Skip to contentSkip to conversation

Add online file check feature

PATCH Superseded Gang He 6 participants 40 messages Oct 28, 2015 – Dec 18, 2015 OCFS2

v2 of 4 Superseded Latest is v4 →

lore
Share & tools
lore

Gang He

author
#0 lore raw

v2 cover Add online file check feature

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Compare with first version, I use strncasecmp instead of double strncmp functions. Second, update the source file contribution vendor.

Gang He (4):
ocfs2: export ocfs2_kset for online file check
ocfs2: sysfile interfaces for online file check
ocfs2: create/remove sysfile for online file check
ocfs2: check/fix inode block for online file check

 fs/ocfs2/Makefile      |   3 +-
 fs/ocfs2/filecheck.c   | 566 +++++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/filecheck.h   |  48 +++++
 fs/ocfs2/inode.c       | 196 ++++++++++++++++-
 fs/ocfs2/inode.h       |   3 +
 fs/ocfs2/ocfs2_trace.h |   2 +
 fs/ocfs2/stackglue.c   |   3 +-
 fs/ocfs2/stackglue.h   |   2 +
 fs/ocfs2/super.c       |   5 +
 9 files changed, 820 insertions(+), 8 deletions(-)
 create mode 100644 fs/ocfs2/filecheck.c
 create mode 100644 fs/ocfs2/filecheck.h
Signature · 1 line
2.1.2
author
#1 lore raw

PATCH v2 1/4 ocfs2: export ocfs2_kset for online file check

Export ocfs2_kset object from ocfs2_stackglue kernel module, then online file check code will create the related sysfiles under ocfs2_kset object.

---
 fs/ocfs2/stackglue.c | 3 ++-
 fs/ocfs2/stackglue.h | 2 ++
 2 files changed, 4 insertions(+), 1 deletion(-)
fs/ocfs2/stackglue.c +2 -1
index 5d965e8..13219ed 100644
@@ -629,7 +629,8 @@ static struct attribute_group ocfs2_attr_group = {
 	.attrs = ocfs2_attrs,
 };
 
-static struct kset *ocfs2_kset;
+struct kset *ocfs2_kset;
+EXPORT_SYMBOL_GPL(ocfs2_kset);
 
 static void ocfs2_sysfs_exit(void)
 {
fs/ocfs2/stackglue.h +2 -0
index 66334a3..f2dce10 100644
@@ -298,4 +298,6 @@ void ocfs2_stack_glue_set_max_proto_version(struct ocfs2_protocol_version *max_p
 int ocfs2_stack_glue_register(struct ocfs2_stack_plugin *plugin);
 void ocfs2_stack_glue_unregister(struct ocfs2_stack_plugin *plugin);
 
+extern struct kset *ocfs2_kset;
+
 #endif  /* STACKGLUE_H */
Signature · 1 line
2.1.2
Reviewed-by on v2 1/4
#2 lore raw

Re: [PATCH v2 1/4] ocfs2: export ocfs2_kset for online file check

Quoted 5 lines · Gang He

Export ocfs2_kset object from ocfs2_stackglue kernel module, then online file check code will create the related sysfiles under ocfs2_kset object.

--
Mark Fasheh

author
#3 lore raw

PATCH v2 2/4 ocfs2: sysfile interfaces for online file check

In reply to Gang He

Implement online file check sysfile interfaces, e.g. how to create the related sysfile according to device name, how to display/handle file check request from the sysfile.

---
 fs/ocfs2/Makefile    |   3 +-
 fs/ocfs2/filecheck.c | 566 +++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/filecheck.h |  48 +++++
 fs/ocfs2/inode.h     |   3 +
 4 files changed, 619 insertions(+), 1 deletion(-)
 create mode 100644 fs/ocfs2/filecheck.c
 create mode 100644 fs/ocfs2/filecheck.h
fs/ocfs2/Makefile +2 -1
index ce210d4..e27e652 100644
@@ -41,7 +41,8 @@ ocfs2-objs := \
 	quota_local.o		\
 	quota_global.o		\
 	xattr.o			\
-	acl.o
+	acl.o	\
+	filecheck.o
 
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
fs/ocfs2/filecheck.c new file +566 -0

566 lines not on this page · raw message

fs/ocfs2/filecheck.h new file +48 -0

48 lines not on this page · raw message

fs/ocfs2/inode.h +3 -0
index 5e86b24..abd1018 100644
@@ -139,6 +139,9 @@ int ocfs2_drop_inode(struct inode *inode);
 /* Flags for ocfs2_iget() */
 #define OCFS2_FI_FLAG_SYSFILE		0x1
 #define OCFS2_FI_FLAG_ORPHAN_RECOVERY	0x2
+#define OCFS2_FI_FLAG_FILECHECK_CHK	0x4
+#define OCFS2_FI_FLAG_FILECHECK_FIX	0x8
+
 struct inode *ocfs2_ilookup(struct super_block *sb, u64 feoff);
 struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned flags,
 			 int sysfile_type);
Signature · 1 line
2.1.2
#4 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Hi Gang,

I didn't see a need to add a sysfs file for the check and repair. This leaves a hard problem for customer to decide. How they decide whether they should repair the bad inode since this may cause corruption even harder?
I think the error should be fixed by this feature automaticlly if repair helps, of course this can be done only when error=continue is enabled or add some mount option for it.

Thanks,
Junxiao.

Quoted 669 lines · Gang He

Implement online file check sysfile interfaces, e.g. how to create the related sysfile according to device name, how to display/handle file check request from the sysfile.

---
 fs/ocfs2/Makefile    |   3 +-
 fs/ocfs2/filecheck.c | 566 +++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/ocfs2/filecheck.h |  48 +++++
 fs/ocfs2/inode.h     |   3 +
 4 files changed, 619 insertions(+), 1 deletion(-)
 create mode 100644 fs/ocfs2/filecheck.c
 create mode 100644 fs/ocfs2/filecheck.h
fs/ocfs2/Makefile +2 -1
index ce210d4..e27e652 100644
@@ -41,7 +41,8 @@ ocfs2-objs := \
 	quota_local.o		\
 	quota_global.o		\
 	xattr.o			\
-	acl.o
+	acl.o	\
+	filecheck.o
 
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
fs/ocfs2/filecheck.c new file +566 -0
new file mode 100644
index 0000000..f12ed1f
@@ -0,0 +1,566 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.c
+ *
+ * Code which implements online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/kmod.h>
+#include <linux/fs.h>
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+#include <linux/sysctl.h>
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+#include "ocfs2_fs.h"
+#include "stackglue.h"
+#include "inode.h"
+
+#include "filecheck.h"
+
+
+/* File check error strings,
+ * must correspond with error number in header file.
+ */
+static const char * const ocfs2_filecheck_errs[] = {
+	"SUCCESS",
+	"FAILED",
+	"INPROGRESS",
+	"READONLY",
+	"INVALIDINO",
+	"BLOCKECC",
+	"BLOCKNO",
+	"VALIDFLAG",
+	"GENERATION",
+	"UNSUPPORTED"
+};
+
+static DEFINE_SPINLOCK(ocfs2_filecheck_sysfs_lock);
+static LIST_HEAD(ocfs2_filecheck_sysfs_list);
+
+struct ocfs2_filecheck {
+	struct list_head fc_head;	/* File check entry list head */
+	spinlock_t fc_lock;
+	unsigned int fc_max;	/* Maximum number of entry in list */
+	unsigned int fc_size;	/* Current entry count in list */
+	unsigned int fc_done;	/* File check entries are done in list */
+};
+
+struct ocfs2_filecheck_sysfs_entry {
+	struct list_head fs_list;
+	atomic_t fs_count;
+	struct super_block *fs_sb;
+	struct kset *fs_kset;
+	struct ocfs2_filecheck *fs_fcheck;
+};
+
+#define OCFS2_FILECHECK_MAXSIZE		100
+#define OCFS2_FILECHECK_MINSIZE		10
+
+/* File check operation type */
+enum {
+	OCFS2_FILECHECK_TYPE_CHK = 0,	/* Check a file */
+	OCFS2_FILECHECK_TYPE_FIX,	/* Fix a file */
+	OCFS2_FILECHECK_TYPE_SET = 100	/* Set file check options */
+};
+
+struct ocfs2_filecheck_entry {
+	struct list_head fe_list;
+	unsigned long fe_ino;
+	unsigned int fe_type;
+	unsigned short fe_done:1;
+	unsigned short fe_status:15;
+};
+
+struct ocfs2_filecheck_args {
+	unsigned int fa_type;
+	union {
+		unsigned long fa_ino;
+		unsigned int fa_len;
+	};
+};
+
+static const char *
+ocfs2_filecheck_error(int errno)
+{
+	if (!errno)
+		return ocfs2_filecheck_errs[errno];
+
+	BUG_ON(errno < OCFS2_FILECHECK_ERR_START ||
+			errno > OCFS2_FILECHECK_ERR_END);
+	return ocfs2_filecheck_errs[errno - OCFS2_FILECHECK_ERR_START + 1];
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf);
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					const char *buf, size_t count);
+static struct kobj_attribute ocfs2_attr_filecheck =
+					__ATTR(filecheck, S_IRUSR | S_IWUSR,
+					ocfs2_filecheck_show,
+					ocfs2_filecheck_store);
+
+static int ocfs2_filecheck_sysfs_wait(atomic_t *p)
+{
+	schedule();
+	return 0;
+}
+
+static void
+ocfs2_filecheck_sysfs_free(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	if (!atomic_dec_and_test(&entry->fs_count))
+		wait_on_atomic_t(&entry->fs_count, ocfs2_filecheck_sysfs_wait,
+						TASK_UNINTERRUPTIBLE);
+
+	spin_lock(&entry->fs_fcheck->fc_lock);
+	while (!list_empty(&entry->fs_fcheck->fc_head)) {
+		p = list_first_entry(&entry->fs_fcheck->fc_head,
+				struct ocfs2_filecheck_entry, fe_list);
+		list_del(&p->fe_list);
+		BUG_ON(!p->fe_done); /* To free a undone file check entry */
+		kfree(p);
+	}
+	spin_unlock(&entry->fs_fcheck->fc_lock);
+
+	kset_unregister(entry->fs_kset);
+	kfree(entry->fs_fcheck);
+	kfree(entry);
+}
+
+static void
+ocfs2_filecheck_sysfs_add(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_add_tail(&entry->fs_list, &ocfs2_filecheck_sysfs_list);
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+}
+
+static int ocfs2_filecheck_sysfs_del(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			list_del(&p->fs_list);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			ocfs2_filecheck_sysfs_free(p);
+			return 0;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return 1;
+}
+
+static void
+ocfs2_filecheck_sysfs_put(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	if (atomic_dec_and_test(&entry->fs_count))
+		wake_up_atomic_t(&entry->fs_count);
+}
+
+static struct ocfs2_filecheck_sysfs_entry *
+ocfs2_filecheck_sysfs_get(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p = NULL;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			atomic_inc(&p->fs_count);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			return p;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return NULL;
+}
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb)
+{
+	int ret = 0;
+	struct kset *ocfs2_filecheck_kset = NULL;
+	struct ocfs2_filecheck *fcheck = NULL;
+	struct ocfs2_filecheck_sysfs_entry *entry = NULL;
+	struct attribute **attrs = NULL;
+	struct attribute_group attrgp;
+
+	if (!ocfs2_kset)
+		return -ENOMEM;
+
+	attrs = kmalloc(sizeof(struct attribute *) * 2, GFP_NOFS);
+	if (!attrs) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		attrs[0] = &ocfs2_attr_filecheck.attr;
+		attrs[1] = NULL;
+		memset(&attrgp, 0, sizeof(attrgp));
+		attrgp.attrs = attrs;
+	}
+
+	fcheck = kmalloc(sizeof(struct ocfs2_filecheck), GFP_NOFS);
+	if (!fcheck) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		INIT_LIST_HEAD(&fcheck->fc_head);
+		spin_lock_init(&fcheck->fc_lock);
+		fcheck->fc_max = OCFS2_FILECHECK_MINSIZE;
+		fcheck->fc_size = 0;
+		fcheck->fc_done = 0;
+	}
+
+	if (strlen(sb->s_id) <= 0) {
+		mlog(ML_ERROR,
+		"Cannot get device basename when create filecheck sysfs\n");
+		ret = -ENODEV;
+		goto error;
+	}
+
+	ocfs2_filecheck_kset = kset_create_and_add(sb->s_id, NULL,
+						&ocfs2_kset->kobj);
+	if (!ocfs2_filecheck_kset) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	ret = sysfs_create_group(&ocfs2_filecheck_kset->kobj, &attrgp);
+	if (ret)
+		goto error;
+
+	entry = kmalloc(sizeof(struct ocfs2_filecheck_sysfs_entry), GFP_NOFS);
+	if (!entry) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		atomic_set(&entry->fs_count, 1);
+		entry->fs_sb = sb;
+		entry->fs_kset = ocfs2_filecheck_kset;
+		entry->fs_fcheck = fcheck;
+		ocfs2_filecheck_sysfs_add(entry);
+	}
+
+	kfree(attrs);
+	return 0;
+
+error:
+	kfree(attrs);
+	kfree(entry);
+	kfree(fcheck);
+	kset_unregister(ocfs2_filecheck_kset);
+	return ret;
+}
+
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb)
+{
+	return ocfs2_filecheck_sysfs_del(sb->s_id);
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count);
+static int
+ocfs2_filecheck_adjust_max(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int len)
+{
+	int ret;
+
+	if ((len < OCFS2_FILECHECK_MINSIZE) || (len > OCFS2_FILECHECK_MAXSIZE))
+		return -EINVAL;
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if (len < (ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done)) {
+		mlog(ML_ERROR,
+		"Cannot set online file check maximum entry number "
+		"to %u due to too much pending entries(%u)\n",
+		len, ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done);
+		ret = -EBUSY;
+	} else {
+		if (len < ent->fs_fcheck->fc_size)
+			BUG_ON(!ocfs2_filecheck_erase_entries(ent,
+				ent->fs_fcheck->fc_size - len));
+
+		ent->fs_fcheck->fc_max = len;
+		ret = 0;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	return ret;
+}
+
+#define OCFS2_FILECHECK_ARGS_LEN	32
+static int
+ocfs2_filecheck_args_get_long(const char *buf, size_t count,
+				unsigned long *val)
+{
+	char buffer[OCFS2_FILECHECK_ARGS_LEN];
+
+	if (count < 1)
+		return 1;
+
+	memcpy(buffer, buf, count);
+	buffer[count] = '\0';
+
+	if (kstrtoul(buffer, 0, val))
+		return 1;
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_args_parse(const char *buf, size_t count,
+				struct ocfs2_filecheck_args *args)
+{
+	unsigned long val = 0;
+
+	/* too short/long args length */
+	if ((count < 5) || (count > OCFS2_FILECHECK_ARGS_LEN))
+		return 1;
+
+	if (!strncasecmp(buf, "FIX ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_FIX;
+		args->fa_ino = val;
+		return 0;
+	} else if ((count > 6) && !strncasecmp(buf, "CHECK ", 6)) {
+		if (ocfs2_filecheck_args_get_long(buf + 6, count - 6, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_CHK;
+		args->fa_ino = val;
+		return 0;
+	} else if (!strncasecmp(buf, "SET ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_SET;
+		args->fa_len = (unsigned int)val;
+		return 0;
+	} else { /* invalid args */
+		return 1;
+	}
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf)
+{
+
+	ssize_t ret = 0, total = 0, remain = PAGE_SIZE;
+	struct ocfs2_filecheck_entry *p;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ret = snprintf(buf, remain, "INO\t\tTYPE\tDONE\tERROR\n");
+	total += ret;
+	remain -= ret;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		ret = snprintf(buf + total, remain, "%lu\t\t%u\t%u\t%s\n",
+			p->fe_ino, p->fe_type, p->fe_done,
+			ocfs2_filecheck_error(p->fe_status));
+		if (ret < 0) {
+			total = ret;
+			break;
+		}
+		if (ret == remain) {
+			/* snprintf() didn't fit */
+			total = -E2BIG;
+			break;
+		}
+		total += ret;
+		remain -= ret;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return total;
+}
+
+static int
+ocfs2_filecheck_erase_entry(struct ocfs2_filecheck_sysfs_entry *ent)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		if (p->fe_done) {
+			list_del(&p->fe_list);
+			kfree(p);
+			ent->fs_fcheck->fc_size--;
+			ent->fs_fcheck->fc_done--;
+			return 1;
+		}
+	}
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count)
+{
+	unsigned int i = 0;
+	unsigned int ret = 0;
+
+	while (i++ < count) {
+		if (ocfs2_filecheck_erase_entry(ent))
+			ret++;
+		else
+			break;
+	}
+
+	return (ret == count ? 1 : 0);
+}
+
+static void
+ocfs2_filecheck_done_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	entry->fe_done = 1;
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ent->fs_fcheck->fc_done++;
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+}
+
+static unsigned short
+ocfs2_filecheck_handle(struct super_block *sb,
+				unsigned long ino, unsigned int flags)
+{
+	unsigned short ret = OCFS2_FILECHECK_ERR_SUCCESS;
+	struct inode *inode = NULL;
+	int rc;
+
+	inode = ocfs2_iget(OCFS2_SB(sb), ino, flags, 0);
+	if (IS_ERR(inode)) {
+		rc = (int)(-(long)inode);
+		if (rc >= OCFS2_FILECHECK_ERR_START &&
+			rc < OCFS2_FILECHECK_ERR_END)
+			ret = rc;
+		else
+			ret = OCFS2_FILECHECK_ERR_FAILED;
+	} else
+		iput(inode);
+
+	return ret;
+}
+
+static void
+ocfs2_filecheck_handle_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	if (entry->fe_type == OCFS2_FILECHECK_TYPE_CHK)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_CHK);
+	else if (entry->fe_type == OCFS2_FILECHECK_TYPE_FIX)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_FIX);
+	else
+		entry->fe_status = OCFS2_FILECHECK_ERR_UNSUPPORTED;
+
+	ocfs2_filecheck_done_entry(ent, entry);
+}
+
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+				struct kobj_attribute *attr,
+				const char *buf, size_t count)
+{
+	struct ocfs2_filecheck_args args;
+	struct ocfs2_filecheck_entry *entry = NULL;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+	ssize_t ret = 0;
+
+	if (count == 0)
+		return count;
+
+	if (ocfs2_filecheck_args_parse(buf, count, &args)) {
+		mlog(ML_ERROR, "Invalid arguments for online file check\n");
+		return -EINVAL;
+	}
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	if (args.fa_type == OCFS2_FILECHECK_TYPE_SET) {
+		ret = ocfs2_filecheck_adjust_max(ent, args.fa_len);
+		ocfs2_filecheck_sysfs_put(ent);
+		return (!ret ? count : ret);
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+		(ent->fs_fcheck->fc_done == 0)) {
+		mlog(ML_ERROR,
+		"Online file check queue(%u) is full\n",
+		ent->fs_fcheck->fc_max);
+		ret = -EBUSY;
+	} else {
+		if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+			(ent->fs_fcheck->fc_done > 0)) {
+			/* Delete the oldest entry which was done,
+			 * make sure the entry size in list does
+			 * not exceed maximum value
+			 */
+			BUG_ON(!ocfs2_filecheck_erase_entry(ent));
+		}
+
+		entry = kmalloc(sizeof(struct ocfs2_filecheck_entry), GFP_NOFS);
+		if (entry) {
+			entry->fe_ino = args.fa_ino;
+			entry->fe_type = args.fa_type;
+			entry->fe_done = 0;
+			entry->fe_status = OCFS2_FILECHECK_ERR_INPROGRESS;
+			list_add_tail(&entry->fe_list,
+					&ent->fs_fcheck->fc_head);
+
+			ent->fs_fcheck->fc_size++;
+			ret = count;
+		} else {
+			ret = -ENOMEM;
+		}
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	if (entry)
+		ocfs2_filecheck_handle_entry(ent, entry);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return ret;
+}
fs/ocfs2/filecheck.h new file +48 -0
new file mode 100644
index 0000000..5ec331b
@@ -0,0 +1,48 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.h
+ *
+ * Online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+
+#ifndef FILECHECK_H
+#define FILECHECK_H
+
+#include <linux/types.h>
+#include <linux/list.h>
+
+
+/* File check errno */
+enum {
+	OCFS2_FILECHECK_ERR_SUCCESS = 0,	/* Success */
+	OCFS2_FILECHECK_ERR_FAILED = 1000,	/* Other failure */
+	OCFS2_FILECHECK_ERR_INPROGRESS,		/* In progress */
+	OCFS2_FILECHECK_ERR_READONLY,		/* Read only */
+	OCFS2_FILECHECK_ERR_INVALIDINO,		/* Invalid ino */
+	OCFS2_FILECHECK_ERR_BLOCKECC,		/* Block ecc */
+	OCFS2_FILECHECK_ERR_BLOCKNO,		/* Block number */
+	OCFS2_FILECHECK_ERR_VALIDFLAG,		/* Inode valid flag */
+	OCFS2_FILECHECK_ERR_GENERATION,		/* Inode generation */
+	OCFS2_FILECHECK_ERR_UNSUPPORTED		/* Unsupported */
+};
+
+#define OCFS2_FILECHECK_ERR_START	OCFS2_FILECHECK_ERR_FAILED
+#define OCFS2_FILECHECK_ERR_END		OCFS2_FILECHECK_ERR_UNSUPPORTED
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb);
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb);
+
+#endif  /* FILECHECK_H */
fs/ocfs2/inode.h +3 -0
index 5e86b24..abd1018 100644
@@ -139,6 +139,9 @@ int ocfs2_drop_inode(struct inode *inode);
 /* Flags for ocfs2_iget() */
 #define OCFS2_FI_FLAG_SYSFILE		0x1
 #define OCFS2_FI_FLAG_ORPHAN_RECOVERY	0x2
+#define OCFS2_FI_FLAG_FILECHECK_CHK	0x4
+#define OCFS2_FI_FLAG_FILECHECK_FIX	0x8
+
 struct inode *ocfs2_ilookup(struct super_block *sb, u64 feoff);
 struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned flags,
 			 int sysfile_type);
author
#5 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via pass a ino number).
But, this operation is manually triggered by user, instead of automatically fix in the kernel.
Why?
1) we should let users make this decision, since some users do not want to fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.
2) frankly speaking, this feature will probably bring a second corruption if there is some error in the code, I do not suggest to use automatically fix by default in the first version.
3) in the future, if this feature is well proved, we can add a mount option to make this automatically fix enabled.

