]> git.kernelconcepts.de Git - karo-tx-linux.git/commitdiff
x86: UV BAU and nodes with no memory
authorCliff Wickman <cpw@sgi.com>
Fri, 3 Apr 2009 13:34:05 +0000 (08:34 -0500)
committerIngo Molnar <mingo@elte.hu>
Fri, 3 Apr 2009 16:25:26 +0000 (18:25 +0200)
This patch fixes BAU initialization for systems containing
nodes with no memory and for systems with non-consecutive
node numbers.

Fixes and clarifies situations where pnode should be used instead
of node id.

Tested on the UV hardware simulator.

Signed-off-by: Cliff Wickman <cpw@sgi.com>
LKML-Reference: <E1LpjX3-00007N-12@eag09.americas.sgi.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
arch/x86/kernel/tlb_uv.c

index 79c073247284cbdb334cf2a03b4e7ffaae8d0bcd..b833bc634d1729cab02da6916abb33734df638cd 100644 (file)
@@ -31,6 +31,34 @@ static unsigned long         uv_mmask __read_mostly;
 static DEFINE_PER_CPU(struct ptc_stats, ptcstats);
 static DEFINE_PER_CPU(struct bau_control, bau_control);
 
+/*
+ * Determine the first node on a blade.
+ */
+static int __init blade_to_first_node(int blade)
+{
+       int node, b;
+
+       for_each_online_node(node) {
+               b = uv_node_to_blade_id(node);
+               if (blade == b)
+                       return node;
+       }
+       BUG();
+}
+
+/*
+ * Determine the apicid of the first cpu on a blade.
+ */
+static int __init blade_to_first_apicid(int blade)
+{
+       int cpu;
+
+       for_each_present_cpu(cpu)
+               if (blade == uv_cpu_to_blade_id(cpu))
+                       return per_cpu(x86_cpu_to_apicid, cpu);
+       return -1;
+}
+
 /*
  * Free a software acknowledge hardware resource by clearing its Pending
  * bit. This will return a reply to the sender.
@@ -67,7 +95,7 @@ static void uv_bau_process_message(struct bau_payload_queue_entry *msg,
        msp = __get_cpu_var(bau_control).msg_statuses + msg_slot;
        cpu = uv_blade_processor_id();
        msg->number_of_cpus =
-           uv_blade_nr_online_cpus(uv_node_to_blade_id(numa_node_id()));
+               uv_blade_nr_online_cpus(uv_node_to_blade_id(numa_node_id()));
        this_cpu_mask = 1UL << cpu;
        if (msp->seen_by.bits & this_cpu_mask)
                return;
@@ -215,14 +243,14 @@ static int uv_wait_completion(struct bau_desc *bau_desc,
  * Returns @flush_mask if some remote flushing remains to be done. The
  * mask will have some bits still set.
  */
-const struct cpumask *uv_flush_send_and_wait(int cpu, int this_blade,
+const struct cpumask *uv_flush_send_and_wait(int cpu, int this_pnode,
                                             struct bau_desc *bau_desc,
                                             struct cpumask *flush_mask)
 {
        int completion_status = 0;
        int right_shift;
        int tries = 0;
-       int blade;
+       int pnode;
        int bit;
        unsigned long mmr_offset;
        unsigned long index;
@@ -265,8 +293,8 @@ const struct cpumask *uv_flush_send_and_wait(int cpu, int this_blade,
         * use the IPI method of shootdown on them.
         */
        for_each_cpu(bit, flush_mask) {
-               blade = uv_cpu_to_blade_id(bit);
-               if (blade == this_blade)
+               pnode = uv_cpu_to_pnode(bit);
+               if (pnode == this_pnode)
                        continue;
                cpumask_clear_cpu(bit, flush_mask);
        }
@@ -308,16 +336,16 @@ const struct cpumask *uv_flush_tlb_others(const struct cpumask *cpumask,
        struct cpumask *flush_mask = &__get_cpu_var(flush_tlb_mask);
        int i;
        int bit;
-       int blade;
+       int pnode;
        int uv_cpu;
-       int this_blade;
+       int this_pnode;
        int locals = 0;
        struct bau_desc *bau_desc;
 
        cpumask_andnot(flush_mask, cpumask, cpumask_of(cpu));
 
        uv_cpu = uv_blade_processor_id();
-       this_blade = uv_numa_blade_id();
+       this_pnode = uv_hub_info->pnode;
        bau_desc = __get_cpu_var(bau_control).descriptor_base;
        bau_desc += UV_ITEMS_PER_DESCRIPTOR * uv_cpu;
 
