Add a simple KVM tool to dump the kernel state of UMA, including walking

the keg/zone lists, summarizing cache state, and walking bucket lists in
each zone.  I seem to get inconsistent results on SMP, possibly due to
local header problems, but it seems to work quite well on UP.  This tool
requires sufficient privilege to read /dev/mem (or a core dump), and is
for debugging purposes rather than administrative monitoring purposes
(use vmstat instead).

2 files changed, 381 insertions, 0 deletions

diff --git a/tools/tools/umastat/Makefile b/tools/tools/umastat/Makefile
new file mode 100644
index 0000000..98f8adc
--- /dev/null
+++ b/tools/tools/umastat/Makefile
@@ -0,0 +1,11 @@
+# $FreeBSD$
+
+PROG=	umastat
+NO_MAN=	
+
+DPADD=	${LIBKVM}
+LDADD=	-lkvm
+
+WARNS=	3
+
+.include <bsd.prog.mk>
diff --git a/tools/tools/umastat/umastat.c b/tools/tools/umastat/umastat.c
new file mode 100644
index 0000000..b1b685d
--- /dev/null
+++ b/tools/tools/umastat/umastat.c
@@ -0,0 +1,370 @@
+/*-
+ * Copyright (c) 2005 Robert N. M. Watson
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#include <sys/param.h>
+
+#define LIBMEMSTAT	/* Cause vm_page.h not to include opt_vmpage.h */
+#include <vm/vm.h>
+#include <vm/vm_page.h>
+
+#include <vm/uma.h>
+#include <vm/uma_int.h>
+
+#include <err.h>
+#include <kvm.h>
+#include <memstat.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+static struct nlist namelist[] = {
+#define X_UMA_KEGS      0
+	{ .n_name = "_uma_kegs" },
+#define X_MP_MAXID      1
+	{ .n_name = "_mp_maxid" },
+#define	X_ALLCPU	2
+	{ .n_name = "_all_cpus" },
+	{ .n_name = "" },
+};
+
+static void
+usage(void)
+{
+
+	fprintf(stderr, "umastat\n");
+	exit(-1);
+}
+
+static int
+kread(kvm_t *kvm, void *kvm_pointer, void *address, size_t size,
+    size_t offset)
+{
+	ssize_t ret;
+
+	ret = kvm_read(kvm, (unsigned long)kvm_pointer + offset, address,
+	    size);
+	if (ret < 0)
+		return (MEMSTAT_ERROR_KVM);
+	if ((size_t)ret != size)
+		return (MEMSTAT_ERROR_KVM_SHORTREAD);
+	return (0);
+}
+
+static int
+kread_string(kvm_t *kvm, void *kvm_pointer, char *buffer, int buflen)
+{
+	ssize_t ret;
+	int i;
+
+	for (i = 0; i < buflen; i++) {
+		ret = kvm_read(kvm, (unsigned long)kvm_pointer + i,
+		    &(buffer[i]), sizeof(char));
+		if (ret < 0)
+			return (MEMSTAT_ERROR_KVM);
+		if ((size_t)ret != sizeof(char))
+			return (MEMSTAT_ERROR_KVM_SHORTREAD);
+		if (buffer[i] == '\0')
+			return (0);
+	}
+	/* Truncate. */
+	buffer[i-1] = '\0';
+	return (0);
+}
+
+static int
+kread_symbol(kvm_t *kvm, int index, void *address, size_t size,
+    size_t offset)
+{
+	ssize_t ret;
+
+	ret = kvm_read(kvm, namelist[index].n_value + offset, address, size);
+	if (ret < 0)
+		return (MEMSTAT_ERROR_KVM);
+	if ((size_t)ret != size)
+		return (MEMSTAT_ERROR_KVM_SHORTREAD);
+	return (0);
+}
+
+static const struct flaginfo {
+	u_int32_t	 fi_flag;
+	const char	*fi_name;
+} flaginfo[] = {
+	{ UMA_ZFLAG_PRIVALLOC, "privalloc" },
+	{ UMA_ZFLAG_INTERNAL, "internal" },
+	{ UMA_ZFLAG_FULL, "full" },
+	{ UMA_ZFLAG_CACHEONLY, "cacheonly" },
+	{ UMA_ZONE_PAGEABLE, "pageable" },
+	{ UMA_ZONE_ZINIT, "zinit" },
+	{ UMA_ZONE_STATIC, "static" },
+	{ UMA_ZONE_OFFPAGE, "offpage" },
+	{ UMA_ZONE_MALLOC, "malloc" },
+	{ UMA_ZONE_NOFREE, "nofree" },