Thanks
Gang

Quoted 686 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

I didn't see a need to add a sysfs file for the check and repair. This leaves a hard problem for customer to decide. How they decide whether they should repair the bad inode since this may cause corruption even harder?
I think the error should be fixed by this feature automaticlly if repair helps, of course this can be done only when error=continue is enabled or add some mount option for it.

Thanks,
Junxiao.

Quoted 8 lines · Gang He

Implement online file check sysfile interfaces, e.g. how to create the related sysfile according to device name, how to display/handle file check request from the sysfile.

---
 fs/ocfs2/Makefile    |   3 +-
 fs/ocfs2/filecheck.c | 566 

+++++++++++++++++++++++++++++++++++++++++++++++++++

Quoted 651 lines
 fs/ocfs2/filecheck.h |  48 +++++
 fs/ocfs2/inode.h     |   3 +
 4 files changed, 619 insertions(+), 1 deletion(-)
 create mode 100644 fs/ocfs2/filecheck.c
 create mode 100644 fs/ocfs2/filecheck.h
fs/ocfs2/Makefile +2 -1
index ce210d4..e27e652 100644
@@ -41,7 +41,8 @@ ocfs2-objs := \
 	quota_local.o		\
 	quota_global.o		\
 	xattr.o			\
-	acl.o
+	acl.o	\
+	filecheck.o
 
