* Copyright (c) 1999, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
- * Copyright (c) 2011, 2015, Intel Corporation.
+ * Copyright (c) 2011, 2017, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
#include <linux/miscdevice.h>
#include <linux/user_namespace.h>
-#ifdef HAVE_UIDGID_HEADER
-# include <linux/uidgid.h>
-#endif
+#include <linux/uidgid.h>
#include <linux/atomic.h>
#include <linux/list.h>
EXPORT_SYMBOL(obd_dump_on_timeout);
unsigned int obd_dump_on_eviction;
EXPORT_SYMBOL(obd_dump_on_eviction);
+unsigned int obd_lbug_on_eviction;
+EXPORT_SYMBOL(obd_lbug_on_eviction);
unsigned long obd_max_dirty_pages;
EXPORT_SYMBOL(obd_max_dirty_pages);
atomic_long_t obd_dirty_pages;
int at_extra = 30;
EXPORT_SYMBOL(at_extra);
-atomic_long_t obd_dirty_transit_pages;
-EXPORT_SYMBOL(obd_dirty_transit_pages);
-
#ifdef CONFIG_PROC_FS
struct lprocfs_stats *obd_memory = NULL;
EXPORT_SYMBOL(obd_memory);
RETURN(rc);
}
-int class_handle_ioctl(unsigned int cmd, unsigned long arg)
+#define OBD_MAX_IOCTL_BUFFER 8192
+
+static int obd_ioctl_is_invalid(struct obd_ioctl_data *data)
{
- char *buf = NULL;
- struct obd_ioctl_data *data;
- struct libcfs_debug_ioctl_data *debug_data;
- struct obd_device *obd = NULL;
- int err = 0, len = 0;
- ENTRY;
+ if (data->ioc_len > BIT(30)) {
+ CERROR("OBD ioctl: ioc_len larger than 1<<30\n");
+ return 1;
+ }
- /* only for debugging */
- if (cmd == LIBCFS_IOC_DEBUG_MASK) {
- debug_data = (struct libcfs_debug_ioctl_data*)arg;
- libcfs_subsystem_debug = debug_data->subs;
- libcfs_debug = debug_data->debug;
- return 0;
- }
+ if (data->ioc_inllen1 > BIT(30)) {
+ CERROR("OBD ioctl: ioc_inllen1 larger than 1<<30\n");
+ return 1;
+ }
+
+ if (data->ioc_inllen2 > BIT(30)) {
+ CERROR("OBD ioctl: ioc_inllen2 larger than 1<<30\n");
+ return 1;
+ }
+
+ if (data->ioc_inllen3 > BIT(30)) {
+ CERROR("OBD ioctl: ioc_inllen3 larger than 1<<30\n");
+ return 1;
+ }
+
+ if (data->ioc_inllen4 > BIT(30)) {
+ CERROR("OBD ioctl: ioc_inllen4 larger than 1<<30\n");
+ return 1;
+ }
+
+ if (data->ioc_inlbuf1 && data->ioc_inllen1 == 0) {
+ CERROR("OBD ioctl: inlbuf1 pointer but 0 length\n");
+ return 1;
+ }
+
+ if (data->ioc_inlbuf2 && data->ioc_inllen2 == 0) {
+ CERROR("OBD ioctl: inlbuf2 pointer but 0 length\n");
+ return 1;
+ }
+
+ if (data->ioc_inlbuf3 && data->ioc_inllen3 == 0) {
+ CERROR("OBD ioctl: inlbuf3 pointer but 0 length\n");
+ return 1;
+ }
+
+ if (data->ioc_inlbuf4 && data->ioc_inllen4 == 0) {
+ CERROR("OBD ioctl: inlbuf4 pointer but 0 length\n");
+ return 1;
+ }
+
+ if (data->ioc_pbuf1 && data->ioc_plen1 == 0) {
+ CERROR("OBD ioctl: pbuf1 pointer but 0 length\n");
+ return 1;
