Merge branch 'linux-2.6' into for-2.6.22

13 files changed, 159 insertions, 112 deletions

diff --git a/arch/powerpc/platforms/52xx/mpc52xx_pic.c b/arch/powerpc/platforms/52xx/mpc52xx_pic.c
index c75192567e55..fbfff95b4437 100644
--- a/arch/powerpc/platforms/52xx/mpc52xx_pic.c
+++ b/arch/powerpc/platforms/52xx/mpc52xx_pic.c
@@ -128,7 +128,7 @@ static void mpc52xx_main_mask(unsigned int virq)
 
         pr_debug("%s: irq=%x. l2=%d\n", __func__, irq, l2irq);
 
-        io_be_setbit(&intr->main_mask, 15 - l2irq);
+        io_be_setbit(&intr->main_mask, 16 - l2irq);
 }
 
 static void mpc52xx_main_unmask(unsigned int virq)
@@ -141,7 +141,7 @@ static void mpc52xx_main_unmask(unsigned int virq)
 
         pr_debug("%s: irq=%x. l2=%d\n", __func__, irq, l2irq);
 
-        io_be_clrbit(&intr->main_mask, 15 - l2irq);
+        io_be_clrbit(&intr->main_mask, 16 - l2irq);
 }
 
 static struct irq_chip mpc52xx_main_irqchip = {
diff --git a/arch/powerpc/platforms/83xx/mpc834x_mds.c b/arch/powerpc/platforms/83xx/mpc834x_mds.c
index e5d819166874..8aa9a93e2aa2 100644
--- a/arch/powerpc/platforms/83xx/mpc834x_mds.c
+++ b/arch/powerpc/platforms/83xx/mpc834x_mds.c
@@ -55,9 +55,9 @@ static int mpc834x_usb_cfg(void)
         struct device_node *np = NULL;
         int port0_is_dr = 0;
 
-        if ((np = of_find_compatible_node(np, "usb", "fsl-usb2-dr")) != NULL)
+        if ((np = of_find_compatible_node(NULL, "usb", "fsl-usb2-dr")) != NULL)
                 port0_is_dr = 1;
-        if ((np = of_find_compatible_node(np, "usb", "fsl-usb2-mph")) != NULL){
+        if ((np = of_find_compatible_node(NULL, "usb", "fsl-usb2-mph")) != NULL){
                 if (port0_is_dr) {
                         printk(KERN_WARNING
                                 "There is only one USB port on PB board! \n");
@@ -103,8 +103,8 @@ static int mpc834x_usb_cfg(void)
                 return -1;
 
         /*
-         * if MDS board is plug into PIB board,
-         * force to use the PHY on MDS board
+         * if Processor Board is plugged into PIB board,
+         * force to use the PHY on Processor Board
          */
         bcsr5 = in_8(bcsr_regs + 5);
         if (!(bcsr5 & BCSR5_INT_USB))
diff --git a/arch/powerpc/platforms/85xx/Kconfig b/arch/powerpc/platforms/85xx/Kconfig
index 124e2c595972..329fcd48669a 100644
--- a/arch/powerpc/platforms/85xx/Kconfig
+++ b/arch/powerpc/platforms/85xx/Kconfig
@@ -53,6 +53,7 @@ config MPC85xx
         bool
         select PPC_UDBG_16550
         select PPC_INDIRECT_PCI
+        select SERIAL_8250_SHARE_IRQ if SERIAL_8250
         default y if MPC8540_ADS || MPC85xx_CDS || MPC8560_ADS \
                 || MPC85xx_MDS || MPC8544_DS
 
diff --git a/arch/powerpc/platforms/cell/spu_base.c b/arch/powerpc/platforms/cell/spu_base.c
index c43999a10deb..eba7a2641dce 100644
--- a/arch/powerpc/platforms/cell/spu_base.c
+++ b/arch/powerpc/platforms/cell/spu_base.c
@@ -38,8 +38,61 @@
 const struct spu_management_ops *spu_management_ops;
 const struct spu_priv1_ops *spu_priv1_ops;
 