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
fs/ocfs2/filecheck.c new file +566 -0
new file mode 100644
index 0000000..f12ed1f
@@ -0,0 +1,566 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.c
+ *
+ * Code which implements online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/kmod.h>
+#include <linux/fs.h>
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+#include <linux/sysctl.h>
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+#include "ocfs2_fs.h"
+#include "stackglue.h"
+#include "inode.h"
+
+#include "filecheck.h"
+
+
+/* File check error strings,
+ * must correspond with error number in header file.
+ */
+static const char * const ocfs2_filecheck_errs[] = {
+	"SUCCESS",
+	"FAILED",
+	"INPROGRESS",
+	"READONLY",
+	"INVALIDINO",
+	"BLOCKECC",
+	"BLOCKNO",
+	"VALIDFLAG",
+	"GENERATION",
+	"UNSUPPORTED"
+};
+
+static DEFINE_SPINLOCK(ocfs2_filecheck_sysfs_lock);
+static LIST_HEAD(ocfs2_filecheck_sysfs_list);
+
+struct ocfs2_filecheck {
+	struct list_head fc_head;	/* File check entry list head */
+	spinlock_t fc_lock;
+	unsigned int fc_max;	/* Maximum number of entry in list */
+	unsigned int fc_size;	/* Current entry count in list */
+	unsigned int fc_done;	/* File check entries are done in list */
+};
+
+struct ocfs2_filecheck_sysfs_entry {
+	struct list_head fs_list;
+	atomic_t fs_count;
+	struct super_block *fs_sb;
+	struct kset *fs_kset;
+	struct ocfs2_filecheck *fs_fcheck;
+};
+
+#define OCFS2_FILECHECK_MAXSIZE		100
+#define OCFS2_FILECHECK_MINSIZE		10
+
+/* File check operation type */
+enum {
+	OCFS2_FILECHECK_TYPE_CHK = 0,	/* Check a file */
+	OCFS2_FILECHECK_TYPE_FIX,	/* Fix a file */
+	OCFS2_FILECHECK_TYPE_SET = 100	/* Set file check options */
+};
+
+struct ocfs2_filecheck_entry {
+	struct list_head fe_list;
+	unsigned long fe_ino;
+	unsigned int fe_type;
+	unsigned short fe_done:1;
+	unsigned short fe_status:15;
+};
+
+struct ocfs2_filecheck_args {
+	unsigned int fa_type;
+	union {
+		unsigned long fa_ino;
+		unsigned int fa_len;
+	};
+};
+
+static const char *
+ocfs2_filecheck_error(int errno)
+{
+	if (!errno)
+		return ocfs2_filecheck_errs[errno];
+
+	BUG_ON(errno < OCFS2_FILECHECK_ERR_START ||
+			errno > OCFS2_FILECHECK_ERR_END);
+	return ocfs2_filecheck_errs[errno - OCFS2_FILECHECK_ERR_START + 1];
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf);
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					const char *buf, size_t count);
+static struct kobj_attribute ocfs2_attr_filecheck =
+					__ATTR(filecheck, S_IRUSR | S_IWUSR,
+					ocfs2_filecheck_show,
+					ocfs2_filecheck_store);
+
+static int ocfs2_filecheck_sysfs_wait(atomic_t *p)
+{
+	schedule();
+	return 0;
+}
+
+static void
+ocfs2_filecheck_sysfs_free(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	if (!atomic_dec_and_test(&entry->fs_count))
+		wait_on_atomic_t(&entry->fs_count, ocfs2_filecheck_sysfs_wait,
+						TASK_UNINTERRUPTIBLE);
+
+	spin_lock(&entry->fs_fcheck->fc_lock);
+	while (!list_empty(&entry->fs_fcheck->fc_head)) {
+		p = list_first_entry(&entry->fs_fcheck->fc_head,
+				struct ocfs2_filecheck_entry, fe_list);
+		list_del(&p->fe_list);
+		BUG_ON(!p->fe_done); /* To free a undone file check entry */
+		kfree(p);
+	}
+	spin_unlock(&entry->fs_fcheck->fc_lock);
+
+	kset_unregister(entry->fs_kset);
+	kfree(entry->fs_fcheck);
+	kfree(entry);
+}
+
+static void
+ocfs2_filecheck_sysfs_add(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_add_tail(&entry->fs_list, &ocfs2_filecheck_sysfs_list);
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+}
+
+static int ocfs2_filecheck_sysfs_del(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			list_del(&p->fs_list);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			ocfs2_filecheck_sysfs_free(p);
+			return 0;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return 1;
+}
+
+static void
+ocfs2_filecheck_sysfs_put(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	if (atomic_dec_and_test(&entry->fs_count))
+		wake_up_atomic_t(&entry->fs_count);
+}
+
+static struct ocfs2_filecheck_sysfs_entry *
+ocfs2_filecheck_sysfs_get(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p = NULL;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			atomic_inc(&p->fs_count);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			return p;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return NULL;
+}
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb)
+{
+	int ret = 0;
+	struct kset *ocfs2_filecheck_kset = NULL;
+	struct ocfs2_filecheck *fcheck = NULL;
+	struct ocfs2_filecheck_sysfs_entry *entry = NULL;
+	struct attribute **attrs = NULL;
+	struct attribute_group attrgp;
+
+	if (!ocfs2_kset)
+		return -ENOMEM;
+
+	attrs = kmalloc(sizeof(struct attribute *) * 2, GFP_NOFS);
+	if (!attrs) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		attrs[0] = &ocfs2_attr_filecheck.attr;
+		attrs[1] = NULL;
+		memset(&attrgp, 0, sizeof(attrgp));
+		attrgp.attrs = attrs;
+	}
+
+	fcheck = kmalloc(sizeof(struct ocfs2_filecheck), GFP_NOFS);
+	if (!fcheck) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		INIT_LIST_HEAD(&fcheck->fc_head);
+		spin_lock_init(&fcheck->fc_lock);
+		fcheck->fc_max = OCFS2_FILECHECK_MINSIZE;
+		fcheck->fc_size = 0;
+		fcheck->fc_done = 0;
+	}
+
+	if (strlen(sb->s_id) <= 0) {
+		mlog(ML_ERROR,
+		"Cannot get device basename when create filecheck sysfs\n");
+		ret = -ENODEV;
+		goto error;
+	}
+
+	ocfs2_filecheck_kset = kset_create_and_add(sb->s_id, NULL,
+						&ocfs2_kset->kobj);
+	if (!ocfs2_filecheck_kset) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	ret = sysfs_create_group(&ocfs2_filecheck_kset->kobj, &attrgp);
+	if (ret)
+		goto error;
+
+	entry = kmalloc(sizeof(struct ocfs2_filecheck_sysfs_entry), GFP_NOFS);
+	if (!entry) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		atomic_set(&entry->fs_count, 1);
+		entry->fs_sb = sb;
+		entry->fs_kset = ocfs2_filecheck_kset;
+		entry->fs_fcheck = fcheck;
+		ocfs2_filecheck_sysfs_add(entry);
+	}
+
+	kfree(attrs);
+	return 0;
+
+error:
+	kfree(attrs);
+	kfree(entry);
+	kfree(fcheck);
+	kset_unregister(ocfs2_filecheck_kset);
+	return ret;
+}
+
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb)
+{
+	return ocfs2_filecheck_sysfs_del(sb->s_id);
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count);
+static int
+ocfs2_filecheck_adjust_max(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int len)
+{
+	int ret;
+
+	if ((len < OCFS2_FILECHECK_MINSIZE) || (len > OCFS2_FILECHECK_MAXSIZE))
+		return -EINVAL;
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if (len < (ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done)) {
+		mlog(ML_ERROR,
+		"Cannot set online file check maximum entry number "
+		"to %u due to too much pending entries(%u)\n",
+		len, ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done);
+		ret = -EBUSY;
+	} else {
+		if (len < ent->fs_fcheck->fc_size)
+			BUG_ON(!ocfs2_filecheck_erase_entries(ent,
+				ent->fs_fcheck->fc_size - len));
+
+		ent->fs_fcheck->fc_max = len;
+		ret = 0;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	return ret;
+}
+
+#define OCFS2_FILECHECK_ARGS_LEN	32
+static int
+ocfs2_filecheck_args_get_long(const char *buf, size_t count,
+				unsigned long *val)
+{
+	char buffer[OCFS2_FILECHECK_ARGS_LEN];
+
+	if (count < 1)
+		return 1;
+
+	memcpy(buffer, buf, count);
+	buffer[count] = '\0';
+
+	if (kstrtoul(buffer, 0, val))
+		return 1;
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_args_parse(const char *buf, size_t count,
+				struct ocfs2_filecheck_args *args)
+{
+	unsigned long val = 0;
+
+	/* too short/long args length */
+	if ((count < 5) || (count > OCFS2_FILECHECK_ARGS_LEN))
+		return 1;
+
+	if (!strncasecmp(buf, "FIX ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_FIX;
+		args->fa_ino = val;
+		return 0;
+	} else if ((count > 6) && !strncasecmp(buf, "CHECK ", 6)) {
+		if (ocfs2_filecheck_args_get_long(buf + 6, count - 6, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_CHK;
+		args->fa_ino = val;
+		return 0;
+	} else if (!strncasecmp(buf, "SET ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_SET;
+		args->fa_len = (unsigned int)val;
+		return 0;
+	} else { /* invalid args */
+		return 1;
+	}
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf)
+{
+
+	ssize_t ret = 0, total = 0, remain = PAGE_SIZE;
+	struct ocfs2_filecheck_entry *p;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ret = snprintf(buf, remain, "INO\t\tTYPE\tDONE\tERROR\n");
+	total += ret;
+	remain -= ret;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		ret = snprintf(buf + total, remain, "%lu\t\t%u\t%u\t%s\n",
+			p->fe_ino, p->fe_type, p->fe_done,
+			ocfs2_filecheck_error(p->fe_status));
+		if (ret < 0) {
+			total = ret;
+			break;
+		}
+		if (ret == remain) {
+			/* snprintf() didn't fit */
+			total = -E2BIG;
+			break;
+		}
+		total += ret;
+		remain -= ret;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return total;
+}
+
+static int
+ocfs2_filecheck_erase_entry(struct ocfs2_filecheck_sysfs_entry *ent)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		if (p->fe_done) {
+			list_del(&p->fe_list);
+			kfree(p);
+			ent->fs_fcheck->fc_size--;
+			ent->fs_fcheck->fc_done--;
+			return 1;
+		}
+	}
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count)
+{
+	unsigned int i = 0;
+	unsigned int ret = 0;
+
+	while (i++ < count) {
+		if (ocfs2_filecheck_erase_entry(ent))
+			ret++;
+		else
+			break;
+	}
+
+	return (ret == count ? 1 : 0);
+}
+
+static void
+ocfs2_filecheck_done_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	entry->fe_done = 1;
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ent->fs_fcheck->fc_done++;
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+}
+
+static unsigned short
+ocfs2_filecheck_handle(struct super_block *sb,
+				unsigned long ino, unsigned int flags)
+{
+	unsigned short ret = OCFS2_FILECHECK_ERR_SUCCESS;
+	struct inode *inode = NULL;
+	int rc;
+
+	inode = ocfs2_iget(OCFS2_SB(sb), ino, flags, 0);
+	if (IS_ERR(inode)) {
+		rc = (int)(-(long)inode);
+		if (rc >= OCFS2_FILECHECK_ERR_START &&
+			rc < OCFS2_FILECHECK_ERR_END)
+			ret = rc;
+		else
+			ret = OCFS2_FILECHECK_ERR_FAILED;
+	} else
+		iput(inode);
+
+	return ret;
+}
+
+static void
+ocfs2_filecheck_handle_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	if (entry->fe_type == OCFS2_FILECHECK_TYPE_CHK)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_CHK);
+	else if (entry->fe_type == OCFS2_FILECHECK_TYPE_FIX)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_FIX);
+	else
+		entry->fe_status = OCFS2_FILECHECK_ERR_UNSUPPORTED;
+
+	ocfs2_filecheck_done_entry(ent, entry);
+}
+
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+				struct kobj_attribute *attr,
+				const char *buf, size_t count)
+{
+	struct ocfs2_filecheck_args args;
+	struct ocfs2_filecheck_entry *entry = NULL;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+	ssize_t ret = 0;
+
+	if (count == 0)
+		return count;
+
+	if (ocfs2_filecheck_args_parse(buf, count, &args)) {
+		mlog(ML_ERROR, "Invalid arguments for online file check\n");
+		return -EINVAL;
+	}
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	if (args.fa_type == OCFS2_FILECHECK_TYPE_SET) {
+		ret = ocfs2_filecheck_adjust_max(ent, args.fa_len);
+		ocfs2_filecheck_sysfs_put(ent);
+		return (!ret ? count : ret);
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+		(ent->fs_fcheck->fc_done == 0)) {
+		mlog(ML_ERROR,
+		"Online file check queue(%u) is full\n",
+		ent->fs_fcheck->fc_max);
+		ret = -EBUSY;
+	} else {
+		if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+			(ent->fs_fcheck->fc_done > 0)) {
+			/* Delete the oldest entry which was done,
+			 * make sure the entry size in list does
+			 * not exceed maximum value
+			 */
+			BUG_ON(!ocfs2_filecheck_erase_entry(ent));
+		}
+
+		entry = kmalloc(sizeof(struct ocfs2_filecheck_entry), GFP_NOFS);
+		if (entry) {
+			entry->fe_ino = args.fa_ino;
+			entry->fe_type = args.fa_type;
+			entry->fe_done = 0;
+			entry->fe_status = OCFS2_FILECHECK_ERR_INPROGRESS;
+			list_add_tail(&entry->fe_list,
+					&ent->fs_fcheck->fc_head);
+
+			ent->fs_fcheck->fc_size++;
+			ret = count;
+		} else {
+			ret = -ENOMEM;
+		}
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	if (entry)
+		ocfs2_filecheck_handle_entry(ent, entry);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return ret;
+}
fs/ocfs2/filecheck.h new file +48 -0
new file mode 100644
index 0000000..5ec331b
@@ -0,0 +1,48 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.h
+ *
+ * Online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+
+#ifndef FILECHECK_H
+#define FILECHECK_H
+
+#include <linux/types.h>
+#include <linux/list.h>
+
+
+/* File check errno */
+enum {
+	OCFS2_FILECHECK_ERR_SUCCESS = 0,	/* Success */
+	OCFS2_FILECHECK_ERR_FAILED = 1000,	/* Other failure */
+	OCFS2_FILECHECK_ERR_INPROGRESS,		/* In progress */
+	OCFS2_FILECHECK_ERR_READONLY,		/* Read only */
+	OCFS2_FILECHECK_ERR_INVALIDINO,		/* Invalid ino */
+	OCFS2_FILECHECK_ERR_BLOCKECC,		/* Block ecc */
+	OCFS2_FILECHECK_ERR_BLOCKNO,		/* Block number */
+	OCFS2_FILECHECK_ERR_VALIDFLAG,		/* Inode valid flag */
+	OCFS2_FILECHECK_ERR_GENERATION,		/* Inode generation */
+	OCFS2_FILECHECK_ERR_UNSUPPORTED		/* Unsupported */
+};
+
+#define OCFS2_FILECHECK_ERR_START	OCFS2_FILECHECK_ERR_FAILED
+#define OCFS2_FILECHECK_ERR_END		OCFS2_FILECHECK_ERR_UNSUPPORTED
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb);
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb);
+
+#endif  /* FILECHECK_H */
fs/ocfs2/inode.h +0 -0
index 5e86b24..abd1018 100644

fs/ocfs2/inode.h int ocfs2_drop_inode(struct inode *inode); · line 146

Quoted 8 lines
fs/ocfs2/inode.h +3 -0
@@ -139,6 +139,9 @@ int ocfs2_drop_inode(struct inode *inode);
 /* Flags for ocfs2_iget() */
 #define OCFS2_FI_FLAG_SYSFILE		0x1
 #define OCFS2_FI_FLAG_ORPHAN_RECOVERY	0x2
+#define OCFS2_FI_FLAG_FILECHECK_CHK	0x4
+#define OCFS2_FI_FLAG_FILECHECK_FIX	0x8
+
 struct inode *ocfs2_ilookup(struct super_block *sb, u64 feoff);
 struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned 

flags,

Quoted 2 lines
fs/ocfs2/inode.h +0 -0
 			 int sysfile_type);
#6 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Hi Gang,

Quoted 7 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via pass a ino number).
But, this operation is manually triggered by user, instead of automatically fix in the kernel.
Why?
1) we should let users make this decision, since some users do not want to fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Quoted 1 line

2) frankly speaking, this feature will probably bring a second corruption if there is some error in the code, I do not suggest to use automatically fix by default in the first version.

I think if this feature could bring more corruption, then this should be fixed first.

Thanks,
Junxiao

Quoted 694 lines

3) in the future, if this feature is well proved, we can add a mount option to make this automatically fix enabled.

Thanks
Gang

Quoted 686 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

I didn't see a need to add a sysfs file for the check and repair. This leaves a hard problem for customer to decide. How they decide whether they should repair the bad inode since this may cause corruption even harder?
I think the error should be fixed by this feature automaticlly if repair helps, of course this can be done only when error=continue is enabled or add some mount option for it.

Thanks,
Junxiao.

Quoted 8 lines · Gang He

Implement online file check sysfile interfaces, e.g. how to create the related sysfile according to device name, how to display/handle file check request from the sysfile.

---
 fs/ocfs2/Makefile    |   3 +-
 fs/ocfs2/filecheck.c | 566 

+++++++++++++++++++++++++++++++++++++++++++++++++++

Quoted 651 lines
 fs/ocfs2/filecheck.h |  48 +++++
 fs/ocfs2/inode.h     |   3 +
 4 files changed, 619 insertions(+), 1 deletion(-)
 create mode 100644 fs/ocfs2/filecheck.c
 create mode 100644 fs/ocfs2/filecheck.h
fs/ocfs2/Makefile +2 -1
index ce210d4..e27e652 100644
@@ -41,7 +41,8 @@ ocfs2-objs := \
 	quota_local.o		\
 	quota_global.o		\
 	xattr.o			\
-	acl.o
+	acl.o	\
+	filecheck.o
 