+ }
+
+ if (data->ioc_pbuf2 && data->ioc_plen2 == 0) {
+ CERROR("OBD ioctl: pbuf2 pointer but 0 length\n");
+ return 1;
+ }
+
+ if (!data->ioc_pbuf1 && data->ioc_plen1 != 0) {
+ CERROR("OBD ioctl: plen1 set but NULL pointer\n");
+ return 1;
+ }
+
+ if (!data->ioc_pbuf2 && data->ioc_plen2 != 0) {
+ CERROR("OBD ioctl: plen2 set but NULL pointer\n");
+ return 1;
+ }
+
+ if (obd_ioctl_packlen(data) > data->ioc_len) {
+ CERROR("OBD ioctl: packlen exceeds ioc_len (%d > %d)\n",
+ obd_ioctl_packlen(data), data->ioc_len);
+ return 1;
+ }
+
+ return 0;
+}
- CDEBUG(D_IOCTL, "cmd = %x\n", cmd);
+/* buffer MUST be at least the size of obd_ioctl_hdr */
+int obd_ioctl_getdata(char **buf, int *len, void __user *arg)
+{
+ struct obd_ioctl_hdr hdr;
+ struct obd_ioctl_data *data;
+ int offset = 0;
+
+ ENTRY;
+ if (copy_from_user(&hdr, arg, sizeof(hdr)))
+ RETURN(-EFAULT);
+
+ if (hdr.ioc_version != OBD_IOCTL_VERSION) {
+ CERROR("Version mismatch kernel (%x) vs application (%x)\n",
+ OBD_IOCTL_VERSION, hdr.ioc_version);
+ RETURN(-EINVAL);
+ }
+
+ if (hdr.ioc_len > OBD_MAX_IOCTL_BUFFER) {
+ CERROR("User buffer len %d exceeds %d max buffer\n",
+ hdr.ioc_len, OBD_MAX_IOCTL_BUFFER);
+ RETURN(-EINVAL);
+ }
+
+ if (hdr.ioc_len < sizeof(struct obd_ioctl_data)) {
+ CERROR("User buffer too small for ioctl (%d)\n", hdr.ioc_len);
+ RETURN(-EINVAL);
+ }
+
+ /* When there are lots of processes calling vmalloc on multi-core
+ * system, the high lock contention will hurt performance badly,
+ * obdfilter-survey is an example, which relies on ioctl. So we'd
+ * better avoid vmalloc on ioctl path. LU-66
+ */
+ OBD_ALLOC_LARGE(*buf, hdr.ioc_len);
+ if (!*buf) {
+ CERROR("Cannot allocate control buffer of len %d\n",
+ hdr.ioc_len);
+ RETURN(-EINVAL);
+ }
+ *len = hdr.ioc_len;
+ data = (struct obd_ioctl_data *)*buf;
+
+ if (copy_from_user(*buf, arg, hdr.ioc_len)) {
+ OBD_FREE_LARGE(*buf, hdr.ioc_len);
+ RETURN(-EFAULT);
+ }
+
+ if (obd_ioctl_is_invalid(data)) {
+ CERROR("ioctl not correctly formatted\n");
+ OBD_FREE_LARGE(*buf, hdr.ioc_len);
+ RETURN(-EINVAL);
+ }
+
+ if (data->ioc_inllen1) {
+ data->ioc_inlbuf1 = &data->ioc_bulk[0];
+ offset += cfs_size_round(data->ioc_inllen1);
+ }
+
+ if (data->ioc_inllen2) {
+ data->ioc_inlbuf2 = &data->ioc_bulk[0] + offset;
+ offset += cfs_size_round(data->ioc_inllen2);
+ }
+
+ if (data->ioc_inllen3) {
+ data->ioc_inlbuf3 = &data->ioc_bulk[0] + offset;
+ offset += cfs_size_round(data->ioc_inllen3);
+ }
+
+ if (data->ioc_inllen4)
+ data->ioc_inlbuf4 = &data->ioc_bulk[0] + offset;
+
+ RETURN(0);
+}
+EXPORT_SYMBOL(obd_ioctl_getdata);
+
+int class_handle_ioctl(unsigned int cmd, unsigned long arg)
+{
+ char *buf = NULL;
+ struct obd_ioctl_data *data;
+ struct obd_device *obd = NULL;
+ int err = 0, len = 0;
+
+ ENTRY;
+ CDEBUG(D_IOCTL, "cmd = %x\n", cmd);
if (obd_ioctl_getdata(&buf, &len, (void __user *)arg)) {
- CERROR("OBD ioctl: data error\n");
- RETURN(-EINVAL);
- }
- data = (struct obd_ioctl_data *)buf;
+ CERROR("OBD ioctl: data error\n");
+ RETURN(-EINVAL);
+ }
+ data = (struct obd_ioctl_data *)buf;
switch (cmd) {
case OBD_IOC_PROCESS_CFG: {
if (!obd)
GOTO(out, err = -ENOENT);
- if (obd->obd_stopping)
- status = "ST";
- else if (obd->obd_set_up)
- status = "UP";
- else if (obd->obd_attached)
- status = "AT";
- else
- status = "--";
+ if (obd->obd_stopping)
+ status = "ST";
+ else if (obd->obd_inactive)
+ status = "IN";
+ else if (obd->obd_set_up)
+ status = "UP";
+ else if (obd->obd_attached)
+ status = "AT";
+ else
+ status = "--";
+
str = (char *)data->ioc_bulk;
snprintf(str, len - sizeof(*data), "%3d %s %s %s %s %d",
(int)index, status, obd->obd_type->typ_name,
RETURN(err);
} /* class_handle_ioctl */
-static int obd_init_checks(void)
+/* to control /dev/obd */
+static long obd_class_ioctl(struct file *filp, unsigned int cmd,
+ unsigned long arg)
+{
+ int err = 0;
+
+ ENTRY;
+ /* Allow non-root access for some limited ioctls */
+ if (!cfs_capable(CFS_CAP_SYS_ADMIN))
+ RETURN(err = -EACCES);
+
+ if ((cmd & 0xffffff00) == ((int)'T') << 8) /* ignore all tty ioctls */
+ RETURN(err = -ENOTTY);
+
+ err = class_handle_ioctl(cmd, (unsigned long)arg);
+
+ RETURN(err);
+}
+
+/* declare character device */
+static struct file_operations obd_psdev_fops = {
+ .owner = THIS_MODULE,
+ .unlocked_ioctl = obd_class_ioctl, /* unlocked_ioctl */
+};
+
+/* modules setup */
+struct miscdevice obd_psdev = {
+ .minor = MISC_DYNAMIC_MINOR,
+ .name = OBD_DEV_NAME,
+ .fops = &obd_psdev_fops,
+};
+
+#define test_string_to_size_err(value, expect, def_unit, __rc) \
+({ \
+ u64 __size; \
+ int __ret; \
+ \
+ BUILD_BUG_ON(sizeof(value) >= 23); \
+ __ret = sysfs_memparse(value, sizeof(value) - 1, &__size, def_unit); \
+ if (__ret != __rc) \
+ CERROR("string_helper: parsing '%s' expect rc %d != got %d\n", \
+ value, __rc, __ret); \
+ else if (!__ret && (u64)expect != __size) \
+ CERROR("string_helper: parsing '%s' expect %llu != got %llu\n",\
+ value, (u64)expect, __size); \
+ __ret; \
+})
+#define test_string_to_size_one(value, expect, def_unit) \
+ test_string_to_size_err(value, expect, def_unit, 0)
+
+static int __init obd_init_checks(void)
{
- __u64 u64val, div64val;
- char buf[64];
- int len, ret = 0;
+ __u64 u64val, div64val;
+ char buf[64];
+ int len, ret = 0;