+static struct list_head spu_list[MAX_NUMNODES];
+static LIST_HEAD(spu_full_list);
+static DEFINE_MUTEX(spu_mutex);
+static spinlock_t spu_list_lock = SPIN_LOCK_UNLOCKED;
+
 EXPORT_SYMBOL_GPL(spu_priv1_ops);
 
+void spu_invalidate_slbs(struct spu *spu)
+{
+        struct spu_priv2 __iomem *priv2 = spu->priv2;
+
+        if (spu_mfc_sr1_get(spu) & MFC_STATE1_RELOCATE_MASK)
+                out_be64(&priv2->slb_invalidate_all_W, 0UL);
+}
+EXPORT_SYMBOL_GPL(spu_invalidate_slbs);
+
+/* This is called by the MM core when a segment size is changed, to
+ * request a flush of all the SPEs using a given mm
+ */
+void spu_flush_all_slbs(struct mm_struct *mm)
+{
+        struct spu *spu;
+        unsigned long flags;
+
+        spin_lock_irqsave(&spu_list_lock, flags);
+        list_for_each_entry(spu, &spu_full_list, full_list) {
+                if (spu->mm == mm)
+                        spu_invalidate_slbs(spu);
+        }
+        spin_unlock_irqrestore(&spu_list_lock, flags);
+}
+
+/* The hack below stinks... try to do something better one of
+ * these days... Does it even work properly with NR_CPUS == 1 ?
+ */
+static inline void mm_needs_global_tlbie(struct mm_struct *mm)
+{
+        int nr = (NR_CPUS > 1) ? NR_CPUS : NR_CPUS + 1;
+
+        /* Global TLBIE broadcast required with SPEs. */
+        __cpus_setall(&mm->cpu_vm_mask, nr);
+}
+
+void spu_associate_mm(struct spu *spu, struct mm_struct *mm)
+{
+        unsigned long flags;
+
+        spin_lock_irqsave(&spu_list_lock, flags);
+        spu->mm = mm;
+        spin_unlock_irqrestore(&spu_list_lock, flags);
+        if (mm)
+                mm_needs_global_tlbie(mm);
+}
+EXPORT_SYMBOL_GPL(spu_associate_mm);
+
 static int __spu_trap_invalid_dma(struct spu *spu)
 {
         pr_debug("%s\n", __FUNCTION__);
@@ -74,6 +127,7 @@ static int __spu_trap_data_seg(struct spu *spu, unsigned long ea)
         struct spu_priv2 __iomem *priv2 = spu->priv2;
         struct mm_struct *mm = spu->mm;
         u64 esid, vsid, llp;
+        int psize;
 
         pr_debug("%s\n", __FUNCTION__);
 
@@ -90,22 +144,25 @@ static int __spu_trap_data_seg(struct spu *spu, unsigned long ea)
         case USER_REGION_ID:
 #ifdef CONFIG_HUGETLB_PAGE
                 if (in_hugepage_area(mm->context, ea))
-                        llp = mmu_psize_defs[mmu_huge_psize].sllp;
+                        psize = mmu_huge_psize;
                 else
 #endif
-                        llp = mmu_psize_defs[mmu_virtual_psize].sllp;
+                        psize = mm->context.user_psize;
                 vsid = (get_vsid(mm->context.id, ea) << SLB_VSID_SHIFT) |
-                                SLB_VSID_USER | llp;
+                                SLB_VSID_USER;
                 break;
         case VMALLOC_REGION_ID:
-                llp = mmu_psize_defs[mmu_virtual_psize].sllp;
+                if (ea < VMALLOC_END)
+                        psize = mmu_vmalloc_psize;
+                else
+                        psize = mmu_io_psize;
                 vsid = (get_kernel_vsid(ea) << SLB_VSID_SHIFT) |
-                        SLB_VSID_KERNEL | llp;
+                        SLB_VSID_KERNEL;
                 break;
         case KERNEL_REGION_ID:
-                llp = mmu_psize_defs[mmu_linear_psize].sllp;
+                psize = mmu_linear_psize;
                 vsid = (get_kernel_vsid(ea) << SLB_VSID_SHIFT) |
-                        SLB_VSID_KERNEL | llp;
+                        SLB_VSID_KERNEL;
                 break;
         default:
                 /* Future: support kernel segments so that drivers
@@ -114,9 +171,10 @@ static int __spu_trap_data_seg(struct spu *spu, unsigned long ea)
                 pr_debug("invalid region access at %016lx\n", ea);
                 return 1;
         }
+        llp = mmu_psize_defs[psize].sllp;
 