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
fs/ocfs2/filecheck.c new file +566 -0
new file mode 100644
index 0000000..f12ed1f
@@ -0,0 +1,566 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.c
+ *
+ * Code which implements online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/kmod.h>
+#include <linux/fs.h>
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+#include <linux/sysctl.h>
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+#include "ocfs2_fs.h"
+#include "stackglue.h"
+#include "inode.h"
+
+#include "filecheck.h"
+
+
+/* File check error strings,
+ * must correspond with error number in header file.
+ */
+static const char * const ocfs2_filecheck_errs[] = {
+	"SUCCESS",
+	"FAILED",
+	"INPROGRESS",
+	"READONLY",
+	"INVALIDINO",
+	"BLOCKECC",
+	"BLOCKNO",
+	"VALIDFLAG",
+	"GENERATION",
+	"UNSUPPORTED"
+};
+
+static DEFINE_SPINLOCK(ocfs2_filecheck_sysfs_lock);
+static LIST_HEAD(ocfs2_filecheck_sysfs_list);
+
+struct ocfs2_filecheck {
+	struct list_head fc_head;	/* File check entry list head */
+	spinlock_t fc_lock;
+	unsigned int fc_max;	/* Maximum number of entry in list */
+	unsigned int fc_size;	/* Current entry count in list */
+	unsigned int fc_done;	/* File check entries are done in list */
+};
+
+struct ocfs2_filecheck_sysfs_entry {
+	struct list_head fs_list;
+	atomic_t fs_count;
+	struct super_block *fs_sb;
+	struct kset *fs_kset;
+	struct ocfs2_filecheck *fs_fcheck;
+};
+
+#define OCFS2_FILECHECK_MAXSIZE		100
+#define OCFS2_FILECHECK_MINSIZE		10
+
+/* File check operation type */
+enum {
+	OCFS2_FILECHECK_TYPE_CHK = 0,	/* Check a file */
+	OCFS2_FILECHECK_TYPE_FIX,	/* Fix a file */
+	OCFS2_FILECHECK_TYPE_SET = 100	/* Set file check options */
+};
+
+struct ocfs2_filecheck_entry {
+	struct list_head fe_list;
+	unsigned long fe_ino;
+	unsigned int fe_type;
+	unsigned short fe_done:1;
+	unsigned short fe_status:15;
+};
+
+struct ocfs2_filecheck_args {
+	unsigned int fa_type;
+	union {
+		unsigned long fa_ino;
+		unsigned int fa_len;
+	};
+};
+
+static const char *
+ocfs2_filecheck_error(int errno)
+{
+	if (!errno)
+		return ocfs2_filecheck_errs[errno];
+
+	BUG_ON(errno < OCFS2_FILECHECK_ERR_START ||
+			errno > OCFS2_FILECHECK_ERR_END);
+	return ocfs2_filecheck_errs[errno - OCFS2_FILECHECK_ERR_START + 1];
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf);
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					const char *buf, size_t count);
+static struct kobj_attribute ocfs2_attr_filecheck =
+					__ATTR(filecheck, S_IRUSR | S_IWUSR,
+					ocfs2_filecheck_show,
+					ocfs2_filecheck_store);
+
+static int ocfs2_filecheck_sysfs_wait(atomic_t *p)
+{
+	schedule();
+	return 0;
+}
+
+static void
+ocfs2_filecheck_sysfs_free(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	if (!atomic_dec_and_test(&entry->fs_count))
+		wait_on_atomic_t(&entry->fs_count, ocfs2_filecheck_sysfs_wait,
+						TASK_UNINTERRUPTIBLE);
+
+	spin_lock(&entry->fs_fcheck->fc_lock);
+	while (!list_empty(&entry->fs_fcheck->fc_head)) {
+		p = list_first_entry(&entry->fs_fcheck->fc_head,
+				struct ocfs2_filecheck_entry, fe_list);
+		list_del(&p->fe_list);
+		BUG_ON(!p->fe_done); /* To free a undone file check entry */
+		kfree(p);
+	}
+	spin_unlock(&entry->fs_fcheck->fc_lock);
+
+	kset_unregister(entry->fs_kset);
+	kfree(entry->fs_fcheck);
+	kfree(entry);
+}
+
+static void
+ocfs2_filecheck_sysfs_add(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_add_tail(&entry->fs_list, &ocfs2_filecheck_sysfs_list);
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+}
+
+static int ocfs2_filecheck_sysfs_del(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			list_del(&p->fs_list);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			ocfs2_filecheck_sysfs_free(p);
+			return 0;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return 1;
+}
+
+static void
+ocfs2_filecheck_sysfs_put(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	if (atomic_dec_and_test(&entry->fs_count))
+		wake_up_atomic_t(&entry->fs_count);
+}
+
+static struct ocfs2_filecheck_sysfs_entry *
+ocfs2_filecheck_sysfs_get(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p = NULL;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			atomic_inc(&p->fs_count);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			return p;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return NULL;
+}
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb)
+{
+	int ret = 0;
+	struct kset *ocfs2_filecheck_kset = NULL;
+	struct ocfs2_filecheck *fcheck = NULL;
+	struct ocfs2_filecheck_sysfs_entry *entry = NULL;
+	struct attribute **attrs = NULL;
+	struct attribute_group attrgp;
+
+	if (!ocfs2_kset)
+		return -ENOMEM;
+
+	attrs = kmalloc(sizeof(struct attribute *) * 2, GFP_NOFS);
+	if (!attrs) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		attrs[0] = &ocfs2_attr_filecheck.attr;
+		attrs[1] = NULL;
+		memset(&attrgp, 0, sizeof(attrgp));
+		attrgp.attrs = attrs;
+	}
+
+	fcheck = kmalloc(sizeof(struct ocfs2_filecheck), GFP_NOFS);
+	if (!fcheck) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		INIT_LIST_HEAD(&fcheck->fc_head);
+		spin_lock_init(&fcheck->fc_lock);
+		fcheck->fc_max = OCFS2_FILECHECK_MINSIZE;
+		fcheck->fc_size = 0;
+		fcheck->fc_done = 0;
+	}
+
+	if (strlen(sb->s_id) <= 0) {
+		mlog(ML_ERROR,
+		"Cannot get device basename when create filecheck sysfs\n");
+		ret = -ENODEV;
+		goto error;
+	}
+
+	ocfs2_filecheck_kset = kset_create_and_add(sb->s_id, NULL,
+						&ocfs2_kset->kobj);
+	if (!ocfs2_filecheck_kset) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	ret = sysfs_create_group(&ocfs2_filecheck_kset->kobj, &attrgp);
+	if (ret)
+		goto error;
+
+	entry = kmalloc(sizeof(struct ocfs2_filecheck_sysfs_entry), GFP_NOFS);
+	if (!entry) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		atomic_set(&entry->fs_count, 1);
+		entry->fs_sb = sb;
+		entry->fs_kset = ocfs2_filecheck_kset;
+		entry->fs_fcheck = fcheck;
+		ocfs2_filecheck_sysfs_add(entry);
+	}
+
+	kfree(attrs);
+	return 0;
+
+error:
+	kfree(attrs);
+	kfree(entry);
+	kfree(fcheck);
+	kset_unregister(ocfs2_filecheck_kset);
+	return ret;
+}
+
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb)
+{
+	return ocfs2_filecheck_sysfs_del(sb->s_id);
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count);
+static int
+ocfs2_filecheck_adjust_max(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int len)
+{
+	int ret;
+
+	if ((len < OCFS2_FILECHECK_MINSIZE) || (len > OCFS2_FILECHECK_MAXSIZE))
+		return -EINVAL;
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if (len < (ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done)) {
+		mlog(ML_ERROR,
+		"Cannot set online file check maximum entry number "
+		"to %u due to too much pending entries(%u)\n",
+		len, ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done);
+		ret = -EBUSY;
+	} else {
+		if (len < ent->fs_fcheck->fc_size)
+			BUG_ON(!ocfs2_filecheck_erase_entries(ent,
+				ent->fs_fcheck->fc_size - len));
+
+		ent->fs_fcheck->fc_max = len;
+		ret = 0;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	return ret;
+}
+
+#define OCFS2_FILECHECK_ARGS_LEN	32
+static int
+ocfs2_filecheck_args_get_long(const char *buf, size_t count,
+				unsigned long *val)
+{
+	char buffer[OCFS2_FILECHECK_ARGS_LEN];
+
+	if (count < 1)
+		return 1;
+
+	memcpy(buffer, buf, count);
+	buffer[count] = '\0';
+
+	if (kstrtoul(buffer, 0, val))
+		return 1;
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_args_parse(const char *buf, size_t count,
+				struct ocfs2_filecheck_args *args)
+{
+	unsigned long val = 0;
+
+	/* too short/long args length */
+	if ((count < 5) || (count > OCFS2_FILECHECK_ARGS_LEN))
+		return 1;
+
+	if (!strncasecmp(buf, "FIX ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_FIX;
+		args->fa_ino = val;
+		return 0;
+	} else if ((count > 6) && !strncasecmp(buf, "CHECK ", 6)) {
+		if (ocfs2_filecheck_args_get_long(buf + 6, count - 6, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_CHK;
+		args->fa_ino = val;
+		return 0;
+	} else if (!strncasecmp(buf, "SET ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_SET;
+		args->fa_len = (unsigned int)val;
+		return 0;
+	} else { /* invalid args */
+		return 1;
+	}
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf)
+{
+
+	ssize_t ret = 0, total = 0, remain = PAGE_SIZE;
+	struct ocfs2_filecheck_entry *p;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ret = snprintf(buf, remain, "INO\t\tTYPE\tDONE\tERROR\n");
+	total += ret;
+	remain -= ret;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		ret = snprintf(buf + total, remain, "%lu\t\t%u\t%u\t%s\n",
+			p->fe_ino, p->fe_type, p->fe_done,
+			ocfs2_filecheck_error(p->fe_status));
+		if (ret < 0) {
+			total = ret;
+			break;
+		}
+		if (ret == remain) {
+			/* snprintf() didn't fit */
+			total = -E2BIG;
+			break;
+		}
+		total += ret;
+		remain -= ret;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return total;
+}
+
+static int
+ocfs2_filecheck_erase_entry(struct ocfs2_filecheck_sysfs_entry *ent)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		if (p->fe_done) {
+			list_del(&p->fe_list);
+			kfree(p);
+			ent->fs_fcheck->fc_size--;
+			ent->fs_fcheck->fc_done--;
+			return 1;
+		}
+	}
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count)
+{
+	unsigned int i = 0;
+	unsigned int ret = 0;
+
+	while (i++ < count) {
+		if (ocfs2_filecheck_erase_entry(ent))
+			ret++;
+		else
+			break;
+	}
+
+	return (ret == count ? 1 : 0);
+}
+
+static void
+ocfs2_filecheck_done_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	entry->fe_done = 1;
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ent->fs_fcheck->fc_done++;
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+}
+
+static unsigned short
+ocfs2_filecheck_handle(struct super_block *sb,
+				unsigned long ino, unsigned int flags)
+{
+	unsigned short ret = OCFS2_FILECHECK_ERR_SUCCESS;
+	struct inode *inode = NULL;
+	int rc;
+
+	inode = ocfs2_iget(OCFS2_SB(sb), ino, flags, 0);
+	if (IS_ERR(inode)) {
+		rc = (int)(-(long)inode);
+		if (rc >= OCFS2_FILECHECK_ERR_START &&
+			rc < OCFS2_FILECHECK_ERR_END)
+			ret = rc;
+		else
+			ret = OCFS2_FILECHECK_ERR_FAILED;
+	} else
+		iput(inode);
+
+	return ret;
+}
+
+static void
+ocfs2_filecheck_handle_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	if (entry->fe_type == OCFS2_FILECHECK_TYPE_CHK)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_CHK);
+	else if (entry->fe_type == OCFS2_FILECHECK_TYPE_FIX)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_FIX);
+	else
+		entry->fe_status = OCFS2_FILECHECK_ERR_UNSUPPORTED;
+
+	ocfs2_filecheck_done_entry(ent, entry);
+}
+
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+				struct kobj_attribute *attr,
+				const char *buf, size_t count)
+{
+	struct ocfs2_filecheck_args args;
+	struct ocfs2_filecheck_entry *entry = NULL;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+	ssize_t ret = 0;
+
+	if (count == 0)
+		return count;
+
+	if (ocfs2_filecheck_args_parse(buf, count, &args)) {
+		mlog(ML_ERROR, "Invalid arguments for online file check\n");
+		return -EINVAL;
+	}
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	if (args.fa_type == OCFS2_FILECHECK_TYPE_SET) {
+		ret = ocfs2_filecheck_adjust_max(ent, args.fa_len);
+		ocfs2_filecheck_sysfs_put(ent);
+		return (!ret ? count : ret);
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+		(ent->fs_fcheck->fc_done == 0)) {
+		mlog(ML_ERROR,
+		"Online file check queue(%u) is full\n",
+		ent->fs_fcheck->fc_max);
+		ret = -EBUSY;
+	} else {
+		if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+			(ent->fs_fcheck->fc_done > 0)) {
+			/* Delete the oldest entry which was done,
+			 * make sure the entry size in list does
+			 * not exceed maximum value
+			 */
+			BUG_ON(!ocfs2_filecheck_erase_entry(ent));
+		}
+
+		entry = kmalloc(sizeof(struct ocfs2_filecheck_entry), GFP_NOFS);
+		if (entry) {
+			entry->fe_ino = args.fa_ino;
+			entry->fe_type = args.fa_type;
+			entry->fe_done = 0;
+			entry->fe_status = OCFS2_FILECHECK_ERR_INPROGRESS;
+			list_add_tail(&entry->fe_list,
+					&ent->fs_fcheck->fc_head);
+
+			ent->fs_fcheck->fc_size++;
+			ret = count;
+		} else {
+			ret = -ENOMEM;
+		}
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	if (entry)
+		ocfs2_filecheck_handle_entry(ent, entry);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return ret;
+}
fs/ocfs2/filecheck.h new file +48 -0
new file mode 100644
index 0000000..5ec331b
@@ -0,0 +1,48 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.h
+ *
+ * Online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+
+#ifndef FILECHECK_H
+#define FILECHECK_H
+
+#include <linux/types.h>
+#include <linux/list.h>
+
+
+/* File check errno */
+enum {
+	OCFS2_FILECHECK_ERR_SUCCESS = 0,	/* Success */
+	OCFS2_FILECHECK_ERR_FAILED = 1000,	/* Other failure */
+	OCFS2_FILECHECK_ERR_INPROGRESS,		/* In progress */
+	OCFS2_FILECHECK_ERR_READONLY,		/* Read only */
+	OCFS2_FILECHECK_ERR_INVALIDINO,		/* Invalid ino */
+	OCFS2_FILECHECK_ERR_BLOCKECC,		/* Block ecc */
+	OCFS2_FILECHECK_ERR_BLOCKNO,		/* Block number */
+	OCFS2_FILECHECK_ERR_VALIDFLAG,		/* Inode valid flag */
+	OCFS2_FILECHECK_ERR_GENERATION,		/* Inode generation */
+	OCFS2_FILECHECK_ERR_UNSUPPORTED		/* Unsupported */
+};
+
+#define OCFS2_FILECHECK_ERR_START	OCFS2_FILECHECK_ERR_FAILED
+#define OCFS2_FILECHECK_ERR_END		OCFS2_FILECHECK_ERR_UNSUPPORTED
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb);
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb);
+
+#endif  /* FILECHECK_H */
fs/ocfs2/inode.h +0 -0
index 5e86b24..abd1018 100644

fs/ocfs2/inode.h int ocfs2_drop_inode(struct inode *inode); · line 146

Quoted 8 lines
fs/ocfs2/inode.h +3 -0
@@ -139,6 +139,9 @@ int ocfs2_drop_inode(struct inode *inode);
 /* Flags for ocfs2_iget() */
 #define OCFS2_FI_FLAG_SYSFILE		0x1
 #define OCFS2_FI_FLAG_ORPHAN_RECOVERY	0x2
+#define OCFS2_FI_FLAG_FILECHECK_CHK	0x4
+#define OCFS2_FI_FLAG_FILECHECK_FIX	0x8
+
 struct inode *ocfs2_ilookup(struct super_block *sb, u64 feoff);
 struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned 

flags,

Quoted 2 lines
fs/ocfs2/inode.h +0 -0
 			 int sysfile_type);
author
#7 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Hi Junxiao,

Quoted 17 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

Quoted 4 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via

pass a ino number).

Quoted 1 line

But, this operation is manually triggered by user, instead of automatically

fix in the kernel.

Quoted 2 lines

Why?
1) we should let users make this decision, since some users do not want to

fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.
If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Quoted 5 lines
Quoted 1 line

2) frankly speaking, this feature will probably bring a second corruption if

there is some error in the code, I do not suggest to use automatically fix by default in the first version.
I think if this feature could bring more corruption, then this should be fixed first.

In theory, this feature will avoid bringing any second corruption after our detailed reviewing and discussion.
but, my means is that if there is any carelessness due to our experience limitation, it will probably bring a accident second corruption. this is why, I do not suggest to use automatically fix by default in the kernel when a feature is firstly introduced.

Quoted 698 lines

Thanks,
Junxiao

Quoted 1 line

3) in the future, if this feature is well proved, we can add a mount option

to make this automatically fix enabled.

Quoted 693 lines

Thanks
Gang

Quoted 686 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

I didn't see a need to add a sysfs file for the check and repair. This leaves a hard problem for customer to decide. How they decide whether they should repair the bad inode since this may cause corruption even harder?
I think the error should be fixed by this feature automaticlly if repair helps, of course this can be done only when error=continue is enabled or add some mount option for it.

Thanks,
Junxiao.

Quoted 8 lines · Gang He

Implement online file check sysfile interfaces, e.g. how to create the related sysfile according to device name, how to display/handle file check request from the sysfile.

---
 fs/ocfs2/Makefile    |   3 +-
 fs/ocfs2/filecheck.c | 566 

+++++++++++++++++++++++++++++++++++++++++++++++++++

Quoted 651 lines
 fs/ocfs2/filecheck.h |  48 +++++
 fs/ocfs2/inode.h     |   3 +
 4 files changed, 619 insertions(+), 1 deletion(-)
 create mode 100644 fs/ocfs2/filecheck.c
 create mode 100644 fs/ocfs2/filecheck.h
fs/ocfs2/Makefile +2 -1
index ce210d4..e27e652 100644
@@ -41,7 +41,8 @@ ocfs2-objs := \
 	quota_local.o		\
 	quota_global.o		\
 	xattr.o			\
-	acl.o
+	acl.o	\
+	filecheck.o
 