CDEBUG(D_INFO, "OBD_OBJECT_EOF = %#llx\n", (__u64)OBD_OBJECT_EOF);
- u64val = OBD_OBJECT_EOF;
+ u64val = OBD_OBJECT_EOF;
CDEBUG(D_INFO, "u64val OBD_OBJECT_EOF = %#llx\n", u64val);
- if (u64val != OBD_OBJECT_EOF) {
+ if (u64val != OBD_OBJECT_EOF) {
CERROR("__u64 %#llx(%d) != 0xffffffffffffffff\n",
- u64val, (int)sizeof(u64val));
- ret = -EINVAL;
- }
+ u64val, (int)sizeof(u64val));
+ ret = -EINVAL;
+ }
len = snprintf(buf, sizeof(buf), "%#llx", u64val);
- if (len != 18) {
- CWARN("u64 hex wrong length! strlen(%s)=%d != 18\n", buf, len);
- ret = -EINVAL;
- }
+ if (len != 18) {
+ CERROR("u64 hex wrong length, strlen(%s)=%d != 18\n", buf, len);
+ ret = -EINVAL;
+ }
- div64val = OBD_OBJECT_EOF;
+ div64val = OBD_OBJECT_EOF;
CDEBUG(D_INFO, "u64val OBD_OBJECT_EOF = %#llx\n", u64val);
- if (u64val != OBD_OBJECT_EOF) {
+ if (u64val != OBD_OBJECT_EOF) {
CERROR("__u64 %#llx(%d) != 0xffffffffffffffff\n",
- u64val, (int)sizeof(u64val));
- ret = -EOVERFLOW;
- }
- if (u64val >> 8 != OBD_OBJECT_EOF >> 8) {
+ u64val, (int)sizeof(u64val));
+ ret = -EOVERFLOW;
+ }
+ if (u64val >> 8 != OBD_OBJECT_EOF >> 8) {
CERROR("__u64 %#llx(%d) != 0xffffffffffffffff\n",
- u64val, (int)sizeof(u64val));
- return -EOVERFLOW;
- }
- if (do_div(div64val, 256) != (u64val & 255)) {
+ u64val, (int)sizeof(u64val));
+ ret = -EOVERFLOW;
+ }
+ if (do_div(div64val, 256) != (u64val & 255)) {
CERROR("do_div(%#llx,256) != %llu\n", u64val, u64val & 255);
- return -EOVERFLOW;
- }
- if (u64val >> 8 != div64val) {
+ ret = -EOVERFLOW;
+ }
+ if (u64val >> 8 != div64val) {
CERROR("do_div(%#llx,256) %llu != %llu\n",
- u64val, div64val, u64val >> 8);
- return -EOVERFLOW;
- }
+ u64val, div64val, u64val >> 8);
+ ret = -EOVERFLOW;
+ }
len = snprintf(buf, sizeof(buf), "%#llx", u64val);
- if (len != 18) {
- CWARN("u64 hex wrong length! strlen(%s)=%d != 18\n", buf, len);
- ret = -EINVAL;
- }
+ if (len != 18) {
+ CERROR("u64 hex wrong length! strlen(%s)=%d != 18\n", buf, len);
+ ret = -EINVAL;
+ }
len = snprintf(buf, sizeof(buf), "%llu", u64val);
- if (len != 20) {
- CWARN("u64 wrong length! strlen(%s)=%d != 20\n", buf, len);
- ret = -EINVAL;
- }
+ if (len != 20) {
+ CERROR("u64 wrong length! strlen(%s)=%d != 20\n", buf, len);
+ ret = -EINVAL;
+ }
len = snprintf(buf, sizeof(buf), "%lld", u64val);
- if (len != 2) {
- CWARN("s64 wrong length! strlen(%s)=%d != 2\n", buf, len);
- ret = -EINVAL;
- }
+ if (len != 2) {
+ CERROR("s64 wrong length! strlen(%s)=%d != 2\n", buf, len);
+ ret = -EINVAL;
+ }
if ((u64val & ~PAGE_MASK) >= PAGE_SIZE) {
- CWARN("mask failed: u64val %llu >= %llu\n", u64val,
- (__u64)PAGE_SIZE);
- ret = -EINVAL;
- }
+ CERROR("mask failed: u64val %llu >= %llu\n", u64val,