         out_be64(&priv2->slb_index_W, spu->slb_replace);
-        out_be64(&priv2->slb_vsid_RW, vsid);
+        out_be64(&priv2->slb_vsid_RW, vsid | llp);
         out_be64(&priv2->slb_esid_RW, esid);
 
         spu->slb_replace++;
@@ -330,10 +388,6 @@ static void spu_free_irqs(struct spu *spu)
                 free_irq(spu->irqs[2], spu);
 }
 
-static struct list_head spu_list[MAX_NUMNODES];
-static LIST_HEAD(spu_full_list);
-static DEFINE_MUTEX(spu_mutex);
-
 static void spu_init_channels(struct spu *spu)
 {
         static const struct {
@@ -593,6 +647,7 @@ static int __init create_spu(void *data)
         struct spu *spu;
         int ret;
         static int number;
+        unsigned long flags;
 
         ret = -ENOMEM;
         spu = kzalloc(sizeof (*spu), GFP_KERNEL);
@@ -620,8 +675,10 @@ static int __init create_spu(void *data)
                 goto out_free_irqs;
 
         mutex_lock(&spu_mutex);
+        spin_lock_irqsave(&spu_list_lock, flags);
         list_add(&spu->list, &spu_list[spu->node]);
         list_add(&spu->full_list, &spu_full_list);
+        spin_unlock_irqrestore(&spu_list_lock, flags);
         mutex_unlock(&spu_mutex);
 
         goto out;
diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c
index b00653d69c01..505266a568d4 100644
--- a/arch/powerpc/platforms/cell/spufs/file.c
+++ b/arch/powerpc/platforms/cell/spufs/file.c
@@ -63,8 +63,8 @@ static ssize_t
 spufs_mem_read(struct file *file, char __user *buffer,
                                 size_t size, loff_t *pos)
 {
-        int ret;
         struct spu_context *ctx = file->private_data;
+        ssize_t ret;
 
         spu_acquire(ctx);
         ret = __spufs_mem_read(ctx, buffer, size, pos);
@@ -74,25 +74,29 @@ spufs_mem_read(struct file *file, char __user *buffer,
 
 static ssize_t
 spufs_mem_write(struct file *file, const char __user *buffer,
-                                        size_t size, loff_t *pos)
+                                        size_t size, loff_t *ppos)
 {
         struct spu_context *ctx = file->private_data;
         char *local_store;
+        loff_t pos = *ppos;
         int ret;
 
-        size = min_t(ssize_t, LS_SIZE - *pos, size);
-        if (size <= 0)
+        if (pos < 0)
+                return -EINVAL;
+        if (pos > LS_SIZE)
                 return -EFBIG;
-        *pos += size;
+        if (size > LS_SIZE - pos)
+                size = LS_SIZE - pos;
 
         spu_acquire(ctx);
-
         local_store = ctx->ops->get_ls(ctx);
-        ret = copy_from_user(local_store + *pos - size,
-                             buffer, size) ? -EFAULT : size;
-
+        ret = copy_from_user(local_store + pos, buffer, size);
         spu_release(ctx);
-        return ret;
+
+        if (ret)
+                return -EFAULT;
+        *ppos = pos + size;
+        return size;
 }
 
 static unsigned long spufs_mem_mmap_nopfn(struct vm_area_struct *vma,
diff --git a/arch/powerpc/platforms/cell/spufs/run.c b/arch/powerpc/platforms/cell/spufs/run.c
index 353a8fa07ab8..f95a611ca362 100644
--- a/arch/powerpc/platforms/cell/spufs/run.c
+++ b/arch/powerpc/platforms/cell/spufs/run.c
@@ -143,7 +143,7 @@ static inline int spu_run_init(struct spu_context *ctx, u32 * npc)
         int ret;
         unsigned long runcntl = SPU_RUNCNTL_RUNNABLE;
 