 ocfs2_stackglue-objs := stackglue.o
 ocfs2_stack_o2cb-objs := stack_o2cb.o
fs/ocfs2/filecheck.c new file +566 -0
new file mode 100644
index 0000000..f12ed1f
@@ -0,0 +1,566 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.c
+ *
+ * Code which implements online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+#include <linux/list.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <linux/kmod.h>
+#include <linux/fs.h>
+#include <linux/kobject.h>
+#include <linux/sysfs.h>
+#include <linux/sysctl.h>
+#include <cluster/masklog.h>
+
+#include "ocfs2.h"
+#include "ocfs2_fs.h"
+#include "stackglue.h"
+#include "inode.h"
+
+#include "filecheck.h"
+
+
+/* File check error strings,
+ * must correspond with error number in header file.
+ */
+static const char * const ocfs2_filecheck_errs[] = {
+	"SUCCESS",
+	"FAILED",
+	"INPROGRESS",
+	"READONLY",
+	"INVALIDINO",
+	"BLOCKECC",
+	"BLOCKNO",
+	"VALIDFLAG",
+	"GENERATION",
+	"UNSUPPORTED"
+};
+
+static DEFINE_SPINLOCK(ocfs2_filecheck_sysfs_lock);
+static LIST_HEAD(ocfs2_filecheck_sysfs_list);
+
+struct ocfs2_filecheck {
+	struct list_head fc_head;	/* File check entry list head */
+	spinlock_t fc_lock;
+	unsigned int fc_max;	/* Maximum number of entry in list */
+	unsigned int fc_size;	/* Current entry count in list */
+	unsigned int fc_done;	/* File check entries are done in list */
+};
+
+struct ocfs2_filecheck_sysfs_entry {
+	struct list_head fs_list;
+	atomic_t fs_count;
+	struct super_block *fs_sb;
+	struct kset *fs_kset;
+	struct ocfs2_filecheck *fs_fcheck;
+};
+
+#define OCFS2_FILECHECK_MAXSIZE		100
+#define OCFS2_FILECHECK_MINSIZE		10
+
+/* File check operation type */
+enum {
+	OCFS2_FILECHECK_TYPE_CHK = 0,	/* Check a file */
+	OCFS2_FILECHECK_TYPE_FIX,	/* Fix a file */
+	OCFS2_FILECHECK_TYPE_SET = 100	/* Set file check options */
+};
+
+struct ocfs2_filecheck_entry {
+	struct list_head fe_list;
+	unsigned long fe_ino;
+	unsigned int fe_type;
+	unsigned short fe_done:1;
+	unsigned short fe_status:15;
+};
+
+struct ocfs2_filecheck_args {
+	unsigned int fa_type;
+	union {
+		unsigned long fa_ino;
+		unsigned int fa_len;
+	};
+};
+
+static const char *
+ocfs2_filecheck_error(int errno)
+{
+	if (!errno)
+		return ocfs2_filecheck_errs[errno];
+
+	BUG_ON(errno < OCFS2_FILECHECK_ERR_START ||
+			errno > OCFS2_FILECHECK_ERR_END);
+	return ocfs2_filecheck_errs[errno - OCFS2_FILECHECK_ERR_START + 1];
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf);
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					const char *buf, size_t count);
+static struct kobj_attribute ocfs2_attr_filecheck =
+					__ATTR(filecheck, S_IRUSR | S_IWUSR,
+					ocfs2_filecheck_show,
+					ocfs2_filecheck_store);
+
+static int ocfs2_filecheck_sysfs_wait(atomic_t *p)
+{
+	schedule();
+	return 0;
+}
+
+static void
+ocfs2_filecheck_sysfs_free(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	if (!atomic_dec_and_test(&entry->fs_count))
+		wait_on_atomic_t(&entry->fs_count, ocfs2_filecheck_sysfs_wait,
+						TASK_UNINTERRUPTIBLE);
+
+	spin_lock(&entry->fs_fcheck->fc_lock);
+	while (!list_empty(&entry->fs_fcheck->fc_head)) {
+		p = list_first_entry(&entry->fs_fcheck->fc_head,
+				struct ocfs2_filecheck_entry, fe_list);
+		list_del(&p->fe_list);
+		BUG_ON(!p->fe_done); /* To free a undone file check entry */
+		kfree(p);
+	}
+	spin_unlock(&entry->fs_fcheck->fc_lock);
+
+	kset_unregister(entry->fs_kset);
+	kfree(entry->fs_fcheck);
+	kfree(entry);
+}
+
+static void
+ocfs2_filecheck_sysfs_add(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_add_tail(&entry->fs_list, &ocfs2_filecheck_sysfs_list);
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+}
+
+static int ocfs2_filecheck_sysfs_del(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			list_del(&p->fs_list);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			ocfs2_filecheck_sysfs_free(p);
+			return 0;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return 1;
+}
+
+static void
+ocfs2_filecheck_sysfs_put(struct ocfs2_filecheck_sysfs_entry *entry)
+{
+	if (atomic_dec_and_test(&entry->fs_count))
+		wake_up_atomic_t(&entry->fs_count);
+}
+
+static struct ocfs2_filecheck_sysfs_entry *
+ocfs2_filecheck_sysfs_get(const char *devname)
+{
+	struct ocfs2_filecheck_sysfs_entry *p = NULL;
+
+	spin_lock(&ocfs2_filecheck_sysfs_lock);
+	list_for_each_entry(p, &ocfs2_filecheck_sysfs_list, fs_list) {
+		if (!strcmp(p->fs_sb->s_id, devname)) {
+			atomic_inc(&p->fs_count);
+			spin_unlock(&ocfs2_filecheck_sysfs_lock);
+			return p;
+		}
+	}
+	spin_unlock(&ocfs2_filecheck_sysfs_lock);
+	return NULL;
+}
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb)
+{
+	int ret = 0;
+	struct kset *ocfs2_filecheck_kset = NULL;
+	struct ocfs2_filecheck *fcheck = NULL;
+	struct ocfs2_filecheck_sysfs_entry *entry = NULL;
+	struct attribute **attrs = NULL;
+	struct attribute_group attrgp;
+
+	if (!ocfs2_kset)
+		return -ENOMEM;
+
+	attrs = kmalloc(sizeof(struct attribute *) * 2, GFP_NOFS);
+	if (!attrs) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		attrs[0] = &ocfs2_attr_filecheck.attr;
+		attrs[1] = NULL;
+		memset(&attrgp, 0, sizeof(attrgp));
+		attrgp.attrs = attrs;
+	}
+
+	fcheck = kmalloc(sizeof(struct ocfs2_filecheck), GFP_NOFS);
+	if (!fcheck) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		INIT_LIST_HEAD(&fcheck->fc_head);
+		spin_lock_init(&fcheck->fc_lock);
+		fcheck->fc_max = OCFS2_FILECHECK_MINSIZE;
+		fcheck->fc_size = 0;
+		fcheck->fc_done = 0;
+	}
+
+	if (strlen(sb->s_id) <= 0) {
+		mlog(ML_ERROR,
+		"Cannot get device basename when create filecheck sysfs\n");
+		ret = -ENODEV;
+		goto error;
+	}
+
+	ocfs2_filecheck_kset = kset_create_and_add(sb->s_id, NULL,
+						&ocfs2_kset->kobj);
+	if (!ocfs2_filecheck_kset) {
+		ret = -ENOMEM;
+		goto error;
+	}
+
+	ret = sysfs_create_group(&ocfs2_filecheck_kset->kobj, &attrgp);
+	if (ret)
+		goto error;
+
+	entry = kmalloc(sizeof(struct ocfs2_filecheck_sysfs_entry), GFP_NOFS);
+	if (!entry) {
+		ret = -ENOMEM;
+		goto error;
+	} else {
+		atomic_set(&entry->fs_count, 1);
+		entry->fs_sb = sb;
+		entry->fs_kset = ocfs2_filecheck_kset;
+		entry->fs_fcheck = fcheck;
+		ocfs2_filecheck_sysfs_add(entry);
+	}
+
+	kfree(attrs);
+	return 0;
+
+error:
+	kfree(attrs);
+	kfree(entry);
+	kfree(fcheck);
+	kset_unregister(ocfs2_filecheck_kset);
+	return ret;
+}
+
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb)
+{
+	return ocfs2_filecheck_sysfs_del(sb->s_id);
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count);
+static int
+ocfs2_filecheck_adjust_max(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int len)
+{
+	int ret;
+
+	if ((len < OCFS2_FILECHECK_MINSIZE) || (len > OCFS2_FILECHECK_MAXSIZE))
+		return -EINVAL;
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if (len < (ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done)) {
+		mlog(ML_ERROR,
+		"Cannot set online file check maximum entry number "
+		"to %u due to too much pending entries(%u)\n",
+		len, ent->fs_fcheck->fc_size - ent->fs_fcheck->fc_done);
+		ret = -EBUSY;
+	} else {
+		if (len < ent->fs_fcheck->fc_size)
+			BUG_ON(!ocfs2_filecheck_erase_entries(ent,
+				ent->fs_fcheck->fc_size - len));
+
+		ent->fs_fcheck->fc_max = len;
+		ret = 0;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	return ret;
+}
+
+#define OCFS2_FILECHECK_ARGS_LEN	32
+static int
+ocfs2_filecheck_args_get_long(const char *buf, size_t count,
+				unsigned long *val)
+{
+	char buffer[OCFS2_FILECHECK_ARGS_LEN];
+
+	if (count < 1)
+		return 1;
+
+	memcpy(buffer, buf, count);
+	buffer[count] = '\0';
+
+	if (kstrtoul(buffer, 0, val))
+		return 1;
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_args_parse(const char *buf, size_t count,
+				struct ocfs2_filecheck_args *args)
+{
+	unsigned long val = 0;
+
+	/* too short/long args length */
+	if ((count < 5) || (count > OCFS2_FILECHECK_ARGS_LEN))
+		return 1;
+
+	if (!strncasecmp(buf, "FIX ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_FIX;
+		args->fa_ino = val;
+		return 0;
+	} else if ((count > 6) && !strncasecmp(buf, "CHECK ", 6)) {
+		if (ocfs2_filecheck_args_get_long(buf + 6, count - 6, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_CHK;
+		args->fa_ino = val;
+		return 0;
+	} else if (!strncasecmp(buf, "SET ", 4)) {
+		if (ocfs2_filecheck_args_get_long(buf + 4, count - 4, &val))
+			return 1;
+
+		args->fa_type = OCFS2_FILECHECK_TYPE_SET;
+		args->fa_len = (unsigned int)val;
+		return 0;
+	} else { /* invalid args */
+		return 1;
+	}
+}
+
+static ssize_t ocfs2_filecheck_show(struct kobject *kobj,
+					struct kobj_attribute *attr,
+					char *buf)
+{
+
+	ssize_t ret = 0, total = 0, remain = PAGE_SIZE;
+	struct ocfs2_filecheck_entry *p;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ret = snprintf(buf, remain, "INO\t\tTYPE\tDONE\tERROR\n");
+	total += ret;
+	remain -= ret;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		ret = snprintf(buf + total, remain, "%lu\t\t%u\t%u\t%s\n",
+			p->fe_ino, p->fe_type, p->fe_done,
+			ocfs2_filecheck_error(p->fe_status));
+		if (ret < 0) {
+			total = ret;
+			break;
+		}
+		if (ret == remain) {
+			/* snprintf() didn't fit */
+			total = -E2BIG;
+			break;
+		}
+		total += ret;
+		remain -= ret;
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return total;
+}
+
+static int
+ocfs2_filecheck_erase_entry(struct ocfs2_filecheck_sysfs_entry *ent)
+{
+	struct ocfs2_filecheck_entry *p;
+
+	list_for_each_entry(p, &ent->fs_fcheck->fc_head, fe_list) {
+		if (p->fe_done) {
+			list_del(&p->fe_list);
+			kfree(p);
+			ent->fs_fcheck->fc_size--;
+			ent->fs_fcheck->fc_done--;
+			return 1;
+		}
+	}
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_erase_entries(struct ocfs2_filecheck_sysfs_entry *ent,
+				unsigned int count)
+{
+	unsigned int i = 0;
+	unsigned int ret = 0;
+
+	while (i++ < count) {
+		if (ocfs2_filecheck_erase_entry(ent))
+			ret++;
+		else
+			break;
+	}
+
+	return (ret == count ? 1 : 0);
+}
+
+static void
+ocfs2_filecheck_done_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	entry->fe_done = 1;
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	ent->fs_fcheck->fc_done++;
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+}
+
+static unsigned short
+ocfs2_filecheck_handle(struct super_block *sb,
+				unsigned long ino, unsigned int flags)
+{
+	unsigned short ret = OCFS2_FILECHECK_ERR_SUCCESS;
+	struct inode *inode = NULL;
+	int rc;
+
+	inode = ocfs2_iget(OCFS2_SB(sb), ino, flags, 0);
+	if (IS_ERR(inode)) {
+		rc = (int)(-(long)inode);
+		if (rc >= OCFS2_FILECHECK_ERR_START &&
+			rc < OCFS2_FILECHECK_ERR_END)
+			ret = rc;
+		else
+			ret = OCFS2_FILECHECK_ERR_FAILED;
+	} else
+		iput(inode);
+
+	return ret;
+}
+
+static void
+ocfs2_filecheck_handle_entry(struct ocfs2_filecheck_sysfs_entry *ent,
+				struct ocfs2_filecheck_entry *entry)
+{
+	if (entry->fe_type == OCFS2_FILECHECK_TYPE_CHK)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_CHK);
+	else if (entry->fe_type == OCFS2_FILECHECK_TYPE_FIX)
+		entry->fe_status = ocfs2_filecheck_handle(ent->fs_sb,
+				entry->fe_ino, OCFS2_FI_FLAG_FILECHECK_FIX);
+	else
+		entry->fe_status = OCFS2_FILECHECK_ERR_UNSUPPORTED;
+
+	ocfs2_filecheck_done_entry(ent, entry);
+}
+
+static ssize_t ocfs2_filecheck_store(struct kobject *kobj,
+				struct kobj_attribute *attr,
+				const char *buf, size_t count)
+{
+	struct ocfs2_filecheck_args args;
+	struct ocfs2_filecheck_entry *entry = NULL;
+	struct ocfs2_filecheck_sysfs_entry *ent;
+	ssize_t ret = 0;
+
+	if (count == 0)
+		return count;
+
+	if (ocfs2_filecheck_args_parse(buf, count, &args)) {
+		mlog(ML_ERROR, "Invalid arguments for online file check\n");
+		return -EINVAL;
+	}
+
+	ent = ocfs2_filecheck_sysfs_get(kobj->name);
+	if (!ent) {
+		mlog(ML_ERROR,
+		"Cannot get the corresponding entry via device basename %s\n",
+		kobj->name);
+		return -ENODEV;
+	}
+
+	if (args.fa_type == OCFS2_FILECHECK_TYPE_SET) {
+		ret = ocfs2_filecheck_adjust_max(ent, args.fa_len);
+		ocfs2_filecheck_sysfs_put(ent);
+		return (!ret ? count : ret);
+	}
+
+	spin_lock(&ent->fs_fcheck->fc_lock);
+	if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+		(ent->fs_fcheck->fc_done == 0)) {
+		mlog(ML_ERROR,
+		"Online file check queue(%u) is full\n",
+		ent->fs_fcheck->fc_max);
+		ret = -EBUSY;
+	} else {
+		if ((ent->fs_fcheck->fc_size >= ent->fs_fcheck->fc_max) &&
+			(ent->fs_fcheck->fc_done > 0)) {
+			/* Delete the oldest entry which was done,
+			 * make sure the entry size in list does
+			 * not exceed maximum value
+			 */
+			BUG_ON(!ocfs2_filecheck_erase_entry(ent));
+		}
+
+		entry = kmalloc(sizeof(struct ocfs2_filecheck_entry), GFP_NOFS);
+		if (entry) {
+			entry->fe_ino = args.fa_ino;
+			entry->fe_type = args.fa_type;
+			entry->fe_done = 0;
+			entry->fe_status = OCFS2_FILECHECK_ERR_INPROGRESS;
+			list_add_tail(&entry->fe_list,
+					&ent->fs_fcheck->fc_head);
+
+			ent->fs_fcheck->fc_size++;
+			ret = count;
+		} else {
+			ret = -ENOMEM;
+		}
+	}
+	spin_unlock(&ent->fs_fcheck->fc_lock);
+
+	if (entry)
+		ocfs2_filecheck_handle_entry(ent, entry);
+
+	ocfs2_filecheck_sysfs_put(ent);
+	return ret;
+}
fs/ocfs2/filecheck.h new file +48 -0
new file mode 100644
index 0000000..5ec331b
@@ -0,0 +1,48 @@
+/* -*- mode: c; c-basic-offset: 8; -*-
+ * vim: noexpandtab sw=8 ts=8 sts=0:
+ *
+ * filecheck.h
+ *
+ * Online file check.
+ *
+ * Copyright (C) 2015 Novell.  All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public
+ * License as published by the Free Software Foundation, version 2.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ */
+
+
+#ifndef FILECHECK_H
+#define FILECHECK_H
+
+#include <linux/types.h>
+#include <linux/list.h>
+
+
+/* File check errno */
+enum {
+	OCFS2_FILECHECK_ERR_SUCCESS = 0,	/* Success */
+	OCFS2_FILECHECK_ERR_FAILED = 1000,	/* Other failure */
+	OCFS2_FILECHECK_ERR_INPROGRESS,		/* In progress */
+	OCFS2_FILECHECK_ERR_READONLY,		/* Read only */
+	OCFS2_FILECHECK_ERR_INVALIDINO,		/* Invalid ino */
+	OCFS2_FILECHECK_ERR_BLOCKECC,		/* Block ecc */
+	OCFS2_FILECHECK_ERR_BLOCKNO,		/* Block number */
+	OCFS2_FILECHECK_ERR_VALIDFLAG,		/* Inode valid flag */
+	OCFS2_FILECHECK_ERR_GENERATION,		/* Inode generation */
+	OCFS2_FILECHECK_ERR_UNSUPPORTED		/* Unsupported */
+};
+
+#define OCFS2_FILECHECK_ERR_START	OCFS2_FILECHECK_ERR_FAILED
+#define OCFS2_FILECHECK_ERR_END		OCFS2_FILECHECK_ERR_UNSUPPORTED
+
+int ocfs2_filecheck_create_sysfs(struct super_block *sb);
+int ocfs2_filecheck_remove_sysfs(struct super_block *sb);
+
+#endif  /* FILECHECK_H */
fs/ocfs2/inode.h +0 -0
index 5e86b24..abd1018 100644

fs/ocfs2/inode.h int ocfs2_drop_inode(struct inode *inode); · line 146

Quoted 8 lines
fs/ocfs2/inode.h +3 -0
@@ -139,6 +139,9 @@ int ocfs2_drop_inode(struct inode *inode);
 /* Flags for ocfs2_iget() */
 #define OCFS2_FI_FLAG_SYSFILE		0x1
 #define OCFS2_FI_FLAG_ORPHAN_RECOVERY	0x2
+#define OCFS2_FI_FLAG_FILECHECK_CHK	0x4
+#define OCFS2_FI_FLAG_FILECHECK_FIX	0x8
+
 struct inode *ocfs2_ilookup(struct super_block *sb, u64 feoff);
 struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 feoff, unsigned 

flags,

Quoted 2 lines
fs/ocfs2/inode.h +0 -0
 			 int sysfile_type);
#8 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

In reply to Junxiao Bi
Quoted 5 lines · Junxiao Bi

Hi Gang,

Quoted 2 lines · Gang He

Hi Junxiao,

Quoted 7 lines
Quoted 5 lines

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via pass a ino number).
But, this operation is manually triggered by user, instead of automatically fix in the kernel.
Why?
1) we should let users make this decision, since some users do not want to fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

Quoted 3 lines
Quoted 1 line

2) frankly speaking, this feature will probably bring a second corruption if there is some error in the code, I do not suggest to use automatically fix by default in the first version.

