powerpc/pseries: Add page coalescing support

Adds support for page coalescing, which is a feature on IBM Power servers
which allows for coalescing identical pages between logical partitions.
Hint text pages as coalesce candidates, since they are the most likely
pages to be able to be coalesced between partitions. This patch also
exports some page coalescing statistics available from firmware via
lparcfg.

[BenH: Moved a couple of things around to fix compile problems]

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

8 files changed, 106 insertions, 30 deletions

diff --git a/arch/powerpc/include/asm/firmware.h b/arch/powerpc/include/asm/firmware.h
index 4ef662e4a31d..3a6c586c4e40 100644
--- a/arch/powerpc/include/asm/firmware.h
+++ b/arch/powerpc/include/asm/firmware.h
@@ -47,6 +47,7 @@
 #define FW_FEATURE_BEAT         ASM_CONST(0x0000000001000000)
 #define FW_FEATURE_CMO          ASM_CONST(0x0000000002000000)
 #define FW_FEATURE_VPHN         ASM_CONST(0x0000000004000000)
+#define FW_FEATURE_XCMO         ASM_CONST(0x0000000008000000)
 
 #ifndef __ASSEMBLY__
 
@@ -60,7 +61,7 @@ enum {
                 FW_FEATURE_VIO | FW_FEATURE_RDMA | FW_FEATURE_LLAN |
                 FW_FEATURE_BULK_REMOVE | FW_FEATURE_XDABR |
                 FW_FEATURE_MULTITCE | FW_FEATURE_SPLPAR | FW_FEATURE_LPAR |
-                FW_FEATURE_CMO | FW_FEATURE_VPHN,
+                FW_FEATURE_CMO | FW_FEATURE_VPHN | FW_FEATURE_XCMO,
         FW_FEATURE_PSERIES_ALWAYS = 0,
         FW_FEATURE_ISERIES_POSSIBLE = FW_FEATURE_ISERIES | FW_FEATURE_LPAR,
         FW_FEATURE_ISERIES_ALWAYS = FW_FEATURE_ISERIES | FW_FEATURE_LPAR,
diff --git a/arch/powerpc/include/asm/hvcall.h b/arch/powerpc/include/asm/hvcall.h
index 8edec710cc6d..852b8c1c09db 100644
--- a/arch/powerpc/include/asm/hvcall.h
+++ b/arch/powerpc/include/asm/hvcall.h
@@ -102,6 +102,7 @@
 #define H_ANDCOND               (1UL<<(63-33))
 #define H_ICACHE_INVALIDATE     (1UL<<(63-40))  /* icbi, etc.  (ignored for IO pages) */
 #define H_ICACHE_SYNCHRONIZE    (1UL<<(63-41))  /* dcbst, icbi, etc (ignored for IO pages */
+#define H_COALESCE_CAND (1UL<<(63-42))  /* page is a good candidate for coalescing */
 #define H_ZERO_PAGE             (1UL<<(63-48))  /* zero the page before mapping (ignored for IO pages) */
 #define H_COPY_PAGE             (1UL<<(63-49))
 #define H_N                     (1UL<<(63-61))
@@ -234,6 +235,7 @@
 #define H_GET_MPP               0x2D4
 #define H_HOME_NODE_ASSOCIATIVITY 0x2EC
 #define H_BEST_ENERGY           0x2F4
+#define H_GET_MPP_X             0x314
 #define MAX_HCALL_OPCODE        H_BEST_ENERGY
 
 #ifndef __ASSEMBLY__
@@ -312,6 +314,16 @@ struct hvcall_mpp_data {
 
 int h_get_mpp(struct hvcall_mpp_data *);
 
+struct hvcall_mpp_x_data {
+        unsigned long coalesced_bytes;
+        unsigned long pool_coalesced_bytes;
+        unsigned long pool_purr_cycles;
+        unsigned long pool_spurr_cycles;
+        unsigned long reserved[3];
+};
+
+int h_get_mpp_x(struct hvcall_mpp_x_data *mpp_x_data);
+
 #ifdef CONFIG_PPC_PSERIES
 extern int CMO_PrPSP;
 extern int CMO_SecPSP;
diff --git a/arch/powerpc/include/asm/pSeries_reconfig.h b/arch/powerpc/include/asm/pSeries_reconfig.h
index d4b4bfa26fb3..89d2f99c1bf4 100644
--- a/arch/powerpc/include/asm/pSeries_reconfig.h
+++ b/arch/powerpc/include/asm/pSeries_reconfig.h
@@ -18,13 +18,18 @@
 extern int pSeries_reconfig_notifier_register(struct notifier_block *);
 extern void pSeries_reconfig_notifier_unregister(struct notifier_block *);
 extern struct blocking_notifier_head pSeries_reconfig_chain;
+/* Not the best place to put this, will be fixed when we move some
+ * of the rtas suspend-me stuff to pseries */
+extern void pSeries_coalesce_init(void);
 #else /* !CONFIG_PPC_PSERIES */
 static inline int pSeries_reconfig_notifier_register(struct notifier_block *nb)
 {
         return 0;
 }
 static inline void pSeries_reconfig_notifier_unregister(struct notifier_block *nb) { }
+static inline void pSeries_coalesce_init(void) { }
 #endif /* CONFIG_PPC_PSERIES */
 