-        ret = spu_acquire_runnable(ctx, SPU_ACTIVATE_NOWAKE);
+        ret = spu_acquire_runnable(ctx, 0);
         if (ret)
                 return ret;
 
@@ -155,7 +155,7 @@ static inline int spu_run_init(struct spu_context *ctx, u32 * npc)
                         spu_release(ctx);
                         ret = spu_setup_isolated(ctx);
                         if (!ret)
-                                ret = spu_acquire_runnable(ctx, SPU_ACTIVATE_NOWAKE);
+                                ret = spu_acquire_runnable(ctx, 0);
                 }
 
                 /* if userspace has set the runcntrl register (eg, to issue an
diff --git a/arch/powerpc/platforms/cell/spufs/sched.c b/arch/powerpc/platforms/cell/spufs/sched.c
index 2f25e68b4bac..39823cec0844 100644
--- a/arch/powerpc/platforms/cell/spufs/sched.c
+++ b/arch/powerpc/platforms/cell/spufs/sched.c
@@ -127,14 +127,6 @@ static void spu_remove_from_active_list(struct spu *spu)
         mutex_unlock(&spu_prio->active_mutex[node]);
 }
 
-static inline void mm_needs_global_tlbie(struct mm_struct *mm)
-{
-        int nr = (NR_CPUS > 1) ? NR_CPUS : NR_CPUS + 1;
-
-        /* Global TLBIE broadcast required with SPEs. */
-        __cpus_setall(&mm->cpu_vm_mask, nr);
-}
-
 static BLOCKING_NOTIFIER_HEAD(spu_switch_notifier);
 
 static void spu_switch_notify(struct spu *spu, struct spu_context *ctx)
@@ -167,8 +159,7 @@ static void spu_bind_context(struct spu *spu, struct spu_context *ctx)
         ctx->spu = spu;
         ctx->ops = &spu_hw_ops;
         spu->pid = current->pid;
-        spu->mm = ctx->owner;
-        mm_needs_global_tlbie(spu->mm);
+        spu_associate_mm(spu, ctx->owner);
         spu->ibox_callback = spufs_ibox_callback;
         spu->wbox_callback = spufs_wbox_callback;
         spu->stop_callback = spufs_stop_callback;
@@ -205,7 +196,7 @@ static void spu_unbind_context(struct spu *spu, struct spu_context *ctx)
         spu->stop_callback = NULL;
         spu->mfc_callback = NULL;
         spu->dma_callback = NULL;
-        spu->mm = NULL;
+        spu_associate_mm(spu, NULL);
         spu->pid = 0;
         ctx->ops = &spu_backing_ops;
         ctx->spu = NULL;
@@ -263,7 +254,6 @@ static void spu_prio_wait(struct spu_context *ctx)
 {
         DEFINE_WAIT(wait);
 
-        set_bit(SPU_SCHED_WAKE, &ctx->sched_flags);
         prepare_to_wait_exclusive(&ctx->stop_wq, &wait, TASK_INTERRUPTIBLE);
         if (!signal_pending(current)) {
                 mutex_unlock(&ctx->state_mutex);
@@ -272,7 +262,6 @@ static void spu_prio_wait(struct spu_context *ctx)
         }
         __set_current_state(TASK_RUNNING);
         remove_wait_queue(&ctx->stop_wq, &wait);
-        clear_bit(SPU_SCHED_WAKE, &ctx->sched_flags);
 }
 
 /**
@@ -292,7 +281,7 @@ static void spu_reschedule(struct spu *spu)
         best = sched_find_first_bit(spu_prio->bitmap);
         if (best < MAX_PRIO) {
                 struct spu_context *ctx = spu_grab_context(best);
-                if (ctx && test_bit(SPU_SCHED_WAKE, &ctx->sched_flags))
+                if (ctx)
                         wake_up(&ctx->stop_wq);
         }
         spin_unlock(&spu_prio->runq_lock);
@@ -414,8 +403,7 @@ int spu_activate(struct spu_context *ctx, unsigned long flags)
                 }
 