I think if this feature could bring more corruption, then this should be fixed first.

Btw, I am pretty sure that Gang is referring to the feature being new and thus more likely to have problems. There is nothing I see in here that is file system corrupting.

	--Mark

--
Mark Fasheh

#9 lore raw

Re: [Ocfs2-devel] [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Quoted 6 lines · Mark Fasheh
Quoted 5 lines · Junxiao Bi

Hi Gang,

Quoted 2 lines · Gang He

Hi Junxiao,

Quoted 8 lines
Quoted 7 lines
Quoted 5 lines

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via pass a ino number).
But, this operation is manually triggered by user, instead of automatically fix in the kernel.
Why?
1) we should let users make this decision, since some users do not want to fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

Great idea Mark! I think adding errors=fix would be a good way to address both concerns :) It gives some control if anyone is uncomfortable of things getting checked/fixed automatically.

Quoted 36 lines

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

Quoted 3 lines
Quoted 1 line

2) frankly speaking, this feature will probably bring a second corruption if there is some error in the code, I do not suggest to use automatically fix by default in the first version.

I think if this feature could bring more corruption, then this should be fixed first.

Btw, I am pretty sure that Gang is referring to the feature being new and thus more likely to have problems. There is nothing I see in here that is file system corrupting.

	--Mark

--
Mark Fasheh

_______________________________________________
Ocfs2-devel mailing list
[email protected]
https://oss.oracle.com/mailman/listinfo/ocfs2-devel

author
#10 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

In reply to Mark Fasheh

Hi Mark and Junxiao,

Quoted 31 lines
Quoted 1 line
Quoted 1 line
Quoted 7 lines · Junxiao Bi

Hi Gang,

Quoted 4 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via

pass a ino number).

Quoted 1 line
Quoted 1 line

But, this operation is manually triggered by user, instead of automatically

fix in the kernel.

Quoted 2 lines
Quoted 2 lines

Why?
1) we should let users make this decision, since some users do not want to

fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

Quoted 2 lines

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

Quoted 7 lines

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

That is what I want to say, add a mount option to let users to decide. Here, I want to split "error=fix" mount option task out from online file check feature, I think this part should be a independent feature. We can implement this feature after online file check is done, I want to split the feature into some more detailed features, implement them one by one. Do you agree this point?

Quoted 16 lines
Quoted 3 lines
Quoted 3 lines

2) frankly speaking, this feature will probably bring a second corruption

if there is some error in the code, I do not suggest to use automatically fix by default in the first version.

Quoted 2 lines

I think if this feature could bring more corruption, then this should be fixed first.

Btw, I am pretty sure that Gang is referring to the feature being new and thus more likely to have problems. There is nothing I see in here that is file system corrupting.

	--Mark

--
Mark Fasheh

#11 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Hi Gang,

Quoted 38 lines · Gang He

Hi Mark and Junxiao,

Quoted 35 lines
Quoted 1 line
Quoted 1 line
Quoted 7 lines · Junxiao Bi

Hi Gang,

Quoted 4 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via

pass a ino number).

Quoted 1 line
Quoted 1 line

But, this operation is manually triggered by user, instead of automatically

fix in the kernel.

Quoted 2 lines
Quoted 2 lines

Why?
1) we should let users make this decision, since some users do not want to

fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

Quoted 2 lines

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

Quoted 7 lines
Quoted 3 lines

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

That is what I want to say, add a mount option to let users to decide. Here, I want to split "error=fix" mount option task out from online file check feature, I think this part should be a independent feature. We can implement this feature after online file check is done, I want to split the feature into some more detailed features, implement them one by one. Do you agree this point?

With error=fix, when a possible corruption is found, online fsck will start to check and fix things. So this doesn't looks like a independent feature.

Thanks,
Junxiao.

Quoted 18 lines
Quoted 17 lines
Quoted 4 lines
Quoted 4 lines

2) frankly speaking, this feature will probably bring a second corruption

if there is some error in the code, I do not suggest to use automatically fix by default in the first version.

Quoted 2 lines

I think if this feature could bring more corruption, then this should be fixed first.

Btw, I am pretty sure that Gang is referring to the feature being new and thus more likely to have problems. There is nothing I see in here that is file system corrupting.

	--Mark

--
Mark Fasheh

author
#12 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

Hi Junxiao,

Quoted 47 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

Quoted 34 lines · Gang He

Hi Mark and Junxiao,

Quoted 31 lines
Quoted 1 line
Quoted 1 line
Quoted 7 lines · Junxiao Bi

Hi Gang,

Quoted 4 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via

pass a ino number).

Quoted 1 line
Quoted 1 line

But, this operation is manually triggered by user, instead of automatically

fix in the kernel.

Quoted 2 lines
Quoted 2 lines

Why?
1) we should let users make this decision, since some users do not want to

fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

Quoted 2 lines

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

Quoted 8 lines
Quoted 7 lines

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

That is what I want to say, add a mount option to let users to decide. Here,

I want to split "error=fix"

Quoted 8 lines
Quoted 1 line

mount option task out from online file check feature, I think this part

should be a independent feature.

Quoted 1 line

We can implement this feature after online file check is done, I want to

split the feature into some more

Quoted 1 line

detailed features, implement them one by one. Do you agree this point?

With error=fix, when a possible corruption is found, online fsck will start to check and fix things. So this doesn't looks like a independent feature.

My means is, we can implement online file check by manually triage feature first, then Add a mount option "error=fix" feature, the second feature can be implemented after the first part is done. I want to split them into more detailed items, maybe it is more helpful to be reviewed, but the whole feature ideas are very OK, just need to do one by one.

Quoted 23 lines

Thanks,
Junxiao.

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Quoted 4 lines

2) frankly speaking, this feature will probably bring a second corruption

if there is some error in the code, I do not suggest to use automatically

fix

Quoted 13 lines
Quoted 12 lines

by default in the first version.

Quoted 2 lines

I think if this feature could bring more corruption, then this should be fixed first.

Btw, I am pretty sure that Gang is referring to the feature being new and thus more likely to have problems. There is nothing I see in here that is file system corrupting.

	--Mark

--
Mark Fasheh

#13 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

In reply to Gang He
Quoted 38 lines · Gang He

Hi Mark and Junxiao,

Quoted 35 lines
Quoted 1 line
Quoted 1 line
Quoted 7 lines · Junxiao Bi

Hi Gang,

Quoted 4 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via

pass a ino number).

Quoted 1 line
Quoted 1 line

But, this operation is manually triggered by user, instead of automatically

fix in the kernel.

Quoted 2 lines
Quoted 2 lines

Why?
1) we should let users make this decision, since some users do not want to

fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

Quoted 2 lines

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

Quoted 7 lines
Quoted 3 lines

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

That is what I want to say, add a mount option to let users to decide. Here, I want to split "error=fix" mount option task out from online file check feature, I think this part should be a independent feature. We can implement this feature after online file check is done, I want to split the feature into some more detailed features, implement them one by one. Do you agree this point?

Yeah that's fine, I would have automatic checking turned off though until we have a good plan in place for users who do / don't want this.

	--Mark

--
Mark Fasheh

#14 lore raw

Re: [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

In reply to Mark Fasheh
Quoted 11 lines · Mark Fasheh
Quoted 10 lines · Junxiao Bi

Hi Gang,

Quoted 7 lines · Gang He

Hi Junxiao,

Thank for your reviewing.
Current design, we use a sysfile as a interface to check/fix a file (via pass a ino number).
But, this operation is manually triggered by user, instead of automatically fix in the kernel.
Why?
1) we should let users make this decision, since some users do not want to fix when encountering a file system corruption, maybe they want to keep the file system unchanged for a further investigation.

Quoted 4 lines
Quoted 2 lines

If user don't want this, they should not use error=continue option, let fs go after a corruption is very dangerous.

Maybe we need another errors=XXX flag (maybe errors=fix)?

Sound great. This is a good option since user may have not enough knowledge whether to fix the found issue.

Thanks,
Junxiao.

Quoted 33 lines

You both make good points, here's what I gather from the conversation:

- Some customers would be sad if they have to manually fix corruptions. This takes effort on their part, and if the FS can handle it automatically, it should.

- There are valid concerns that automatically fixing things is a change in behavior that might not be welcome, or worse might lead to unforseeable circumstances.

- I will add that fixing things automatically implies checking them automatically which could introduce some performance impact depending on how much checking we're doing.

So if the user wants errors to be fixed automatically, they could mount with errros=fix, and everyone else would have no change in behavior unless they wanted to make use of the new feature.

Quoted 3 lines
Quoted 1 line

2) frankly speaking, this feature will probably bring a second corruption if there is some error in the code, I do not suggest to use automatically fix by default in the first version.

I think if this feature could bring more corruption, then this should be fixed first.

Btw, I am pretty sure that Gang is referring to the feature being new and thus more likely to have problems. There is nothing I see in here that is file system corrupting.

	--Mark

--
Mark Fasheh

#15 lore raw

Re: [Ocfs2-devel] [PATCH v2 2/4] ocfs2: sysfile interfaces for online file check

In reply to Gang He
Quoted 5 lines · Gang He

Implement online file check sysfile interfaces, e.g. how to create the related sysfile according to device name, how to display/handle file check request from the sysfile.

FYI, This looks generally fine to me however we should address Junxiao's concerns before it goes any further.

	--Mark

--
Mark Fasheh

author
#16 lore raw

PATCH v2 3/4 ocfs2: create/remove sysfile for online file check

In reply to Gang He

Create online file check sysfile when ocfs2 mount, remove the related sysfile when ocfs2 umount.

---
 fs/ocfs2/super.c | 5 +++++
 1 file changed, 5 insertions(+)
fs/ocfs2/super.c +5 -0
index 403c566..7213a94 100644
@@ -74,6 +74,7 @@
 #include "suballoc.h"
 
 #include "buffer_head_io.h"
+#include "filecheck.h"
 
 static struct kmem_cache *ocfs2_inode_cachep;
 struct kmem_cache *ocfs2_dquot_cachep;
@@ -1202,6 +1203,9 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent)
 	/* Start this when the mount is almost sure of being successful */
 	ocfs2_orphan_scan_start(osb);
 
+	/* Create filecheck sysfile /sys/fs/ocfs2/<devname>/filecheck */
+	ocfs2_filecheck_create_sysfs(sb);
+
 	return status;
 
 read_super_error:
@@ -1658,6 +1662,7 @@ static void ocfs2_put_super(struct super_block *sb)
 
 	ocfs2_sync_blockdev(sb);
 	ocfs2_dismount_volume(sb, 0);
+	ocfs2_filecheck_remove_sysfs(sb);
 }
 
 static int ocfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
Signature · 1 line
2.1.2
author
#18 lore raw

PATCH v2 4/4 ocfs2: check/fix inode block for online file check

In reply to Gang He

Implement online check or fix inode block during
reading a inode block to memory.

---
 fs/ocfs2/inode.c       | 196 +++++++++++++++++++++++++++++++++++++++++++++++--
 fs/ocfs2/ocfs2_trace.h |   2 +
 2 files changed, 192 insertions(+), 6 deletions(-)
fs/ocfs2/inode.c +190 -6
index b254416..d811698 100644
@@ -53,6 +53,7 @@
 #include "xattr.h"
 #include "refcounttree.h"
 #include "ocfs2_trace.h"
+#include "filecheck.h"
 
 #include "buffer_head_io.h"
 
@@ -74,6 +75,13 @@ static int ocfs2_truncate_for_delete(struct ocfs2_super *osb,
 				    struct inode *inode,
 				    struct buffer_head *fe_bh);
 
+static int ocfs2_filecheck_read_inode_block_full(struct inode *inode,
+			struct buffer_head **bh, int flags, int type);
+static int ocfs2_filecheck_validate_inode_block(struct super_block *sb,
+			struct buffer_head *bh);
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			struct buffer_head *bh);
+
 void ocfs2_set_inode_flags(struct inode *inode)
 {
 	unsigned int flags = OCFS2_I(inode)->ip_attr;
@@ -127,6 +135,7 @@ struct inode *ocfs2_ilookup(struct super_block *sb, u64 blkno)
 struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
 			 int sysfile_type)
 {
+	int rc = 0;
 	struct inode *inode = NULL;
 	struct super_block *sb = osb->sb;
 	struct ocfs2_find_inode_args args;
@@ -161,12 +170,17 @@ struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned flags,
 	}
 	trace_ocfs2_iget5_locked(inode->i_state);
 	if (inode->i_state & I_NEW) {
-		ocfs2_read_locked_inode(inode, &args);
+		rc = ocfs2_read_locked_inode(inode, &args);
 		unlock_new_inode(inode);
 	}
 	if (is_bad_inode(inode)) {
 		iput(inode);
-		inode = ERR_PTR(-ESTALE);
+		if ((flags & OCFS2_FI_FLAG_FILECHECK_CHK) ||
+			(flags & OCFS2_FI_FLAG_FILECHECK_FIX))
+			/* Return OCFS2_FILECHECK_ERR_XXX related errno */
+			inode = ERR_PTR(rc);
+		else
+			inode = ERR_PTR(-ESTALE);
 		goto bail;
 	}
 
