtas event later. */
         irq_work_queue(&mce_errlog_process_work);