+	{ UMA_ZONE_MTXCLASS, "mtxclass" },
+	{ UMA_ZONE_VM, "vm" },
+	{ UMA_ZONE_HASH, "hash" },
+	{ UMA_ZONE_SECONDARY, "secondary" },
+	{ UMA_ZONE_REFCNT, "refcnt" },
+	{ UMA_ZONE_MAXBUCKET, "maxbucket" },
+};
+static const int flaginfo_count = sizeof(flaginfo) / sizeof(struct flaginfo);
+
+static void
+uma_print_keg_flags(struct uma_keg *ukp, const char *spaces)
+{
+	int count, i;
+
+	if (!ukp->uk_flags) {
+		printf("%suk_flags = 0;\n", spaces);
+		return;
+	}
+
+	printf("%suk_flags = ", spaces);
+	for (i = 0, count = 0; i < flaginfo_count; i++) {
+		if (ukp->uk_flags & flaginfo[i].fi_flag) {
+			if (count++ > 0)
+				printf(" | ");
+			printf(flaginfo[i].fi_name);
+		}
+
+	}
+	printf(";\n");
+}
+
+static void
+uma_print_keg_align(struct uma_keg *ukp, const char *spaces)
+{
+
+	switch(ukp->uk_align) {
+	case UMA_ALIGN_PTR:
+		printf("%suk_align = UMA_ALIGN_PTR;\n", spaces);
+		break;
+
+#if 0
+	case UMA_ALIGN_LONG:
+		printf("%suk_align = UMA_ALIGN_LONG;\n", spaces);
+		break;
+
+	case UMA_ALIGN_INT:
+		printf("%suk_align = UMA_ALIGN_INT;\n", spaces);
+		break;
+#endif
+
+	case UMA_ALIGN_SHORT:
+		printf("%suk_align = UMA_ALIGN_SHORT;\n", spaces);
+		break;
+
+	case UMA_ALIGN_CHAR:
+		printf("%suk_align = UMA_ALIGN_CHAR;\n", spaces);
+		break;
+
+	case UMA_ALIGN_CACHE:
+		printf("%suk_align = UMA_ALIGN_CACHE;\n", spaces);
+		break;
+
+	default:
+		printf("%suk_align = %d\n", spaces, ukp->uk_align);
+	}
+}
+
+LIST_HEAD(bucketlist, uma_bucket);
+
+static void
+uma_print_bucket(struct uma_bucket *ubp, const char *spaces)
+{
+
+	printf("{ ub_cnt = %d, ub_entries = %d }", ubp->ub_cnt,
+	    ubp->ub_entries);
+}
+
+static void
+uma_print_bucketlist(kvm_t *kvm, struct bucketlist *bucketlist,
+    const char *name, const char *spaces)
+{
+	struct uma_bucket *ubp, ub;
+	uint64_t total_entries, total_cnt;
+	int count, ret;
+
+	printf("%s%s {", spaces, name);
+
+	total_entries = total_cnt = 0;
+	count = 0;
+	for (ubp = LIST_FIRST(bucketlist); ubp != NULL; ubp =
+	    LIST_NEXT(&ub, ub_link)) {
+		ret = kread(kvm, ubp, &ub, sizeof(ub), 0);
+		if (ret != 0)
+			errx(-1, "uma_print_bucketlist: %s", kvm_geterr(kvm));
+		if (count % 2 == 0)
+			printf("\n%s  ", spaces);
+		uma_print_bucket(&ub, "");
+		printf(" ");
+		total_entries += ub.ub_entries;
+		total_cnt += ub.ub_cnt;
+		count++;
+	}
+
+	printf("\n");
+	printf("%s};  // total cnt %llu, total entries %llu\n", spaces,
+	    total_cnt, total_entries);
+}
+
+static void
+uma_print_cache(kvm_t *kvm, struct uma_cache *cache, const char *name,
+    int cpu, const char *spaces)
+{
+	struct uma_bucket ub;
+	int ret;
+
+	printf("%s%s[%d] = {\n", spaces, name, cpu);
+	printf("%s  uc_frees = %llu;\n", spaces, cache->uc_frees);
+	printf("%s  uc_allocs = %llu;\n", spaces, cache->uc_allocs);
+
+	if (cache->uc_freebucket != NULL) {
+		ret = kread(kvm, cache->uc_freebucket, &ub, sizeof(ub), 0);
+		if (ret != 0)
+			errx(-1, "uma_print_cache: %s", kvm_geterr(kvm));
+		printf("%s  uc_freebucket ", spaces);
+		uma_print_bucket(&ub, spaces);
+		printf(";\n");
+	} else
+		printf("%s  uc_freebucket = NULL;\n", spaces);
+	if (cache->uc_allocbucket != NULL) {
+		ret = kread(kvm, cache->uc_allocbucket, &ub, sizeof(ub), 0);
+		if (ret != 0)
+			errx(-1, "uma_print_cache: %s", kvm_geterr(kvm));
+		printf("%s  uc_allocbucket ", spaces);
+		uma_print_bucket(&ub, spaces);
+		printf(";\n");
+	} else
+		printf("%s  uc_allocbucket = NULL;\n", spaces);
+	printf("%s};\n", spaces);
+}
+
+int
+main(int argc, char *argv[])
+{
+	LIST_HEAD(, uma_keg) uma_kegs;
+	char name[MEMTYPE_MAXNAME];
+	struct uma_keg *kzp, kz;
+	struct uma_zone *uzp, uz;
+	kvm_t *kvm;
+	int all_cpus, cpu, mp_maxid, ret;
+
+	if (argc != 1)
+		usage();
+
+	kvm = kvm_open(NULL, NULL, NULL, 0, "umastat");
+	if (kvm == NULL)
+		err(-1, "kvm_open");
+
+	if (kvm_nlist(kvm, namelist) != 0)
+		err(-1, "kvm_nlist");
+
+	if (namelist[X_UMA_KEGS].n_type == 0 ||
+	    namelist[X_UMA_KEGS].n_value == 0)
+		errx(-1, "kvm_nlist return");
+
+	ret = kread_symbol(kvm, X_MP_MAXID, &mp_maxid, sizeof(mp_maxid), 0);
+	if (ret != 0)
+		errx(-1, "kread_symbol: %s", kvm_geterr(kvm));
+
+	printf("mp_maxid = %d\n", mp_maxid);
+
+	ret = kread_symbol(kvm, X_ALLCPU, &all_cpus, sizeof(all_cpus), 0);
+	if (ret != 0)
+		errx(-1, "kread_symbol: %s", kvm_geterr(kvm));
+
+	printf("all_cpus = %x\n", all_cpus);
+
+	ret = kread_symbol(kvm, X_UMA_KEGS, &uma_kegs, sizeof(uma_kegs), 0);
+	if (ret != 0)
+		errx(-1, "kread_symbol: %s", kvm_geterr(kvm));
+
+	for (kzp = LIST_FIRST(&uma_kegs); kzp != NULL; kzp =
+	    LIST_NEXT(&kz, uk_link)) {
+		ret = kread(kvm, kzp, &kz, sizeof(kz), 0);
+		if (ret != 0)
+			errx(-1, "kread: %s", kvm_geterr(kvm));
+		printf("Keg {\n");
+
+		printf("  uk_recurse = %d\n", kz.uk_recurse);
+		uma_print_keg_align(&kz, "  ");
+		printf("  uk_pages = %d\n", kz.uk_pages);
+		printf("  uk_free = %d\n", kz.uk_free);
+		printf("  uk_size = %d\n", kz.uk_size);
+		printf("  uk_rsize = %d\n", kz.uk_rsize);
+		printf("  uk_maxpages = %d\n", kz.uk_maxpages);
+
+		printf("  uk_pgoff = %d\n", kz.uk_pgoff);
+		printf("  uk_ppera = %d\n", kz.uk_ppera);
+		printf("  uk_ipers = %d\n", kz.uk_ipers);
+		uma_print_keg_flags(&kz, "  ");
+
+		if (LIST_FIRST(&kz.uk_zones) == NULL) {
+			printf("; No zones.\n");
+			printf("};\n");
+			continue;
+		}
+		for (uzp = LIST_FIRST(&kz.uk_zones); uzp != NULL; uzp =
+		    LIST_NEXT(&uz, uz_link)) {
+			ret = kread(kvm, uzp, &uz, sizeof(uz), 0);
+			if (ret != 0)
+				errx(-1, "kread: %s", kvm_geterr(kvm));
+			ret = kread_string(kvm, uz.uz_name, name,
+			    MEMTYPE_MAXNAME);
+			if (ret != 0)
+				errx(-1, "kread_string: %s", kvm_geterr(kvm));
+			printf("  Zone {\n");
+			printf("    uz_name = \"%s\";\n", name);
+			printf("    uz_allocs = %llu;\n", uz.uz_allocs);
+			printf("    uz_frees = %llu;\n", uz.uz_frees);
+			printf("    uz_fails = %llu;\n", uz.uz_fails);
+			printf("    uz_fills = %u;\n", uz.uz_fills);
+			printf("    uz_count = %u;\n", uz.uz_count);
+			uma_print_bucketlist(kvm,
+			    (struct bucketlist *)&uz.uz_full_bucket,
+			    "uz_full_bucket", "    ");
+			uma_print_bucketlist(kvm,
+			    (struct bucketlist *)&uz.uz_free_bucket,
+			    "uz_free_bucket", "    ");
+
+			if (!(kz.uk_flags & UMA_ZFLAG_INTERNAL)) {
+				for (cpu = 0; cpu < mp_maxid; cpu++) {
+					/* if (CPU_ABSENT(cpu)) */
+					if ((all_cpus & (1 << cpu)) == 0)
+						continue;
+					uma_print_cache(kvm, &uz.uz_cpu[cpu],
+					    "uc_cache", cpu, "    ");
+				}
+			}
+
+			printf("  };\n");
+		}
+		printf("};\n");
+	}
+
+	return (0);
+}