@@ -494,16 +508,32 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 	}
 
 	if (can_lock) {
-		status = ocfs2_read_inode_block_full(inode, &bh,
-						     OCFS2_BH_IGNORE_CACHE);
+		if (args->fi_flags & OCFS2_FI_FLAG_FILECHECK_CHK)
+			status = ocfs2_filecheck_read_inode_block_full(inode,
+						&bh, OCFS2_BH_IGNORE_CACHE, 0);
+		else if (args->fi_flags & OCFS2_FI_FLAG_FILECHECK_FIX)
+			status = ocfs2_filecheck_read_inode_block_full(inode,
+						&bh, OCFS2_BH_IGNORE_CACHE, 1);
+		else
+			status = ocfs2_read_inode_block_full(inode,
+						&bh, OCFS2_BH_IGNORE_CACHE);
 	} else {
 		status = ocfs2_read_blocks_sync(osb, args->fi_blkno, 1, &bh);
 		/*
 		 * If buffer is in jbd, then its checksum may not have been
 		 * computed as yet.
 		 */
-		if (!status && !buffer_jbd(bh))
-			status = ocfs2_validate_inode_block(osb->sb, bh);
+		if (!status && !buffer_jbd(bh)) {
+			if (args->fi_flags & OCFS2_FI_FLAG_FILECHECK_CHK)
+				status = ocfs2_filecheck_validate_inode_block(
+								osb->sb, bh);
+			else if (args->fi_flags & OCFS2_FI_FLAG_FILECHECK_FIX)
+				status = ocfs2_filecheck_repair_inode_block(
+								osb->sb, bh);
+			else
+				status = ocfs2_validate_inode_block(
+								osb->sb, bh);
+		}
 	}
 	if (status < 0) {
 		mlog_errno(status);
@@ -531,6 +561,14 @@ static int ocfs2_read_locked_inode(struct inode *inode,
 
 	BUG_ON(args->fi_blkno != le64_to_cpu(fe->i_blkno));
 
+	if (buffer_dirty(bh)) {
+		status = ocfs2_write_block(osb, bh, INODE_CACHE(inode));
+		if (status < 0) {
+			mlog_errno(status);
+			goto bail;
+		}
+	}
+
 	status = 0;
 
 bail:
@@ -1385,6 +1423,152 @@ bail:
 	return rc;
 }
 
+static int ocfs2_filecheck_validate_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	trace_ocfs2_filecheck_validate_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	BUG_ON(!buffer_uptodate(bh));
+
+	if (!OCFS2_IS_VALID_DINODE(di)) {
+		mlog(ML_ERROR,
+			"Filecheck: invalid dinode #%llu: signature = %.*s\n",
+			(unsigned long long)bh->b_blocknr, 7, di->i_signature);
+		rc = -OCFS2_FILECHECK_ERR_INVALIDINO;
+		goto bail;
+	}
+
+	rc = ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check);
+	if (rc) {
+		mlog(ML_ERROR,
+			"Filecheck: checksum failed for dinode %llu\n",
+			(unsigned long long)bh->b_blocknr);
+		rc = -OCFS2_FILECHECK_ERR_BLOCKECC;
+		goto bail;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		mlog(ML_ERROR,
+			"Filecheck: invalid dinode #%llu: i_blkno is %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+		rc = -OCFS2_FILECHECK_ERR_BLOCKNO;
+		goto bail;
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		mlog(ML_ERROR,
+			"Filecheck: invalid dinode #%llu: OCFS2_VALID_FL not set\n",
+			(unsigned long long)bh->b_blocknr);
+		rc = -OCFS2_FILECHECK_ERR_VALIDFLAG;
+		goto bail;
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		mlog(ML_ERROR,
+			"Filecheck: invalid dinode #%llu: fs_generation is %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+		rc = -OCFS2_FILECHECK_ERR_GENERATION;
+		goto bail;
+	}
+
+bail:
+	return rc;
+}
+
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
+
+static int
+ocfs2_filecheck_read_inode_block_full(struct inode *inode,
+		struct buffer_head **bh, int flags, int type)
+{
+	int rc;
+	struct buffer_head *tmp = *bh;
+
+	if (!type) /* Check inode block */
+		rc = ocfs2_read_blocks(INODE_CACHE(inode),
+				OCFS2_I(inode)->ip_blkno,
+				1, &tmp, flags,
+				ocfs2_filecheck_validate_inode_block);
+	else /* Repair inode block */
+		rc = ocfs2_read_blocks(INODE_CACHE(inode),
+				OCFS2_I(inode)->ip_blkno,
+				1, &tmp, flags,
+				ocfs2_filecheck_repair_inode_block);
+
+	/* If ocfs2_read_blocks() got us a new bh, pass it up. */
+	if (!rc && !*bh)
+		*bh = tmp;
+
+	return rc;
+}
+
 int ocfs2_read_inode_block_full(struct inode *inode, struct buffer_head **bh,
 				int flags)
 {
fs/ocfs2/ocfs2_trace.h +2 -0
index 6cb019b..d9205e0 100644
@@ -1540,6 +1540,8 @@ DEFINE_OCFS2_ULL_INT_EVENT(ocfs2_read_locked_inode);
 DEFINE_OCFS2_INT_INT_EVENT(ocfs2_check_orphan_recovery_state);
 
 DEFINE_OCFS2_ULL_EVENT(ocfs2_validate_inode_block);
+DEFINE_OCFS2_ULL_EVENT(ocfs2_filecheck_validate_inode_block);
+DEFINE_OCFS2_ULL_EVENT(ocfs2_filecheck_repair_inode_block);
 
 TRACE_EVENT(ocfs2_inode_is_valid_to_delete,
 	TP_PROTO(void *task, void *dc_task, unsigned long long ino,
Signature · 1 line
2.1.2
#19 lore raw

Re: [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Hi Gang,

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.
Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Thanks,
Junxiao.

Quoted 61 lines · Gang He
fs/ocfs2/inode.c +61 -0
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
author
#20 lore raw

Re: [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Hello Junxiao,

See my comments inline.

Quoted 9 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

This online file check/fix feature is used to check/fix some light file meta block corruption, instead of turning a file system off and using fsck.ocfs2. e.g. meta ecc error, we really need not to use fsck.ocfs2.
of course, this feature does not replace fsck.ocfs2 and touch some complicated meta block problems, if there is some potential problem in some areas, we can discuss them one by one.

Quoted 5 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Here, if the inode block was corrupted, the file system is not able to load it into the memory. ocfs2_filecheck_repair_inode_block() will able to load it into the memory, since it try to fix these light-level problem before loading. if the fix is OK, the changed meta-block can pass the block-validate function and load into the memory as a inode object.
Since the file system is under a cluster environment, we have to use some existing function and code path to keep these block operation under a cluster lock.

Thanks
Gang

Quoted 65 lines

Thanks,
Junxiao.

Quoted 61 lines · Gang He
fs/ocfs2/inode.c +61 -0
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
#21 lore raw

Re: [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Quoted 8 lines · Gang He

Hello Junxiao,

See my comments inline.

Quoted 3 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

Quoted 7 lines
Quoted 6 lines

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

This online file check/fix feature is used to check/fix some light file meta block corruption, instead of turning a file system off and using fsck.ocfs2.

What's light meta block corruption? Do you have a case about it?

Quoted 11 lines

e.g. meta ecc error, we really need not to use fsck.ocfs2.
of course, this feature does not replace fsck.ocfs2 and touch some complicated meta block problems, if there is some potential problem in some areas, we can discuss them one by one.

Quoted 5 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Here, if the inode block was corrupted, the file system is not able to load it into the memory.

How do you know inode block corrupted? If bh for inode block is overwritten, i mean bh corrupted, the repair will corrupted a good inode block.

Thanks,
Junxiao.

Quoted 74 lines

ocfs2_filecheck_repair_inode_block() will able to load it into the memory, since it try to fix these light-level problem before loading. if the fix is OK, the changed meta-block can pass the block-validate function and load into the memory as a inode object.
Since the file system is under a cluster environment, we have to use some existing function and code path to keep these block operation under a cluster lock.

Thanks
Gang

Quoted 65 lines

Thanks,
Junxiao.

Quoted 61 lines · Gang He
fs/ocfs2/inode.c +61 -0
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
author
#22 lore raw

Re: [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Quoted 34 lines
Quoted 4 lines
Quoted 4 lines
Quoted 8 lines · Gang He

Hello Junxiao,

See my comments inline.

Quoted 3 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

Quoted 7 lines
Quoted 6 lines

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

This online file check/fix feature is used to check/fix some light file meta

block corruption, instead of turning a file system off and using fsck.ocfs2. What's light meta block corruption? Do you have a case about it?

Quoted 2 lines

e.g. meta ecc error, we really need not to use fsck.ocfs2. of course, this feature does not replace fsck.ocfs2 and touch some

complicated meta block problems, if there is some potential problem in some areas, we can discuss them one by one.

Quoted 8 lines
Quoted 8 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Quoted 5 lines
Quoted 1 line

Here, if the inode block was corrupted, the file system is not able to load

it into the memory.
How do you know inode block corrupted? If bh for inode block is overwritten, i mean bh corrupted, the repair will corrupted a good inode block.

You know, the meta block is only validated when the file system loads the block from disk to memory.
If the inode object is in the memory, we consider this inode block is OK.
If the inode is not loaded by the file system via the normal way, the file system will print a kernel error log to tell which ino is corrupted. we will use ocfs2_filecheck_repair_inode_block() function to fix the inode block before loading.

Thanks
Gang

Quoted 82 lines

Thanks,
Junxiao.

Quoted 1 line

ocfs2_filecheck_repair_inode_block() will able to load it into the memory,

since it try to fix these light-level problem before loading.

Quoted 1 line

if the fix is OK, the changed meta-block can pass the block-validate function

and load into the memory as a inode object.

Quoted 1 line

Since the file system is under a cluster environment, we have to use some

existing function and code path to keep these block operation under a cluster lock.

Quoted 71 lines

Thanks
Gang

Quoted 65 lines

Thanks,
Junxiao.

Quoted 61 lines · Gang He
fs/ocfs2/inode.c +61 -0
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
#23 lore raw

Re: [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Quoted 34 lines · Gang He
Quoted 34 lines
Quoted 4 lines
Quoted 4 lines
Quoted 8 lines · Gang He

Hello Junxiao,

See my comments inline.

Quoted 3 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

Quoted 7 lines
Quoted 6 lines

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

This online file check/fix feature is used to check/fix some light file meta

block corruption, instead of turning a file system off and using fsck.ocfs2. What's light meta block corruption? Do you have a case about it?

Quoted 2 lines

e.g. meta ecc error, we really need not to use fsck.ocfs2. of course, this feature does not replace fsck.ocfs2 and touch some

complicated meta block problems, if there is some potential problem in some areas, we can discuss them one by one.

Quoted 8 lines
Quoted 8 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

Here, if the inode block was corrupted, the file system is not able to load

it into the memory.
How do you know inode block corrupted? If bh for inode block is overwritten, i mean bh corrupted, the repair will corrupted a good inode block.

You know, the meta block is only validated when the file system loads the block from disk to memory. If the inode object is in the memory, we consider this inode block is OK.

This assuming is not true as there are always bugs. Bugs can make inode object in memory bad and corrupted the fs when repair the inode.

Thanks,
Junxiao.

Quoted 88 lines

If the inode is not loaded by the file system via the normal way, the file system will print a kernel error log to tell which ino is corrupted. we will use ocfs2_filecheck_repair_inode_block() function to fix the inode block before loading.

Thanks
Gang

Quoted 82 lines

Thanks,
Junxiao.

Quoted 1 line

ocfs2_filecheck_repair_inode_block() will able to load it into the memory,

since it try to fix these light-level problem before loading.

Quoted 1 line

if the fix is OK, the changed meta-block can pass the block-validate function

and load into the memory as a inode object.

Quoted 1 line

Since the file system is under a cluster environment, we have to use some

existing function and code path to keep these block operation under a cluster lock.

Quoted 71 lines

Thanks
Gang

Quoted 65 lines

Thanks,
Junxiao.

Quoted 61 lines · Gang He
fs/ocfs2/inode.c +61 -0
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
author
#24 lore raw

Re: [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Quoted 45 lines
Quoted 4 lines
Quoted 4 lines
Quoted 40 lines · Gang He
Quoted 40 lines
Quoted 4 lines
Quoted 4 lines
Quoted 8 lines · Gang He

Hello Junxiao,

See my comments inline.

Quoted 3 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

Quoted 7 lines
Quoted 6 lines

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

This online file check/fix feature is used to check/fix some light file meta

block corruption, instead of turning a file system off and using fsck.ocfs2. What's light meta block corruption? Do you have a case about it?

Quoted 2 lines

e.g. meta ecc error, we really need not to use fsck.ocfs2. of course, this feature does not replace fsck.ocfs2 and touch some

complicated meta block problems, if there is some potential problem in some areas, we can discuss them one by one.

Quoted 9 lines
Quoted 8 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Here, if the inode block was corrupted, the file system is not able to load

it into the memory.
How do you know inode block corrupted? If bh for inode block is overwritten, i mean bh corrupted, the repair will corrupted a good inode block.

Quoted 5 lines
Quoted 1 line

You know, the meta block is only validated when the file system loads the

block from disk to memory.

Quoted 1 line

If the inode object is in the memory, we consider this inode block is OK.

This assuming is not true as there are always bugs. Bugs can make inode object in memory bad and corrupted the fs when repair the inode.

The inode object in the memory has probably been corrupted in some cases, right.
but in these cases, online file check/fix feature considers this inode object is validated, will not do any further modification and exit. if the next corruption happens in case this inode is flushed into the disk, the bad thing is not made by online file check/fix code. Next (maybe next mount), the inode block is reload into the memory by the file system will fail, with a kernel error log printing, the online file check/fix probably can help to fix at this monent.

Thanks
Gang

Quoted 94 lines

Thanks,
Junxiao.

Quoted 1 line

If the inode is not loaded by the file system via the normal way, the file

system will print a kernel error log to tell which ino is corrupted.

Quoted 1 line

we will use ocfs2_filecheck_repair_inode_block() function to fix the inode

block before loading.

Quoted 14 lines

Thanks
Gang

Quoted 10 lines

Thanks,
Junxiao.

Quoted 1 line

ocfs2_filecheck_repair_inode_block() will able to load it into the memory,

since it try to fix these light-level problem before loading.

Quoted 1 line

if the fix is OK, the changed meta-block can pass the block-validate function

and load into the memory as a inode object.

Quoted 1 line

Since the file system is under a cluster environment, we have to use some

existing function and code path to keep these block operation under a

cluster

Quoted 72 lines
Quoted 72 lines

lock.

Quoted 71 lines

Thanks
Gang

Quoted 65 lines

Thanks,
Junxiao.

Quoted 61 lines · Gang He
fs/ocfs2/inode.c +61 -0
+static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
+			       struct buffer_head *bh)
+{
+	int rc;
+	int changed = 0;
+	struct ocfs2_dinode *di = (struct ocfs2_dinode *)bh->b_data;
+
+	rc = ocfs2_filecheck_validate_inode_block(sb, bh);
+	/* Can't fix invalid inode block */
+	if (!rc || rc == -OCFS2_FILECHECK_ERR_INVALIDINO)
+		return rc;
+
+	trace_ocfs2_filecheck_repair_inode_block(
+		(unsigned long long)bh->b_blocknr);
+
+	if (ocfs2_is_hard_readonly(OCFS2_SB(sb)) ||
+		ocfs2_is_soft_readonly(OCFS2_SB(sb))) {
+		mlog(ML_ERROR,
+			"Filecheck: try to repair dinode #%llu on readonly filesystem\n",
+			(unsigned long long)bh->b_blocknr);
+		return -OCFS2_FILECHECK_ERR_READONLY;
+	}
+
+	if (le64_to_cpu(di->i_blkno) != bh->b_blocknr) {
+		di->i_blkno = cpu_to_le64(bh->b_blocknr);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: i_blkno to %llu\n",
+			(unsigned long long)bh->b_blocknr,
+			(unsigned long long)le64_to_cpu(di->i_blkno));
+	}
+
+	if (!(di->i_flags & cpu_to_le32(OCFS2_VALID_FL))) {
+		di->i_flags |= cpu_to_le32(OCFS2_VALID_FL);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: OCFS2_VALID_FL is set\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	if (le32_to_cpu(di->i_fs_generation) !=
+	    OCFS2_SB(sb)->fs_generation) {
+		di->i_fs_generation = cpu_to_le32(OCFS2_SB(sb)->fs_generation);
+		changed = 1;
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: fs_generation to %u\n",
+			(unsigned long long)bh->b_blocknr,
+			le32_to_cpu(di->i_fs_generation));
+	}
+
+	if (changed ||
+		ocfs2_validate_meta_ecc(sb, bh->b_data, &di->i_check)) {
+		ocfs2_compute_meta_ecc(sb, bh->b_data, &di->i_check);
+		mark_buffer_dirty(bh);
+		mlog(ML_ERROR,
+			"Filecheck: reset dinode #%llu: compute meta ecc\n",
+			(unsigned long long)bh->b_blocknr);
+	}
+
+	return 0;
+}
#25 lore raw

Re: [Ocfs2-devel] [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

In reply to Junxiao Bi

Hi Junxiao,

Quoted 8 lines · Junxiao Bi

Hi Gang,

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

I agree that these are rather uncommon, we might even consider removing the VALID_FL fixup. I definitely don't think we're ready for anything more complicated than this though either. We kind of have to start somewhere too.