+
 #endif /* __KERNEL__ */
 #endif /* _PPC64_PSERIES_RECONFIG_H */
diff --git a/arch/powerpc/kernel/lparcfg.c b/arch/powerpc/kernel/lparcfg.c
index 301db65f05a1..84daabe2fcba 100644
--- a/arch/powerpc/kernel/lparcfg.c
+++ b/arch/powerpc/kernel/lparcfg.c
@@ -132,34 +132,6 @@ static int iseries_lparcfg_data(struct seq_file *m, void *v)
 /*
  * Methods used to fetch LPAR data when running on a pSeries platform.
  */
-/**
- * h_get_mpp
- * H_GET_MPP hcall returns info in 7 parms
- */
-int h_get_mpp(struct hvcall_mpp_data *mpp_data)
-{
-        int rc;
-        unsigned long retbuf[PLPAR_HCALL9_BUFSIZE];
-
-        rc = plpar_hcall9(H_GET_MPP, retbuf);
-
-        mpp_data->entitled_mem = retbuf[0];
-        mpp_data->mapped_mem = retbuf[1];
-
-        mpp_data->group_num = (retbuf[2] >> 2 * 8) & 0xffff;
-        mpp_data->pool_num = retbuf[2] & 0xffff;
-
-        mpp_data->mem_weight = (retbuf[3] >> 7 * 8) & 0xff;
-        mpp_data->unallocated_mem_weight = (retbuf[3] >> 6 * 8) & 0xff;
-        mpp_data->unallocated_entitlement = retbuf[3] & 0xffffffffffff;
-
-        mpp_data->pool_size = retbuf[4];
-        mpp_data->loan_request = retbuf[5];
-        mpp_data->backing_mem = retbuf[6];
-
-        return rc;
-}
-EXPORT_SYMBOL(h_get_mpp);
 
 struct hvcall_ppp_data {
         u64     entitlement;
@@ -345,6 +317,30 @@ static void parse_mpp_data(struct seq_file *m)
         seq_printf(m, "backing_memory=%ld bytes\n", mpp_data.backing_mem);
 }
 
+/**
+ * parse_mpp_x_data
+ * Parse out data returned from h_get_mpp_x
+ */
+static void parse_mpp_x_data(struct seq_file *m)
+{
+        struct hvcall_mpp_x_data mpp_x_data;
+
+        if (!firmware_has_feature(FW_FEATURE_XCMO))
+                return;
+        if (h_get_mpp_x(&mpp_x_data))
+                return;
+
+        seq_printf(m, "coalesced_bytes=%ld\n", mpp_x_data.coalesced_bytes);
+
+        if (mpp_x_data.pool_coalesced_bytes)
+                seq_printf(m, "pool_coalesced_bytes=%ld\n",
+                           mpp_x_data.pool_coalesced_bytes);
+        if (mpp_x_data.pool_purr_cycles)
+                seq_printf(m, "coalesce_pool_purr=%ld\n", mpp_x_data.pool_purr_cycles);
+        if (mpp_x_data.pool_spurr_cycles)
+                seq_printf(m, "coalesce_pool_spurr=%ld\n", mpp_x_data.pool_spurr_cycles);
+}
+
 #define SPLPAR_CHARACTERISTICS_TOKEN 20
 #define SPLPAR_MAXLENGTH 1026*(sizeof(char))
 
@@ -520,6 +516,7 @@ static int pseries_lparcfg_data(struct seq_file *m, void *v)
                 parse_system_parameter_string(m);
                 parse_ppp_data(m);
                 parse_mpp_data(m);
+                parse_mpp_x_data(m);
                 pseries_cmo_data(m);
                 splpar_dispatch_data(m);
 
diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c
index 7839bd7bfd15..c016033ba78d 100644
--- a/arch/powerpc/kernel/prom_init.c
+++ b/arch/powerpc/kernel/prom_init.c
@@ -700,8 +700,10 @@ static void __init early_cmdline_parse(void)
 #endif /* CONFIG_PCI_MSI */
 #ifdef CONFIG_PPC_SMLPAR
 #define OV5_CMO                 0x80    /* Cooperative Memory Overcommitment */
+#define OV5_XCMO                        0x40    /* Page Coalescing */
 #else
 #define OV5_CMO                 0x00
+#define OV5_XCMO                        0x00
 #endif
 #define OV5_TYPE1_AFFINITY      0x80    /* Type 1 NUMA affinity */
 