+ (__u64)PAGE_SIZE);
+ ret = -EINVAL;
+ }
+ if (ret)
+ RETURN(ret);
- return ret;
+ /* invalid string */
+ if (!test_string_to_size_err("256B34", 256, "B", -EINVAL)) {
+ CERROR("string_helpers: format should be number then units\n");
+ ret = -EINVAL;
+ }
+ if (!test_string_to_size_err("132OpQ", 132, "B", -EINVAL)) {
+ CERROR("string_helpers: invalid units should be rejected\n");
+ ret = -EINVAL;
+ }
+ if (!test_string_to_size_err("1.82B", 1, "B", -EINVAL)) {
+ CERROR("string_helpers: 'B' with '.' should be invalid\n");
+ ret = -EINVAL;
+ }
+ if (test_string_to_size_one("343\n", 343, "B")) {
+ CERROR("string_helpers: should ignore newline\n");
+ ret = -EINVAL;
+ }
+ if (ret)
+ RETURN(ret);
+
+ /* memparse unit handling */
+ ret = 0;
+ ret += test_string_to_size_one("0B", 0, "B");
+ ret += test_string_to_size_one("512B", 512, "B");
+ ret += test_string_to_size_one("1.067kB", 1067, "B");
+ ret += test_string_to_size_one("1.042KiB", 1067, "B");
+ ret += test_string_to_size_one("8", 8388608, "M");
+ ret += test_string_to_size_one("65536", 65536, "B");
+ ret += test_string_to_size_one("128", 131072, "K");
+ ret += test_string_to_size_one("1M", 1048576, "B");
+ ret += test_string_to_size_one("0.5T", 549755813888ULL, "T");
+ ret += test_string_to_size_one("256.5G", 275414777856ULL, "G");
+ if (ret)
+ RETURN(ret);
+
+ /* string helper values */
+ ret += test_string_to_size_one("16", 16777216, "MiB");
+ ret += test_string_to_size_one("8.39MB", 8390000, "MiB");
+ ret += test_string_to_size_one("8.00MiB", 8388608, "MiB");
+ ret += test_string_to_size_one("256GB", 256000000000ULL, "GiB");
+ ret += test_string_to_size_one("238.731GiB", 256335459385ULL, "GiB");
+ if (ret)
+ RETURN(ret);
+
+ /* huge values */
+ ret += test_string_to_size_one("0.4TB", 400000000000ULL, "TiB");
+ ret += test_string_to_size_one("12.5TiB", 13743895347200ULL, "TiB");
+ ret += test_string_to_size_one("2PB", 2000000000000000ULL, "PiB");
+ ret += test_string_to_size_one("16PiB", 18014398509481984ULL, "PiB");
+ if (ret)
+ RETURN(ret);
+
+ /* huge values should overflow */
+ if (!test_string_to_size_err("1000EiB", 0, "EiB", -EOVERFLOW)) {
+ CERROR("string_helpers: failed to detect binary overflow\n");
+ ret = -EINVAL;
+ }
+ if (!test_string_to_size_err("1000EB", 0, "EiB", -EOVERFLOW)) {
+ CERROR("string_helpers: failed to detect decimal overflow\n");
+ ret = -EINVAL;
+ }
+
+ return ret;
}
static int __init obdclass_init(void)
libcfs_kkuc_init();
err = obd_init_checks();
- if (err == -EOVERFLOW)
+ if (err)
return err;
#ifdef CONFIG_PROC_FS
/* Default the dirty page cache cap to 1/2 of system memory.