                 spu_add_to_rq(ctx);
-                if (!(flags & SPU_ACTIVATE_NOWAKE))
-                        spu_prio_wait(ctx);
+                spu_prio_wait(ctx);
                 spu_del_from_rq(ctx);
         } while (!signal_pending(current));
 
diff --git a/arch/powerpc/platforms/cell/spufs/spufs.h b/arch/powerpc/platforms/cell/spufs/spufs.h
index 0c437891dfd5..5c4e47d69d79 100644
--- a/arch/powerpc/platforms/cell/spufs/spufs.h
+++ b/arch/powerpc/platforms/cell/spufs/spufs.h
@@ -41,7 +41,7 @@ struct spu_gang;
 
 /* ctx->sched_flags */
 enum {
-        SPU_SCHED_WAKE = 0,
+        SPU_SCHED_WAKE = 0, /* currently unused */
 };
 
 struct spu_context {
@@ -191,9 +191,7 @@ void spu_forget(struct spu_context *ctx);
 int spu_acquire_runnable(struct spu_context *ctx, unsigned long flags);
 void spu_acquire_saved(struct spu_context *ctx);
 int spu_acquire_exclusive(struct spu_context *ctx);
-enum {
-        SPU_ACTIVATE_NOWAKE = 1,
-};
+
 int spu_activate(struct spu_context *ctx, unsigned long flags);
 void spu_deactivate(struct spu_context *ctx);
 void spu_yield(struct spu_context *ctx);
diff --git a/arch/powerpc/platforms/cell/spufs/switch.c b/arch/powerpc/platforms/cell/spufs/switch.c
index c08981ff7fc6..fd91c73de34e 100644
--- a/arch/powerpc/platforms/cell/spufs/switch.c
+++ b/arch/powerpc/platforms/cell/spufs/switch.c
@@ -468,26 +468,6 @@ static inline void wait_purge_complete(struct spu_state *csa, struct spu *spu)
                          MFC_CNTL_PURGE_DMA_COMPLETE);
 }
 
-static inline void save_mfc_slbs(struct spu_state *csa, struct spu *spu)
-{
-        struct spu_priv2 __iomem *priv2 = spu->priv2;
-        int i;
-
-        /* Save, Step 29:
-         *     If MFC_SR1[R]='1', save SLBs in CSA.
-         */
-        if (spu_mfc_sr1_get(spu) & MFC_STATE1_RELOCATE_MASK) {
-                csa->priv2.slb_index_W = in_be64(&priv2->slb_index_W);
-                for (i = 0; i < 8; i++) {
-                        out_be64(&priv2->slb_index_W, i);
-                        eieio();
-                        csa->slb_esid_RW[i] = in_be64(&priv2->slb_esid_RW);
-                        csa->slb_vsid_RW[i] = in_be64(&priv2->slb_vsid_RW);
-                        eieio();
-                }
-        }
-}
-
 static inline void setup_mfc_sr1(struct spu_state *csa, struct spu *spu)
 {
         /* Save, Step 30:
@@ -708,20 +688,6 @@ static inline void resume_mfc_queue(struct spu_state *csa, struct spu *spu)
         out_be64(&priv2->mfc_control_RW, MFC_CNTL_RESUME_DMA_QUEUE);
 }
 
-static inline void invalidate_slbs(struct spu_state *csa, struct spu *spu)
-{
-        struct spu_priv2 __iomem *priv2 = spu->priv2;
-
-        /* Save, Step 45:
-         * Restore, Step 19:
-         *     If MFC_SR1[R]=1, write 0 to SLB_Invalidate_All.
-         */
-        if (spu_mfc_sr1_get(spu) & MFC_STATE1_RELOCATE_MASK) {
-                out_be64(&priv2->slb_invalidate_all_W, 0UL);
-                eieio();
-        }
-}
-
 static inline void get_kernel_slb(u64 ea, u64 slb[2])
 {
         u64 llp;
@@ -765,7 +731,7 @@ static inline void setup_mfc_slbs(struct spu_state *csa, struct spu *spu)
          *     MFC_SR1[R]=1 (in other words, assume that
          *     translation is desired by OS environment).
          */
-        invalidate_slbs(csa, spu);
+        spu_invalidate_slbs(spu);
         get_kernel_slb((unsigned long)&spu_save_code[0], code_slb);
         get_kernel_slb((unsigned long)csa->lscsa, lscsa_slb);
         load_mfc_slb(spu, code_slb, 0);
@@ -1718,27 +1684,6 @@ static inline void check_ppuint_mb_stat(struct spu_state *csa, struct spu *spu)
         }
 }
 
