patch-2.3.43 linux/arch/ia64/kernel/perfmon.c

Next file: linux/arch/ia64/kernel/process.c
Previous file: linux/arch/ia64/kernel/pci.c
Back to the patch index
Back to the overall index

diff -u --recursive --new-file v2.3.42/linux/arch/ia64/kernel/perfmon.c linux/arch/ia64/kernel/perfmon.c
@@ -0,0 +1,227 @@
+#include <linux/config.h>
+
+#include <linux/kernel.h>
+#include <linux/sched.h>
+#include <linux/smp_lock.h>
+
+#include <asm/errno.h>
+#include <asm/irq.h>
+#include <asm/processor.h>
+#include <asm/system.h>
+#include <asm/uaccess.h>
+
+#ifdef CONFIG_PERFMON
+
+#define MAX_PERF_COUNTER	4	/* true for Itanium, at least */
+#define WRITE_PMCS_AND_START	0xa0
+#define WRITE_PMCS		0xa1
+#define READ_PMDS		0xa2
+#define STOP_PMCS		0xa3
+#define IA64_COUNTER_MASK	0xffffffffffffff6f
+#define PERF_OVFL_VAL		0xffffffff
+
+struct perfmon_counter {
+        unsigned long data;
+        int counter_num;
+};
+
+unsigned long pmds[MAX_PERF_COUNTER];
+struct task_struct *perf_owner;
+
+/*
+ * We set dcr.pp, psr.pp, and the appropriate pmc control values with
+ * this.  Notice that we go about modifying _each_ task's pt_regs to
+ * set cr_ipsr.pp.  This will start counting when "current" does an
+ * _rfi_. Also, since each task's cr_ipsr.pp, and cr_ipsr is inherited
+ * across forks, we do _not_ need additional code on context
+ * switches. On stopping of the counters we dont _need_ to go about
+ * changing every task's cr_ipsr back to where it wuz, because we can
+ * just set pmc[0]=1. But we do it anyways becuase we will probably
+ * add thread specific accounting later.
+ *
+ * The obvious problem with this is that on SMP systems, it is a bit
+ * of work (when someone wants to do it) - it would be easier if we
+ * just added code to the context-switch path.  I think we would need
+ * to lock the run queue to ensure no context switches, send an IPI to
+ * each processor, and in that IPI handler, just modify the psr bit of
+ * only the _current_ thread, since we have modified the psr bit
+ * correctly in the kernel stack for every process which is not
+ * running.  Might crash on SMP systems without the
+ * lock_kernel(). Hence the lock..
+ */
+asmlinkage unsigned long
+sys_perfmonctl (int cmd1, int cmd2, void *ptr)
+{
+        struct perfmon_counter tmp, *cptr = ptr;
+        unsigned long pmd, cnum, dcr, flags;
+        struct task_struct *p;
+        struct pt_regs *regs;
+        struct perf_counter;
+        int i;
+
+        switch (cmd1) {
+	      case WRITE_PMCS:           /* Writes to PMC's and clears PMDs */
+	      case WRITE_PMCS_AND_START: /* Also starts counting */
+
+		if (!access_ok(VERIFY_READ, cptr, sizeof(struct perf_counter)*cmd2))
+			return -EFAULT;
+
+		if (cmd2 >= MAX_PERF_COUNTER)
+			return -EFAULT;
+
+		if (perf_owner && perf_owner != current)
+			return -EBUSY;
+		perf_owner = current;
+
+		for (i = 0; i < cmd2; i++, cptr++) {
+			copy_from_user(&tmp, cptr, sizeof(tmp));
+			/* XXX need to check validity of counter_num and perhaps data!! */
+			ia64_set_pmc(tmp.counter_num, tmp.data);
+			ia64_set_pmd(tmp.counter_num, 0);
+			pmds[tmp.counter_num - 4] = 0;
+		}
+
+		if (cmd1 == WRITE_PMCS_AND_START) {
+			local_irq_save(flags);
+			dcr = ia64_get_dcr();
+			dcr |= IA64_DCR_PP;
+			ia64_set_dcr(dcr);
+			local_irq_restore(flags);
+
+			/*
+			 * This is a no can do.  It obviously wouldn't
+			 * work on SMP where another process may not
+			 * be blocked at all.
+			 *
+			 * Perhaps we need a global predicate in the