@@ -756,7 +758,7 @@ static unsigned char ibm_architecture_vec[] = {
         OV5_LPAR | OV5_SPLPAR | OV5_LARGE_PAGES | OV5_DRCONF_MEMORY |
         OV5_DONATE_DEDICATE_CPU | OV5_MSI,
         0,
-        OV5_CMO,
+        OV5_CMO | OV5_XCMO,
         OV5_TYPE1_AFFINITY,
         0,
         0,
diff --git a/arch/powerpc/kernel/rtas.c b/arch/powerpc/kernel/rtas.c
index f48446635c89..271ff6318eda 100644
--- a/arch/powerpc/kernel/rtas.c
+++ b/arch/powerpc/kernel/rtas.c
@@ -42,6 +42,7 @@
 #include <asm/time.h>
 #include <asm/mmu.h>
 #include <asm/topology.h>
+#include <asm/pSeries_reconfig.h>
 
 struct rtas_t rtas = {
         .lock = __ARCH_SPIN_LOCK_UNLOCKED
@@ -731,6 +732,7 @@ static int __rtas_suspend_last_cpu(struct rtas_suspend_me_data *data, int wake_w
 
         atomic_set(&data->error, rc);
         start_topology_update();
+        pSeries_coalesce_init();
 
         if (wake_when_done) {
                 atomic_set(&data->done, 1);
diff --git a/arch/powerpc/platforms/pseries/lpar.c b/arch/powerpc/platforms/pseries/lpar.c
index 6f0ed3aac77f..39e6e0a7b2fa 100644
--- a/arch/powerpc/platforms/pseries/lpar.c
+++ b/arch/powerpc/platforms/pseries/lpar.c
@@ -329,6 +329,8 @@ static long pSeries_lpar_hpte_insert(unsigned long hpte_group,
         /* Make pHyp happy */
         if ((rflags & _PAGE_NO_CACHE) & !(rflags & _PAGE_WRITETHRU))
                 hpte_r &= ~_PAGE_COHERENT;
+        if (firmware_has_feature(FW_FEATURE_XCMO) && !(hpte_r & HPTE_R_N))
+                flags |= H_COALESCE_CAND;
 
         lpar_rc = plpar_pte_enter(flags, hpte_group, hpte_v, hpte_r, &slot);
         if (unlikely(lpar_rc == H_PTEG_FULL)) {
@@ -771,3 +773,47 @@ out:
         local_irq_restore(flags);
 }
 #endif
+
+/**
+ * h_get_mpp
+ * H_GET_MPP hcall returns info in 7 parms
+ */
+int h_get_mpp(struct hvcall_mpp_data *mpp_data)
+{
+        int rc;
+        unsigned long retbuf[PLPAR_HCALL9_BUFSIZE];
+
+        rc = plpar_hcall9(H_GET_MPP, retbuf);
+
+        mpp_data->entitled_mem = retbuf[0];
+        mpp_data->mapped_mem = retbuf[1];
+
+        mpp_data->group_num = (retbuf[2] >> 2 * 8) & 0xffff;
+        mpp_data->pool_num = retbuf[2] & 0xffff;
+
+        mpp_data->mem_weight = (retbuf[3] >> 7 * 8) & 0xff;
+        mpp_data->unallocated_mem_weight = (retbuf[3] >> 6 * 8) & 0xff;
+        mpp_data->unallocated_entitlement = retbuf[3] & 0xffffffffffff;
+
+        mpp_data->pool_size = retbuf[4];
+        mpp_data->loan_request = retbuf[5];
+        mpp_data->backing_mem = retbuf[6];
+
+        return rc;
+}
+EXPORT_SYMBOL(h_get_mpp);
+
+int h_get_mpp_x(struct hvcall_mpp_x_data *mpp_x_data)
+{
+        int rc;
+        unsigned long retbuf[PLPAR_HCALL9_BUFSIZE] = { 0 };
+
+        rc = plpar_hcall9(H_GET_MPP_X, retbuf);
+
+        mpp_x_data->coalesced_bytes = retbuf[0];
+        mpp_x_data->pool_coalesced_bytes = retbuf[1];
+        mpp_x_data->pool_purr_cycles = retbuf[2];
+        mpp_x_data->pool_spurr_cycles = retbuf[3];
+
+        return rc;
+}
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index ab73ad2ff59d..1689adccc6d7 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -405,6 +405,16 @@ static int pseries_set_xdabr(unsigned long dabr)
 #define CMO_CHARACTERISTICS_TOKEN 44
 #define CMO_MAXLENGTH 1026
 
+void pSeries_coalesce_init(void)
+{
+        struct hvcall_mpp_x_data mpp_x_data;
+
+        if (firmware_has_feature(FW_FEATURE_CMO) && !h_get_mpp_x(&mpp_x_data))
+                powerpc_firmware_features |= FW_FEATURE_XCMO;
+        else
+                powerpc_firmware_features &= ~FW_FEATURE_XCMO;
+}
+
 /**
  * fw_cmo_feature_init - FW_FEATURE_CMO is not stored in ibm,hypertas-functions,
  * handle that here. (Stolen from parse_system_parameter_string)
@@ -474,6 +484,7 @@ void pSeries_cmo_feature_init(void)
                 pr_debug("CMO enabled, PrPSP=%d, SecPSP=%d\n", CMO_PrPSP,
                          CMO_SecPSP);
                 powerpc_firmware_features |= FW_FEATURE_CMO;
+                pSeries_coalesce_init();
         } else
                 pr_debug("CMO not enabled, PrPSP=%d, SecPSP=%d\n", CMO_PrPSP,
                          CMO_SecPSP);