x86, UV: Fix macros for multiple coherency domains

Fix bug in the SGI UV macros that support systems with multiple
coherency domains.  The macros used for referencing global MMR
(chipset registers) are failing to correctly "or" the NASID
(node identifier) bits that reside above M+N. These high bits
are supplied automatically by the chipset for memory accesses
coming from the processor socket.

However, the bits must be present for references to the special
global MMR space used to map chipset registers. (See uv_hub.h
for more details ...)

The bug results in references to invalid/incorrect nodes.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Cc: <stable@kernel.org>
LKML-Reference: <20090608154405.GA16395@sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h
index d3a98ea1062ef936c7825f1387a4d2f4b83c65b9..341070f7ad5cb62679f22f7b3fff7316c549b5d9 100644 (file)
--- a/arch/x86/include/asm/uv/uv_hub.h
+++ b/arch/x86/include/asm/uv/uv_hub.h
@@ -133,6 +133,7 @@ struct uv_scir_s {
 struct uv_hub_info_s {
        unsigned long           global_mmr_base;
        unsigned long           gpa_mask;
 struct uv_hub_info_s {
        unsigned long           global_mmr_base;
        unsigned long           gpa_mask;

+       unsigned int            gnode_extra;

        unsigned long           gnode_upper;
        unsigned long           lowmem_remap_top;
        unsigned long           lowmem_remap_base;
        unsigned long           gnode_upper;
        unsigned long           lowmem_remap_top;
        unsigned long           lowmem_remap_base;


@@ -159,7 +160,8 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info);
  *             p -  PNODE (local part of nsids, right shifted 1)
  */
 #define UV_NASID_TO_PNODE(n)           (((n) >> 1) & uv_hub_info->pnode_mask)
  *             p -  PNODE (local part of nsids, right shifted 1)
  */
 #define UV_NASID_TO_PNODE(n)           (((n) >> 1) & uv_hub_info->pnode_mask)

-#define UV_PNODE_TO_NASID(p)           (((p) << 1) | uv_hub_info->gnode_upper)
+#define UV_PNODE_TO_GNODE(p)           ((p) |uv_hub_info->gnode_extra)
+#define UV_PNODE_TO_NASID(p)           (UV_PNODE_TO_GNODE(p) << 1)

 
 #define UV_LOCAL_MMR_BASE              0xf4000000UL
 #define UV_GLOBAL_MMR32_BASE           0xf8000000UL
 
 #define UV_LOCAL_MMR_BASE              0xf4000000UL
 #define UV_GLOBAL_MMR32_BASE           0xf8000000UL


@@ -173,7 +175,7 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info);
 #define UV_GLOBAL_MMR32_PNODE_BITS(p)  ((p) << (UV_GLOBAL_MMR32_PNODE_SHIFT))
 
 #define UV_GLOBAL_MMR64_PNODE_BITS(p)                                  \
 #define UV_GLOBAL_MMR32_PNODE_BITS(p)  ((p) << (UV_GLOBAL_MMR32_PNODE_SHIFT))
 
 #define UV_GLOBAL_MMR64_PNODE_BITS(p)                                  \

-       ((unsigned long)(p) << UV_GLOBAL_MMR64_PNODE_SHIFT)
+       ((unsigned long)(UV_PNODE_TO_GNODE(p)) << UV_GLOBAL_MMR64_PNODE_SHIFT)

 
 #define UV_APIC_PNODE_SHIFT    6
 
 
 #define UV_APIC_PNODE_SHIFT    6
 

diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c
index 2bda693529762b71f30ae2c26fa08a7692ca5cb5..39f2af4b546750702aa78253f3291cc3360d1f05 100644 (file)
--- a/arch/x86/kernel/apic/x2apic_uv_x.c
+++ b/arch/x86/kernel/apic/x2apic_uv_x.c
@@ -562,7 +562,7 @@ void __init uv_system_init(void)
        union uvh_node_id_u node_id;
        unsigned long gnode_upper, lowmem_redir_base, lowmem_redir_size;
        int bytes, nid, cpu, lcpu, pnode, blade, i, j, m_val, n_val;
        union uvh_node_id_u node_id;
        unsigned long gnode_upper, lowmem_redir_base, lowmem_redir_size;
        int bytes, nid, cpu, lcpu, pnode, blade, i, j, m_val, n_val;

-       int max_pnode = 0;
+       int gnode_extra, max_pnode = 0;

        unsigned long mmr_base, present, paddr;
        unsigned short pnode_mask;
 
        unsigned long mmr_base, present, paddr;
        unsigned short pnode_mask;
 


@@ -574,6 +574,13 @@ void __init uv_system_init(void)
        mmr_base =
            uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) &
            ~UV_MMR_ENABLE;
        mmr_base =
            uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) &
            ~UV_MMR_ENABLE;

+       pnode_mask = (1 << n_val) - 1;
+       node_id.v = uv_read_local_mmr(UVH_NODE_ID);
+       gnode_extra = (node_id.s.node_id & ~((1 << n_val) - 1)) >> 1;
+       gnode_upper = ((unsigned long)gnode_extra  << m_val);
+       printk(KERN_DEBUG "UV: N %d, M %d, gnode_upper 0x%lx, gnode_extra 0x%x\n",
+                       n_val, m_val, gnode_upper, gnode_extra);
+

        printk(KERN_DEBUG "UV: global MMR base 0x%lx\n", mmr_base);
 
        for(i = 0; i < UVH_NODE_PRESENT_TABLE_DEPTH; i++)
        printk(KERN_DEBUG "UV: global MMR base 0x%lx\n", mmr_base);
 
        for(i = 0; i < UVH_NODE_PRESENT_TABLE_DEPTH; i++)


@@ -607,11 +614,6 @@ void __init uv_system_init(void)
                }
        }
 
                }
        }
 

-       pnode_mask = (1 << n_val) - 1;
-       node_id.v = uv_read_local_mmr(UVH_NODE_ID);
-       gnode_upper = (((unsigned long)node_id.s.node_id) &
-                      ~((1 << n_val) - 1)) << m_val;
-

        uv_bios_init();
        uv_bios_get_sn_info(0, &uv_type, &sn_partition_id,
                            &sn_coherency_id, &sn_region_size);
        uv_bios_init();
        uv_bios_get_sn_info(0, &uv_type, &sn_partition_id,
                            &sn_coherency_id, &sn_region_size);


@@ -634,6 +636,7 @@ void __init uv_system_init(void)
                uv_cpu_hub_info(cpu)->pnode_mask = pnode_mask;
                uv_cpu_hub_info(cpu)->gpa_mask = (1 << (m_val + n_val)) - 1;
                uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper;
                uv_cpu_hub_info(cpu)->pnode_mask = pnode_mask;
                uv_cpu_hub_info(cpu)->gpa_mask = (1 << (m_val + n_val)) - 1;
                uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper;

+               uv_cpu_hub_info(cpu)->gnode_extra = gnode_extra;

                uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base;
                uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id;
                uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu;
                uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base;
                uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id;
                uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu;