summaryrefslogtreecommitdiffstats
path: root/module
diff options
context:
space:
mode:
authorBrian Behlendorf <[email protected]>2009-02-25 13:20:40 -0800
committerBrian Behlendorf <[email protected]>2009-03-04 10:04:15 -0800
commitd1ff2312b05852d2190afd1da216b4e0e1b0c7d4 (patch)
tree758bd8754c8a730e769f3db8e55b799c0bddae48 /module
parent4327ac3ff93b7aa47d20c184d34fb1f2202b7ab5 (diff)
Linux VM Integration Cleanup
Remove all instances of functions being reimplemented in the SPL. When the prototypes are available in the linux headers but the function address itself is not exported use kallsyms_lookup_name() to find the address. The function name itself can them become a define which calls a function pointer. This is preferable to reimplementing the function in the SPL because it ensures we get the correct version of the function for the running kernel. This is actually pretty safe because the prototype is defined in the headers so we know we are calling the function properly. This patch also includes a rhel5 kernel patch we exports the needed symbols so we don't need to use kallsyms_lookup_name(). There are autoconf checks to detect if the symbol is exported and if so to use it directly. We should add patches for stock upstream kernels as needed if for no other reason than so we can easily track which additional symbols we needed exported. Those patches can also be used by anyone willing to rebuild their kernel, but this should not be a requirement. The rhel5 version of the export-symbols patch has been applied to the chaos kernel. Additional fixes: 1) Implement vmem_size() function using get_vmalloc_info() 2) SPL_CHECK_SYMBOL_EXPORT macro updated to use $LINUX_OBJ instead of $LINUX because Module.symvers is a build product. When $LINUX_OBJ != $LINUX we will not properly detect exported symbols. 3) SPL_LINUX_COMPILE_IFELSE macro updated to add include2 and $LINUX/include search paths to allow proper compilation when the kernel target build directory is not the source directory.
Diffstat (limited to 'module')
-rw-r--r--module/spl/spl-generic.c43
-rw-r--r--module/spl/spl-kmem.c147
-rw-r--r--module/spl/spl-proc.c57
3 files changed, 175 insertions, 72 deletions
diff --git a/module/spl/spl-generic.c b/module/spl/spl-generic.c
index 96a14c62e..c23abf6ff 100644
--- a/module/spl/spl-generic.c
+++ b/module/spl/spl-generic.c
@@ -54,6 +54,10 @@ EXPORT_SYMBOL(hw_serial);
int p0 = 0;
EXPORT_SYMBOL(p0);
+#ifndef HAVE_KALLSYMS_LOOKUP_NAME
+kallsyms_lookup_name_t spl_kallsyms_lookup_name_fn = NULL;
+#endif
+
int
highbit(unsigned long i)
{
@@ -269,6 +273,42 @@ zone_get_hostid(void *zone)
}
EXPORT_SYMBOL(zone_get_hostid);
+#ifdef HAVE_KALLSYMS_LOOKUP_NAME
+#define set_kallsyms_lookup_name() (0)
+#else
+/*
+ * Because kallsyms_lookup_name() is no longer exported in the
+ * mainline kernel we are forced to resort to somewhat drastic
+ * measures. This function replaces the functionality by performing
+ * an upcall to user space where /proc/kallsyms is consulted for
+ * the requested address.
+ */
+#define GET_KALLSYMS_ADDR_CMD \
+ "awk '{ if ( $3 == \"kallsyms_lookup_name\") { print $1 } }' " \
+ "/proc/kallsyms >/proc/sys/kernel/spl/kallsyms_lookup_name"
+
+static int
+set_kallsyms_lookup_name(void)
+{
+ char sh_path[] = "/bin/sh";
+ char *argv[] = { sh_path,
+ "-c",
+ GET_KALLSYMS_ADDR_CMD,
+ NULL };
+ char *envp[] = { "HOME=/",
+ "TERM=linux",
+ "PATH=/sbin:/usr/sbin:/bin:/usr/bin",
+ NULL };
+ int rc;
+
+ rc = call_usermodehelper(sh_path, argv, envp, 1);
+ if (rc)
+ return rc;
+
+ return spl_kmem_init_kallsyms_lookup();
+}
+#endif
+
static int __init spl_init(void)
{
int rc = 0;
@@ -297,6 +337,9 @@ static int __init spl_init(void)
if ((rc = set_hostid()))
GOTO(out7, rc = -EADDRNOTAVAIL);
+ if ((rc = set_kallsyms_lookup_name()))
+ GOTO(out7, rc = -EADDRNOTAVAIL);
+
printk("SPL: Loaded Solaris Porting Layer v%s\n", VERSION);
RETURN(rc);
out7:
diff --git a/module/spl/spl-kmem.c b/module/spl/spl-kmem.c
index 18613e799..b8b006610 100644
--- a/module/spl/spl-kmem.c
+++ b/module/spl/spl-kmem.c
@@ -79,85 +79,29 @@ EXPORT_SYMBOL(zio_alloc_arena);
vmem_t *zio_arena = NULL;
EXPORT_SYMBOL(zio_arena);
+#ifndef HAVE_GET_VMALLOC_INFO
+get_vmalloc_info_t get_vmalloc_info_fn = NULL;
+EXPORT_SYMBOL(get_vmalloc_info_fn);
+#endif /* HAVE_GET_VMALLOC_INFO */
+
#ifndef HAVE_FIRST_ONLINE_PGDAT
-struct pglist_data *
-first_online_pgdat(void)
-{
- return NODE_DATA(first_online_node);
-}
-EXPORT_SYMBOL(first_online_pgdat);
+first_online_pgdat_t first_online_pgdat_fn = NULL;
+EXPORT_SYMBOL(first_online_pgdat_fn);
#endif /* HAVE_FIRST_ONLINE_PGDAT */
#ifndef HAVE_NEXT_ONLINE_PGDAT
-struct pglist_data *
-next_online_pgdat(struct pglist_data *pgdat)
-{
- int nid = next_online_node(pgdat->node_id);
-
- if (nid == MAX_NUMNODES)
- return NULL;
-
- return NODE_DATA(nid);
-}
-EXPORT_SYMBOL(next_online_pgdat);
+next_online_pgdat_t next_online_pgdat_fn = NULL;
+EXPORT_SYMBOL(next_online_pgdat_fn);
#endif /* HAVE_NEXT_ONLINE_PGDAT */
#ifndef HAVE_NEXT_ZONE
-struct zone *
-next_zone(struct zone *zone)
-{
- pg_data_t *pgdat = zone->zone_pgdat;
-
- if (zone < pgdat->node_zones + MAX_NR_ZONES - 1)
- zone++;
- else {
- pgdat = next_online_pgdat(pgdat);
- if (pgdat)
- zone = pgdat->node_zones;
- else
- zone = NULL;
- }
- return zone;
-}
-EXPORT_SYMBOL(next_zone);
+next_zone_t next_zone_fn = NULL;
+EXPORT_SYMBOL(next_zone_fn);
#endif /* HAVE_NEXT_ZONE */
#ifndef HAVE_GET_ZONE_COUNTS
-void
-__get_zone_counts(unsigned long *active, unsigned long *inactive,
- unsigned long *free, struct pglist_data *pgdat)
-{
- struct zone *zones = pgdat->node_zones;
- int i;
-
- *active = 0;
- *inactive = 0;
- *free = 0;
- for (i = 0; i < MAX_NR_ZONES; i++) {
- *active += zones[i].nr_active;
- *inactive += zones[i].nr_inactive;
- *free += zones[i].free_pages;
- }
-}
-
-void
-get_zone_counts(unsigned long *active, unsigned long *inactive,
- unsigned long *free)
-{
- struct pglist_data *pgdat;
-
- *active = 0;
- *inactive = 0;
- *free = 0;
- for_each_online_pgdat(pgdat) {
- unsigned long l, m, n;
- __get_zone_counts(&l, &m, &n, pgdat);
- *active += l;
- *inactive += m;
- *free += n;
- }
-}
-EXPORT_SYMBOL(get_zone_counts);
+get_zone_counts_t get_zone_counts_fn = NULL;
+EXPORT_SYMBOL(get_zone_counts_fn);
#endif /* HAVE_GET_ZONE_COUNTS */
pgcnt_t
@@ -177,11 +121,20 @@ EXPORT_SYMBOL(spl_kmem_availrmem);
size_t
vmem_size(vmem_t *vmp, int typemask)
{
- /* Arena's unsupported */
+ struct vmalloc_info vmi;
+ size_t size = 0;
+
ASSERT(vmp == NULL);
ASSERT(typemask & (VMEM_ALLOC | VMEM_FREE));
- return 0;
+ get_vmalloc_info(&vmi);
+ if (typemask & VMEM_ALLOC)
+ size += (size_t)vmi.used;
+
+ if (typemask & VMEM_FREE)
+ size += (size_t)(VMALLOC_TOTAL - vmi.used);
+
+ return size;
}
EXPORT_SYMBOL(vmem_size);
@@ -1812,6 +1765,57 @@ spl_kmem_init_globals(void)
swapfs_reserve = MIN(4*1024*1024 / PAGE_SIZE, physmem / 16);
}
+/*
+ * Called at module init when it is safe to use spl_kallsyms_lookup_name()
+ */
+int
+spl_kmem_init_kallsyms_lookup(void)
+{
+#ifndef HAVE_GET_VMALLOC_INFO
+ get_vmalloc_info_fn = (get_vmalloc_info_t)
+ spl_kallsyms_lookup_name("get_vmalloc_info");
+ if (!get_vmalloc_info_fn)
+ return -EFAULT;
+#endif /* HAVE_GET_VMALLOC_INFO */
+
+#ifndef HAVE_FIRST_ONLINE_PGDAT
+ first_online_pgdat_fn = (first_online_pgdat_t)
+ spl_kallsyms_lookup_name("first_online_pgdat");
+ if (!first_online_pgdat_fn)
+ return -EFAULT;
+#endif /* HAVE_FIRST_ONLINE_PGDAT */
+
+#ifndef HAVE_NEXT_ONLINE_PGDAT
+ next_online_pgdat_fn = (next_online_pgdat_t)
+ spl_kallsyms_lookup_name("next_online_pgdat");
+ if (!next_online_pgdat_fn)
+ return -EFAULT;
+#endif /* HAVE_NEXT_ONLINE_PGDAT */
+
+#ifndef HAVE_NEXT_ZONE
+ next_zone_fn = (next_zone_t)
+ spl_kallsyms_lookup_name("next_zone");
+ if (!next_zone_fn)
+ return -EFAULT;
+#endif /* HAVE_NEXT_ZONE */
+
+#ifndef HAVE_GET_ZONE_COUNTS
+ get_zone_counts_fn = (get_zone_counts_t)
+ spl_kallsyms_lookup_name("get_zone_counts");
+ if (!get_zone_counts_fn)
+ return -EFAULT;
+#endif /* HAVE_GET_ZONE_COUNTS */
+
+ /*
+ * It is now safe to initialize the global tunings which rely on
+ * the use of the for_each_zone() macro. This macro in turns
+ * depends on the *_pgdat symbols which are now available.
+ */
+ spl_kmem_init_globals();
+
+ return 0;
+}
+
int
spl_kmem_init(void)
{
@@ -1820,7 +1824,6 @@ spl_kmem_init(void)
init_rwsem(&spl_kmem_cache_sem);
INIT_LIST_HEAD(&spl_kmem_cache_list);
- spl_kmem_init_globals();
#ifdef HAVE_SET_SHRINKER
spl_kmem_cache_shrinker = set_shrinker(KMC_DEFAULT_SEEKS,
diff --git a/module/spl/spl-proc.c b/module/spl/spl-proc.c
index 90c89ce31..d0ce23194 100644
--- a/module/spl/spl-proc.c
+++ b/module/spl/spl-proc.c
@@ -68,6 +68,7 @@ struct proc_dir_entry *proc_spl_kstat = NULL;
#define CTL_VERSION CTL_UNNUMBERED /* Version */
#define CTL_HOSTID CTL_UNNUMBERED /* Host id by /usr/bin/hostid */
#define CTL_HW_SERIAL CTL_UNNUMBERED /* HW serial number by hostid */
+#define CTL_KALLSYMS CTL_UNNUMBERED /* kallsyms_lookup_name addr */
#define CTL_DEBUG_SUBSYS CTL_UNNUMBERED /* Debug subsystem */
#define CTL_DEBUG_MASK CTL_UNNUMBERED /* Debug mask */
@@ -123,6 +124,7 @@ enum {
CTL_VERSION = 1, /* Version */
CTL_HOSTID, /* Host id reported by /usr/bin/hostid */
CTL_HW_SERIAL, /* Hardware serial number from hostid */
+ CTL_KALLSYMS, /* Address of kallsyms_lookup_name */
CTL_DEBUG_SUBSYS, /* Debug subsystem */
CTL_DEBUG_MASK, /* Debug mask */
@@ -489,6 +491,51 @@ proc_dohostid(struct ctl_table *table, int write, struct file *filp,
RETURN(rc);
}
+#ifndef HAVE_KALLSYMS_LOOKUP_NAME
+static int
+proc_dokallsyms_lookup_name(struct ctl_table *table, int write,
+ struct file *filp, void __user *buffer,
+ size_t *lenp, loff_t *ppos) {
+ int len, rc = 0;
+ char *end, str[32];
+ ENTRY;
+
+ if (write) {
+ /* This may only be set once at module load time */
+ if (spl_kallsyms_lookup_name_fn)
+ RETURN(-EEXIST);
+
+ /* We can't use proc_doulongvec_minmax() in the write
+ * case hear because the address while a hex value has no
+ * leading 0x which confuses the helper function. */
+ rc = proc_copyin_string(str, sizeof(str), buffer, *lenp);
+ if (rc < 0)
+ RETURN(rc);
+
+ spl_kallsyms_lookup_name_fn =
+ (kallsyms_lookup_name_t)simple_strtoul(str, &end, 16);
+ if (str == end)
+ RETURN(-EINVAL);
+
+ *ppos += *lenp;
+ } else {
+ len = snprintf(str, sizeof(str), "%lx",
+ (unsigned long)spl_kallsyms_lookup_name_fn);
+ if (*ppos >= len)
+ rc = 0;
+ else
+ rc = proc_copyout_string(buffer,*lenp,str+*ppos,"\n");
+
+ if (rc >= 0) {
+ *lenp = rc;
+ *ppos += rc;
+ }
+ }
+
+ RETURN(rc);
+}
+#endif /* HAVE_KALLSYMS_LOOKUP_NAME */
+
static int
proc_doavailrmem(struct ctl_table *table, int write, struct file *filp,
void __user *buffer, size_t *lenp, loff_t *ppos)
@@ -1018,6 +1065,16 @@ static struct ctl_table spl_table[] = {
.mode = 0444,
.proc_handler = &proc_dostring,
},
+#ifndef HAVE_KALLSYMS_LOOKUP_NAME
+ {
+ .ctl_name = CTL_KALLSYMS,
+ .procname = "kallsyms_lookup_name",
+ .data = &spl_kallsyms_lookup_name_fn,
+ .maxlen = sizeof(unsigned long),
+ .mode = 0644,
+ .proc_handler = &proc_dokallsyms_lookup_name,
+ },
+#endif
{
.ctl_name = CTL_SPL_DEBUG,
.procname = "debug",