@@ -325,13 +353,13 @@ const struct cpumask *uv_flush_tlb_others(const struct cpumask *cpumask,
 
        i = 0;
        for_each_cpu(bit, flush_mask) {
-               blade = uv_cpu_to_blade_id(bit);
-               BUG_ON(blade > (UV_DISTRIBUTION_SIZE - 1));
-               if (blade == this_blade) {
+               pnode = uv_cpu_to_pnode(bit);
+               BUG_ON(pnode > (UV_DISTRIBUTION_SIZE - 1));
+               if (pnode == this_pnode) {
                        locals++;
                        continue;
                }
-               bau_node_set(blade, &bau_desc->distribution);
+               bau_node_set(pnode, &bau_desc->distribution);
                i++;
        }
        if (i == 0) {
@@ -349,7 +377,7 @@ const struct cpumask *uv_flush_tlb_others(const struct cpumask *cpumask,
        bau_desc->payload.address = va;
        bau_desc->payload.sending_cpu = cpu;
 
-       return uv_flush_send_and_wait(uv_cpu, this_blade, bau_desc, flush_mask);
+       return uv_flush_send_and_wait(uv_cpu, this_pnode, bau_desc, flush_mask);
 }
 
 /*
@@ -481,8 +509,7 @@ static int uv_ptc_seq_show(struct seq_file *file, void *data)
                           stat->requestee, stat->onetlb, stat->alltlb,
                           stat->s_retry, stat->d_retry, stat->ptc_i);
                seq_printf(file, "%lx %ld %ld %ld %ld %ld %ld\n",
-                          uv_read_global_mmr64(uv_blade_to_pnode
-                                       (uv_cpu_to_blade_id(cpu)),
+                          uv_read_global_mmr64(uv_cpu_to_pnode(cpu),
                                        UVH_LB_BAU_INTD_SOFTWARE_ACKNOWLEDGE),
                           stat->sflush, stat->dflush,
                           stat->retriesok, stat->nomsg,
@@ -616,16 +643,18 @@ static struct bau_control * __init uv_table_bases_init(int blade, int node)
  * finish the initialization of the per-blade control structures
  */
 static void __init
-uv_table_bases_finish(int blade, int node, int cur_cpu,
+uv_table_bases_finish(int blade,
                      struct bau_control *bau_tablesp,
                      struct bau_desc *adp)
 {
        struct bau_control *bcp;
-       int i;
+       int cpu;
 
-       for (i = cur_cpu; i < cur_cpu + uv_blade_nr_possible_cpus(blade); i++) {
-               bcp = (struct bau_control *)&per_cpu(bau_control, i);
+       for_each_present_cpu(cpu) {
+               if (blade != uv_cpu_to_blade_id(cpu))
+                       continue;
 
+               bcp = (struct bau_control *)&per_cpu(bau_control, cpu);
                bcp->bau_msg_head       = bau_tablesp->va_queue_first;
                bcp->va_queue_first     = bau_tablesp->va_queue_first;
                bcp->va_queue_last      = bau_tablesp->va_queue_last;
@@ -648,8 +677,7 @@ uv_activation_descriptor_init(int node, int pnode)
        struct bau_desc *adp;
        struct bau_desc *ad2;
 