+			 * leave_kernel path to control if pp should
+			 * be on or off?
+			 */
+			lock_kernel();
+			for_each_task(p) {
+				regs = (struct pt_regs *) (((char *)p) + IA64_STK_OFFSET) - 1;
+				ia64_psr(regs)->pp = 1;
+			}
+			unlock_kernel();
+			ia64_set_pmc(0, 0);
+		}
+                break;
+
+	      case READ_PMDS:
+		if (cmd2 >= MAX_PERF_COUNTER)
+			return -EFAULT;
+		if (!access_ok(VERIFY_WRITE, cptr, sizeof(struct perf_counter)*cmd2))
+			return -EFAULT;
+		local_irq_save(flags);
+		/* XXX this looks wrong */
+		__asm__ __volatile__("rsm psr.pp\n");
+		dcr = ia64_get_dcr();
+		dcr &= ~IA64_DCR_PP;
+		ia64_set_dcr(dcr);
+		local_irq_restore(flags);
+
+		/*
+		 * We cannot touch pmc[0] to stop counting here, as
+		 * that particular instruction might cause an overflow
+		 * and the mask in pmc[0] might get lost. I'm not very
+		 * sure of the hardware behavior here. So we stop
+		 * counting by psr.pp = 0. And we reset dcr.pp to
+		 * prevent an interrupt from mucking up psr.pp in the
+		 * meanwhile. Perfmon interrupts are pended, hence the
+		 * above code should be ok if one of the above
+		 * instructions cause overflows. Is this ok?  When I
+		 * muck with dcr, is the cli/sti needed??
+		 */
+		for (i = 0, cnum = 4; i < MAX_PERF_COUNTER; i++, cnum++, cptr++) {
+			pmd = pmds[i] + (ia64_get_pmd(cnum) & PERF_OVFL_VAL);
+			put_user(pmd, &cptr->data);
+		}
+		local_irq_save(flags);
+		/* XXX this looks wrong */
+		__asm__ __volatile__("ssm psr.pp");
+		dcr = ia64_get_dcr();
+		dcr |= IA64_DCR_PP;
+		ia64_set_dcr(dcr);
+		local_irq_restore(flags);
+                break;
+
+	      case STOP_PMCS:
+		ia64_set_pmc(0, 1);
+		for (i = 0; i < MAX_PERF_COUNTER; ++i)
+			ia64_set_pmc(i, 0);
+
+		local_irq_save(flags);
+		dcr = ia64_get_dcr();
+		dcr &= ~IA64_DCR_PP;
+		ia64_set_dcr(dcr);
+		local_irq_restore(flags);
+		/*
+		 * This is a no can do.  It obviously wouldn't
+		 * work on SMP where another process may not
+		 * be blocked at all.
+		 *
+		 * Perhaps we need a global predicate in the
+		 * leave_kernel path to control if pp should
+		 * be on or off?
+		 */
+		lock_kernel();
+		for_each_task(p) {
+			regs = (struct pt_regs *) (((char *)p) + IA64_STK_OFFSET) - 1;
+			ia64_psr(regs)->pp = 0;
+		}
+		unlock_kernel();
+		perf_owner = 0;
+		break;
+
+	      default:
+		break;
+        }
+        return 0;
+}
+
+static inline void
+update_counters (void)
+{
+	unsigned long mask, i, cnum, val;
+
+	mask = ia64_get_pmd(0) >> 4;
+	for (i = 0, cnum = 4; i < MAX_PERF_COUNTER; cnum++, i++, mask >>= 1) {
+		if (mask & 0x1) 
+			val = PERF_OVFL_VAL;
+		else
+			/* since we got an interrupt, might as well clear every pmd. */
+			val = ia64_get_pmd(cnum) & PERF_OVFL_VAL;
+		pmds[i] += val;
+		ia64_set_pmd(cnum, 0);
+	}
+}
+
+static void
+perfmon_interrupt (int irq, void *arg, struct pt_regs *regs)
+{
+	update_counters();
+	ia64_set_pmc(0, 0);
+	ia64_srlz_d();
+}
+
+void
+perfmon_init (void)
+{
+        if (request_irq(PERFMON_IRQ, perfmon_interrupt, 0, "perfmon", NULL)) {
+		printk("perfmon_init: could not allocate performance monitor vector %u\n",
+		       PERFMON_IRQ);
+		return;
+	}
+	ia64_set_pmv(PERFMON_IRQ);
+	ia64_srlz_d();
+}
+
+#else /* !CONFIG_PERFMON */
+
+asmlinkage unsigned long
+sys_perfmonctl (int cmd1, int cmd2, void *ptr)
+{
+	return -ENOSYS;
+}
+
+#endif /* !CONFIG_PERFMON */

FUNET's LINUX-ADM group, linux-adm@nic.funet.fi
TCL-scripts by Sam Shen (who was at: slshen@lbl.gov)