-static inline void restore_mfc_slbs(struct spu_state *csa, struct spu *spu)
-{
-        struct spu_priv2 __iomem *priv2 = spu->priv2;
-        int i;
-
-        /* Restore, Step 68:
-         *     If MFC_SR1[R]='1', restore SLBs from CSA.
-         */
-        if (csa->priv1.mfc_sr1_RW & MFC_STATE1_RELOCATE_MASK) {
-                for (i = 0; i < 8; i++) {
-                        out_be64(&priv2->slb_index_W, i);
-                        eieio();
-                        out_be64(&priv2->slb_esid_RW, csa->slb_esid_RW[i]);
-                        out_be64(&priv2->slb_vsid_RW, csa->slb_vsid_RW[i]);
-                        eieio();
-                }
-                out_be64(&priv2->slb_index_W, csa->priv2.slb_index_W);
-                eieio();
-        }
-}
-
 static inline void restore_mfc_sr1(struct spu_state *csa, struct spu *spu)
 {
         /* Restore, Step 69:
@@ -1875,7 +1820,6 @@ static void save_csa(struct spu_state *prev, struct spu *spu)
         set_mfc_tclass_id(prev, spu);   /* Step 26. */
         purge_mfc_queue(prev, spu);     /* Step 27. */
         wait_purge_complete(prev, spu); /* Step 28. */
-        save_mfc_slbs(prev, spu);       /* Step 29. */
         setup_mfc_sr1(prev, spu);       /* Step 30. */
         save_spu_npc(prev, spu);        /* Step 31. */
         save_spu_privcntl(prev, spu);   /* Step 32. */
@@ -1987,7 +1931,7 @@ static void harvest(struct spu_state *prev, struct spu *spu)
         reset_spu_privcntl(prev, spu);          /* Step 16. */
         reset_spu_lslr(prev, spu);              /* Step 17. */
         setup_mfc_sr1(prev, spu);               /* Step 18. */
-        invalidate_slbs(prev, spu);             /* Step 19. */
+        spu_invalidate_slbs(spu);               /* Step 19. */
         reset_ch_part1(prev, spu);              /* Step 20. */
         reset_ch_part2(prev, spu);              /* Step 21. */
         enable_interrupts(prev, spu);           /* Step 22. */
@@ -2055,7 +1999,7 @@ static void restore_csa(struct spu_state *next, struct spu *spu)
         restore_spu_mb(next, spu);              /* Step 65. */
         check_ppu_mb_stat(next, spu);           /* Step 66. */
         check_ppuint_mb_stat(next, spu);        /* Step 67. */
-        restore_mfc_slbs(next, spu);            /* Step 68. */
+        spu_invalidate_slbs(spu);               /* Modified Step 68. */
         restore_mfc_sr1(next, spu);             /* Step 69. */
         restore_other_spu_access(next, spu);    /* Step 70. */
         restore_spu_runcntl(next, spu);         /* Step 71. */
diff --git a/arch/powerpc/platforms/pasemi/iommu.c b/arch/powerpc/platforms/pasemi/iommu.c
index 4587c1dfeda5..07819cd3ac44 100644
--- a/arch/powerpc/platforms/pasemi/iommu.c
+++ b/arch/powerpc/platforms/pasemi/iommu.c
@@ -77,7 +77,7 @@
 #define IOBMAP_L2E_V            0x80000000
 #define IOBMAP_L2E_V_CACHED     0xc0000000
 