* For clients with less memory, a larger fraction is needed
* for other purposes (mostly for BGL). */
- if (totalram_pages <= 512 << (20 - PAGE_SHIFT))
- obd_max_dirty_pages = totalram_pages / 4;
+ if (cfs_totalram_pages() <= 512 << (20 - PAGE_SHIFT))
+ obd_max_dirty_pages = cfs_totalram_pages() / 4;
else
- obd_max_dirty_pages = totalram_pages / 2;
+ obd_max_dirty_pages = cfs_totalram_pages() / 2;
err = obd_init_caches();
if (err)
lu_global_fini();
cleanup_class_procfs:
- obd_sysctl_clean();
class_procfs_clean();
cleanup_caches:
cl_global_fini();
lu_global_fini();
- obd_cleanup_caches();
- obd_sysctl_clean();
+ obd_cleanup_caches();
- class_procfs_clean();
+ class_procfs_clean();
- class_handle_cleanup();
+ class_handle_cleanup();
class_del_uuid(NULL); /* Delete all UUIDs. */
- obd_zombie_impexp_stop();
+ obd_zombie_impexp_stop();
#ifdef CONFIG_PROC_FS
memory_leaked = obd_memory_sum();
memory_max = obd_memory_max();
lprocfs_free_stats(&obd_memory);
+ /* the below message is checked in test-framework.sh check_mem_leak() */
CDEBUG((memory_leaked) ? D_ERROR : D_INFO,
"obd_memory max: %llu, leaked: %llu\n",
memory_max, memory_leaked);
EXIT;
}
+void obd_heat_clear(struct obd_heat_instance *instance, int count)
+{
+ ENTRY;
+
+ memset(instance, 0, sizeof(*instance) * count);
+ RETURN_EXIT;
+}
+EXPORT_SYMBOL(obd_heat_clear);
+
+/*
+ * The file heat is calculated for every time interval period I. The access
+ * frequency during each period is counted. The file heat is only recalculated
+ * at the end of a time period. And a percentage of the former file heat is
+ * lost when recalculated. The recursion formula to calculate the heat of the
+ * file f is as follow:
+ *
+ * Hi+1(f) = (1-P)*Hi(f)+ P*Ci
+ *
+ * Where Hi is the heat value in the period between time points i*I and
+ * (i+1)*I; Ci is the access count in the period; the symbol P refers to the
+ * weight of Ci. The larger the value the value of P is, the more influence Ci
+ * has on the file heat.
+ */
+void obd_heat_decay(struct obd_heat_instance *instance, __u64 time_second,
+ unsigned int weight, unsigned int period_second)
+{
+ u64 second;
+
+ ENTRY;
+
+ if (instance->ohi_time_second > time_second) {
+ obd_heat_clear(instance, 1);
+ RETURN_EXIT;
+ }
+
+ if (instance->ohi_time_second == 0)
+ RETURN_EXIT;
+
+ for (second = instance->ohi_time_second + period_second;
+ second < time_second;
+ second += period_second) {
+ instance->ohi_heat = instance->ohi_heat *
+ (256 - weight) / 256 +
+ instance->ohi_count * weight / 256;
+ instance->ohi_count = 0;
+ instance->ohi_time_second = second;
+ }
+ RETURN_EXIT;
+}
+EXPORT_SYMBOL(obd_heat_decay);
+
+__u64 obd_heat_get(struct obd_heat_instance *instance, unsigned int time_second,
+ unsigned int weight, unsigned int period_second)
+{
+ ENTRY;
+
+ obd_heat_decay(instance, time_second, weight, period_second);
+
+ if (instance->ohi_count == 0)
+ RETURN(instance->ohi_heat);
+
+ RETURN(instance->ohi_heat * (256 - weight) / 256 +
+ instance->ohi_count * weight / 256);
+}
+EXPORT_SYMBOL(obd_heat_get);
+
+void obd_heat_add(struct obd_heat_instance *instance,
+ unsigned int time_second, __u64 count,
+ unsigned int weight, unsigned int period_second)
+{
+ ENTRY;
+
+ obd_heat_decay(instance, time_second, weight, period_second);
+ if (instance->ohi_time_second == 0) {
+ instance->ohi_time_second = time_second;
+ instance->ohi_heat = 0;
+ instance->ohi_count = count;
+ } else {
+ instance->ohi_count += count;
+ }
+ RETURN_EXIT;
+}
+EXPORT_SYMBOL(obd_heat_add);
+
MODULE_AUTHOR("OpenSFS, Inc. <http://www.lustre.org/>");
MODULE_DESCRIPTION("Lustre Class Driver");
MODULE_VERSION(LUSTRE_VERSION_STRING);
git://git.whamcloud.com / fs / lustre-release.git / blobdiff