-       adp = (struct bau_desc *)
-           kmalloc_node(16384, GFP_KERNEL, node);
+       adp = (struct bau_desc *)kmalloc_node(16384, GFP_KERNEL, node);
        BUG_ON(!adp);
 
        pa = __pa((unsigned long)adp);
@@ -666,8 +694,7 @@ uv_activation_descriptor_init(int node, int pnode)
        for (i = 0, ad2 = adp; i < UV_ACTIVATION_DESCRIPTOR_SIZE; i++, ad2++) {
                memset(ad2, 0, sizeof(struct bau_desc));
                ad2->header.sw_ack_flag = 1;
-               ad2->header.base_dest_nodeid =
-                   uv_blade_to_pnode(uv_cpu_to_blade_id(0));
+               ad2->header.base_dest_nodeid = uv_cpu_to_pnode(0);
                ad2->header.command = UV_NET_ENDPOINT_INTD;
                ad2->header.int_both = 1;
                /*
@@ -714,8 +741,9 @@ uv_payload_queue_init(int node, int pnode, struct bau_control *bau_tablesp)
 /*
  * Initialization of each UV blade's structures
  */
-static int __init uv_init_blade(int blade, int node, int cur_cpu)
+static int __init uv_init_blade(int blade)
 {
+       int node;
        int pnode;
        unsigned long pa;
        unsigned long apicid;
@@ -723,16 +751,17 @@ static int __init uv_init_blade(int blade, int node, int cur_cpu)
        struct bau_payload_queue_entry *pqp;
        struct bau_control *bau_tablesp;
 
+       node = blade_to_first_node(blade);
        bau_tablesp = uv_table_bases_init(blade, node);
        pnode = uv_blade_to_pnode(blade);
        adp = uv_activation_descriptor_init(node, pnode);
        pqp = uv_payload_queue_init(node, pnode, bau_tablesp);
-       uv_table_bases_finish(blade, node, cur_cpu, bau_tablesp, adp);
+       uv_table_bases_finish(blade, bau_tablesp, adp);
        /*
         * the below initialization can't be in firmware because the
         * messaging IRQ will be determined by the OS
         */
-       apicid = per_cpu(x86_cpu_to_apicid, cur_cpu);
+       apicid = blade_to_first_apicid(blade);
        pa = uv_read_global_mmr64(pnode, UVH_BAU_DATA_CONFIG);
        if ((pa & 0xff) != UV_BAU_MESSAGE) {
                uv_write_global_mmr64(pnode, UVH_BAU_DATA_CONFIG,
@@ -747,9 +776,7 @@ static int __init uv_init_blade(int blade, int node, int cur_cpu)
 static int __init uv_bau_init(void)
 {
        int blade;
-       int node;
        int nblades;
-       int last_blade;
        int cur_cpu;
 
        if (!is_uv_system())
@@ -758,29 +785,16 @@ static int __init uv_bau_init(void)
        uv_bau_retry_limit = 1;
        uv_nshift = uv_hub_info->n_val;
        uv_mmask = (1UL << uv_hub_info->n_val) - 1;
-       nblades = 0;
-       last_blade = -1;
-       cur_cpu = 0;
-       for_each_online_node(node) {
-               blade = uv_node_to_blade_id(node);
-               if (blade == last_blade)
-                       continue;
-               last_blade = blade;
-               nblades++;
-       }
+       nblades = uv_num_possible_blades();
+
        uv_bau_table_bases = (struct bau_control **)
            kmalloc(nblades * sizeof(struct bau_control *), GFP_KERNEL);
        BUG_ON(!uv_bau_table_bases);
 
-       last_blade = -1;
-       for_each_online_node(node) {
-               blade = uv_node_to_blade_id(node);
-               if (blade == last_blade)
-                       continue;
-               last_blade = blade;
-               uv_init_blade(blade, node, cur_cpu);
-               cur_cpu += uv_blade_nr_possible_cpus(blade);
-       }
+       for (blade = 0; blade < nblades; blade++)
+               if (uv_blade_nr_possible_cpus(blade))
+                       uv_init_blade(blade);
+
        alloc_intr_gate(UV_BAU_MESSAGE, uv_bau_message_intr1);
        uv_enable_timeouts();