-static u32 *iob;
+static u32 __iomem *iob;
 static u32 iob_l1_emptyval;
 static u32 iob_l2_emptyval;
 static u32 *iob_l2_base;
diff --git a/arch/powerpc/platforms/pseries/hvCall.S b/arch/powerpc/platforms/pseries/hvCall.S
index 5c7e38789897..c1427b3634ec 100644
--- a/arch/powerpc/platforms/pseries/hvCall.S
+++ b/arch/powerpc/platforms/pseries/hvCall.S
@@ -30,9 +30,14 @@ END_FTR_SECTION_IFSET(CPU_FTR_PURR);
         
 /*
  * postcall is performed immediately before function return which
- * allows liberal use of volatile registers.
+ * allows liberal use of volatile registers.  We branch around this
+ * in early init (eg when populating the MMU hashtable) by using an
+ * unconditional cpu feature.
  */
 #define HCALL_INST_POSTCALL                                     \
+BEGIN_FTR_SECTION;                                              \
+        b       1f;                                             \
+END_FTR_SECTION(0, 1);                                          \
         ld      r4,STK_PARM(r3)(r1);    /* validate opcode */   \
         cmpldi  cr7,r4,MAX_HCALL_OPCODE;                        \
         bgt-    cr7,1f;                                         \
@@ -123,6 +128,40 @@ _GLOBAL(plpar_hcall)
 
         blr                             /* return r3 = status */
 
+/*
+ * plpar_hcall_raw can be called in real mode. kexec/kdump need some
+ * hypervisor calls to be executed in real mode. So plpar_hcall_raw
+ * does not access the per cpu hypervisor call statistics variables,
+ * since these variables may not be present in the RMO region.
+ */
+_GLOBAL(plpar_hcall_raw)
+        HMT_MEDIUM
+
+        mfcr    r0
+        stw     r0,8(r1)
+
+        std     r4,STK_PARM(r4)(r1)     /* Save ret buffer */
+
+        mr      r4,r5
+        mr      r5,r6
+        mr      r6,r7
+        mr      r7,r8
+        mr      r8,r9
+        mr      r9,r10
+
+        HVSC                            /* invoke the hypervisor */
+
+        ld      r12,STK_PARM(r4)(r1)
+        std     r4,  0(r12)
+        std     r5,  8(r12)
+        std     r6, 16(r12)
+        std     r7, 24(r12)
+
+        lwz     r0,8(r1)
+        mtcrf   0xff,r0
+
+        blr                             /* return r3 = status */
+
 _GLOBAL(plpar_hcall9)
         HMT_MEDIUM
 
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index 7496005566ef..843ee9643211 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -378,7 +378,7 @@ static void pSeries_lpar_hptab_clear(void)
 
         /* TODO: Use bulk call */
         for (i = 0; i < hpte_count; i++)
-                plpar_pte_remove(0, i, 0, &dummy1, &dummy2);
+                plpar_pte_remove_raw(0, i, 0, &dummy1, &dummy2);
 }
 
 /*
diff --git a/arch/powerpc/platforms/pseries/plpar_wrappers.h b/arch/powerpc/platforms/pseries/plpar_wrappers.h
index 3eb7b294d92f..2e4d10c9eea8 100644
--- a/arch/powerpc/platforms/pseries/plpar_wrappers.h
+++ b/arch/powerpc/platforms/pseries/plpar_wrappers.h
@@ -78,6 +78,22 @@ static inline long plpar_pte_remove(unsigned long flags, unsigned long ptex,
         return rc;
 }
 
+/* plpar_pte_remove_raw can be called in real mode. It calls plpar_hcall_raw */
+static inline long plpar_pte_remove_raw(unsigned long flags, unsigned long ptex,
+                unsigned long avpn, unsigned long *old_pteh_ret,
+                unsigned long *old_ptel_ret)
+{
+        long rc;
+        unsigned long retbuf[PLPAR_HCALL_BUFSIZE];
+
+        rc = plpar_hcall_raw(H_REMOVE, retbuf, flags, ptex, avpn);
+
+        *old_pteh_ret = retbuf[0];
+        *old_ptel_ret = retbuf[1];
+
+        return rc;
+}
+
 static inline long plpar_pte_read(unsigned long flags, unsigned long ptex,
                 unsigned long *old_pteh_ret, unsigned long *old_ptel_ret)
 {