Quoted 5 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Your second point (and the last part of your 1st point) makes a good argument for why this shouldn't happen automatically. Some of these corruptions might require a human to look at the log and decide what to do. Especially as you point out, where we might not know where the source of the corruption is. And if the human can't figure it out, then it's probably time to unmount and fsck.

Thanks,

	--Mark

--
Mark Fasheh

#26 lore raw

Re: [Ocfs2-devel] [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Hi Mark,

Quoted 12 lines · Mark Fasheh

Hi Junxiao,

Quoted 8 lines · Junxiao Bi

Hi Gang,

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

Quoted 4 lines

I agree that these are rather uncommon, we might even consider removing the VALID_FL fixup. I definitely don't think we're ready for anything more complicated than this though either. We kind of have to start somewhere too.

Yes, the fix is too simple, and just a start, I think we'd better wait more useful parts done before merging it.

Quoted 7 lines
Quoted 6 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Quoted 6 lines

Your second point (and the last part of your 1st point) makes a good argument for why this shouldn't happen automatically. Some of these corruptions might require a human to look at the log and decide what to do. Especially as you point out, where we might not know where the source of the corruption is. And if the human can't figure it out, then it's probably time to unmount and fsck.

The point is that the fix way is wrong, just flush memory info to disk is not right. I agree online fsck is good feature, but need carefully design, it should not involve more corruptions. A rough idea from mine is that maybe we need some "frezee" mechanism in fs, which can hung all fs op and let fs stop at a safe area. After freeze fs, we can do some fsck work on it and these works should not cost lots time. What's your idea?

Thanks,
Junxiao.

Quoted 7 lines

Thanks,

	--Mark

--
Mark Fasheh

author
#27 lore raw

Re: [Ocfs2-devel] [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Hi Mark and Junxiao,

Quoted 16 lines
Quoted 1 line
Quoted 1 line

Hi Mark,

Quoted 12 lines · Mark Fasheh

Hi Junxiao,

Quoted 8 lines · Junxiao Bi

Hi Gang,

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

Quoted 6 lines
Quoted 4 lines

I agree that these are rather uncommon, we might even consider removing the VALID_FL fixup. I definitely don't think we're ready for anything more complicated than this though either. We kind of have to start somewhere too.

Yes, the fix is too simple, and just a start, I think we'd better wait more useful parts done before merging it.

I agree, just remark VALID_FL flag to fix this field is too simple, we should delay this field fix before I have a flawless solution, I will remove these lines code in the first version patches. In the future submits, I also hope your guys to help review the code carefully, shout out your comments when you doubt somewhere.

Quoted 13 lines
Quoted 7 lines
Quoted 6 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Quoted 6 lines

Your second point (and the last part of your 1st point) makes a good argument for why this shouldn't happen automatically. Some of these corruptions might require a human to look at the log and decide what to do. Especially as you point out, where we might not know where the source of the corruption is. And if the human can't figure it out, then it's probably time to unmount and fsck.

Quoted 6 lines

The point is that the fix way is wrong, just flush memory info to disk is not right. I agree online fsck is good feature, but need carefully design, it should not involve more corruptions. A rough idea from mine is that maybe we need some "frezee" mechanism in fs, which can hung all fs op and let fs stop at a safe area. After freeze fs, we can do some fsck work on it and these works should not cost lots time. What's your idea?

If we need to touch some global data structures, freezing fs can be considered when we can't get any way in case using the locks.
If we only handle some independent problem, we just need to lock the related data structures.

Quoted 11 lines

Thanks,
Junxiao.

Quoted 7 lines

Thanks,

	--Mark

--
Mark Fasheh

#28 lore raw

Re: [Ocfs2-devel] [PATCH v2 4/4] ocfs2: check/fix inode block for online file check

Quoted 23 lines · Gang He

Hi Mark and Junxiao,

Quoted 20 lines
Quoted 1 line
Quoted 1 line

Hi Mark,

Quoted 16 lines · Mark Fasheh

Hi Junxiao,

Quoted 8 lines · Junxiao Bi

Hi Gang,

This is not like a right patch.
First, online file check only checks inode's block number, valid flag, fs generation value, and meta ecc. I never see a real corruption happened only on this field, if these fields are corrupted, that means something bad may happen on other place. So fix this field may not help and even cause corruption more hard.

I agree that these are rather uncommon, we might even consider removing the VALID_FL fixup. I definitely don't think we're ready for anything more complicated than this though either. We kind of have to start somewhere too.

Quoted 5 lines
Quoted 2 lines

Yes, the fix is too simple, and just a start, I think we'd better wait more useful parts done before merging it.

I agree, just remark VALID_FL flag to fix this field is too simple, we should delay this field fix before I have a flawless solution, I will remove these lines code in the first version patches. In the future submits, I also hope your guys to help review the code carefully, shout out your comments when you doubt somewhere.

Sure.

Quoted 22 lines
Quoted 16 lines
Quoted 10 lines
Quoted 9 lines

Second, the repair way is wrong. In ocfs2_filecheck_repair_inode_block(), if these fields in disk don't match the ones in memory, the ones in memory are used to update the disk fields. The question is how do you know these field in memory are right(they may be the real corrupted ones)?

Quoted 6 lines

Your second point (and the last part of your 1st point) makes a good argument for why this shouldn't happen automatically. Some of these corruptions might require a human to look at the log and decide what to do. Especially as you point out, where we might not know where the source of the corruption is. And if the human can't figure it out, then it's probably time to unmount and fsck.

Quoted 6 lines

The point is that the fix way is wrong, just flush memory info to disk is not right. I agree online fsck is good feature, but need carefully design, it should not involve more corruptions. A rough idea from mine is that maybe we need some "frezee" mechanism in fs, which can hung all fs op and let fs stop at a safe area. After freeze fs, we can do some fsck work on it and these works should not cost lots time. What's your idea?

Quoted 3 lines

If we need to touch some global data structures, freezing fs can be considered when we can't get any way in case using the locks.
If we only handle some independent problem, we just need to lock the related data structures.

Hmm, I am not sure whether it's hard to decide an independent issue.

Thanks,
Junxiao.

Quoted 13 lines
Quoted 12 lines

Thanks,
Junxiao.

Quoted 7 lines

Thanks,

	--Mark

--
Mark Fasheh

#29 lore raw
In reply to Gang He

Hi Gang,

thank you for implementing this. I would like to understand this better on where and how it helps ... would you mind sharing couple examples(real scenarios).

Thanks,
--Srini

Quoted 33 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Compare with first version, I use strncasecmp instead of double strncmp functions. Second, update the source file contribution vendor.

Gang He (4):
ocfs2: export ocfs2_kset for online file check
ocfs2: sysfile interfaces for online file check
ocfs2: create/remove sysfile for online file check
ocfs2: check/fix inode block for online file check

  fs/ocfs2/Makefile      |   3 +-
  fs/ocfs2/filecheck.c   | 566 +++++++++++++++++++++++++++++++++++++++++++++++++
  fs/ocfs2/filecheck.h   |  48 +++++
  fs/ocfs2/inode.c       | 196 ++++++++++++++++-
  fs/ocfs2/inode.h       |   3 +
  fs/ocfs2/ocfs2_trace.h |   2 +
  fs/ocfs2/stackglue.c   |   3 +-
  fs/ocfs2/stackglue.h   |   2 +
  fs/ocfs2/super.c       |   5 +
  9 files changed, 820 insertions(+), 8 deletions(-)

create mode 100644 fs/ocfs2/filecheck.c
create mode 100644 fs/ocfs2/filecheck.h

author
#30 lore raw

Hello Srini,

There is a doc about ocfs2 online file check.

OCFS2 online file check
-----------------------

This document will describe OCFS2 online file check feature.

Introduction
============
OCFS2 is often used in high-availaibility systems. However, OCFS2 usually converts the filesystem to read-only on errors. This may not be necessary, since turning the filesystem read-only would affect other running processes as well, decreasing availability. Then, a mount option (errors=continue) was introduced, which would return the EIO to the calling process and terminate furhter processing so that the filesystem is not corrupted further. So,the filesystem is not converted to read-only, and the problematic file's inode number is reported in the kernel log so that the user can try to check/fix this file via online filecheck feature.

Scope
=====
This effort is to check/fix small issues which may hinder day-to-day operations of a cluster filesystem by turning the filesystem read-only. The scope of checking/fixing is at the file level, initially for regular files and eventually to all files (including system files) of the filesystem.

In case of directory to file links is incorrect, the directory inode is reported as erroneous.

This feature is not suited for extravagant checks which involve dependency of other components of the filesystem, such as but not limited to, checking if the bits for file blocks in the allocation has been set. In case of such an error, the offline fsck should/would be recommended.

Finally, such an operation/feature should not be automated lest the filesystem may end up with more damage than before the repair attempt. So, this has to be performed using user interaction and consent.

User interface
==============
When there are errors in the OCFS2 filesystem, they are usually accompanied by the inode number which caused the error. This inode number would be the input to check/fix the file.

There is a sysfs file for each OCFS2 file system mounting:

/sys/fs/ocfs2/<devname>/filecheck

Here, <devname> indicates the name of OCFS2 volumn device which has been already mounted. The file above would accept inode numbers. This could be used to communicate with kernel space, tell which file(inode number) will be checked or fixed. Currently, three operations are supported, which includes checking inode, fixing inode and setting the size of result record history.

1. If you want to know what error exactly happened to <inode> before fixing, do

  # echo "CHECK <inode>" > /sys/fs/ocfs2/<devname>/filecheck
  # cat /sys/fs/ocfs2/<devname>/filecheck

The output is like this:
INO TYPE DONE ERROR
39502 0 1 GENERATION

<INO> lists the inode numbers.
<TYPE> is what kind of operation you've done, 0 for inode check,1 for inode fix.
<DONE> indicates whether the operation has been finished.
<ERROR> says what kind of errors was found. For the details, please refer to the file linux/fs/ocfs2/filecheck.h.

2. If you determine to fix this inode, do

  # echo "FIX <inode>" > /sys/fs/ocfs2/<devname>/filecheck
  # cat /sys/fs/ocfs2/<devname>/filecheck

The output is like this:
INO TYPE DONE ERROR
39502 1 1 SUCCESS

This time, the <ERROR> column indicates whether this fix is successful or not.

3. The record cache is used to store the history of check/fix result. Its defalut size is 10, and can be adjust between the range of 10 ~ 100. You can adjust the size like this:

  # echo "SET <size>" > /sys/fs/ocfs2/<devname>/filecheck

Fixing stuff
============
On receivng the inode, the filesystem would read the inode and the file metadata. In case of errors, the filesystem would fix the errors and report the problems it fixed in the kernel log. As a precautionary measure, the inode must first be checked for errors before performing a final fix.

The inode and the fix history will be maintained temporarily in a small linked list buffer which would contain the last (N) inodes fixed/checked, along with the logs of what errors were reported/fixed.

Thanks
Gang

Quoted 46 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

thank you for implementing this. I would like to understand this better on where and how it helps ... would you mind sharing couple examples(real scenarios).

Thanks,
--Srini

Quoted 14 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Compare with first version, I use strncasecmp instead of double strncmp functions. Second, update the source file contribution vendor.

Quoted 8 lines

Gang He (4):
ocfs2: export ocfs2_kset for online file check
ocfs2: sysfile interfaces for online file check
ocfs2: create/remove sysfile for online file check
ocfs2: check/fix inode block for online file check

  fs/ocfs2/Makefile      |   3 +-
  fs/ocfs2/filecheck.c   | 566 

+++++++++++++++++++++++++++++++++++++++++++++++++

Quoted 11 lines
  fs/ocfs2/filecheck.h   |  48 +++++
  fs/ocfs2/inode.c       | 196 ++++++++++++++++-
  fs/ocfs2/inode.h       |   3 +
  fs/ocfs2/ocfs2_trace.h |   2 +
  fs/ocfs2/stackglue.c   |   3 +-
  fs/ocfs2/stackglue.h   |   2 +
  fs/ocfs2/super.c       |   5 +
  9 files changed, 820 insertions(+), 8 deletions(-)

create mode 100644 fs/ocfs2/filecheck.c
create mode 100644 fs/ocfs2/filecheck.h

#31 lore raw

Hi Gang,

thanks for pointing to explanation of the feature.

What I am curious about is ... what were the real cases that you came across prompted this change and how this change would help in that case.

Thanks,
--Srini

Quoted 148 lines · Gang He

Hello Srini,

There is a doc about ocfs2 online file check.

OCFS2 online file check
-----------------------

This document will describe OCFS2 online file check feature.

Introduction
============
OCFS2 is often used in high-availaibility systems. However, OCFS2 usually converts the filesystem to read-only on errors. This may not be necessary, since turning the filesystem read-only would affect other running processes as well, decreasing availability. Then, a mount option (errors=continue) was introduced, which would return the EIO to the calling process and terminate furhter processing so that the filesystem is not corrupted further. So,the filesystem is not converted to read-only, and the problematic file's inode number is reported in the kernel log so that the user can try to check/fix this file via online filecheck feature.

Scope
=====
This effort is to check/fix small issues which may hinder day-to-day operations of a cluster filesystem by turning the filesystem read-only. The scope of checking/fixing is at the file level, initially for regular files and eventually to all files (including system files) of the filesystem.

In case of directory to file links is incorrect, the directory inode is reported as erroneous.

This feature is not suited for extravagant checks which involve dependency of other components of the filesystem, such as but not limited to, checking if the bits for file blocks in the allocation has been set. In case of such an error, the offline fsck should/would be recommended.

Finally, such an operation/feature should not be automated lest the filesystem may end up with more damage than before the repair attempt. So, this has to be performed using user interaction and consent.

User interface
==============
When there are errors in the OCFS2 filesystem, they are usually accompanied by the inode number which caused the error. This inode number would be the input to check/fix the file.

There is a sysfs file for each OCFS2 file system mounting:

/sys/fs/ocfs2/<devname>/filecheck

Here, <devname> indicates the name of OCFS2 volumn device which has been already mounted. The file above would accept inode numbers. This could be used to communicate with kernel space, tell which file(inode number) will be checked or fixed. Currently, three operations are supported, which includes checking inode, fixing inode and setting the size of result record history.

1. If you want to know what error exactly happened to <inode> before fixing, do

   # echo "CHECK <inode>" > /sys/fs/ocfs2/<devname>/filecheck
   # cat /sys/fs/ocfs2/<devname>/filecheck

The output is like this:

   INO		TYPE		DONE		ERROR

39502 0 1 GENERATION

<INO> lists the inode numbers.
<TYPE> is what kind of operation you've done, 0 for inode check,1 for inode fix.
<DONE> indicates whether the operation has been finished.
<ERROR> says what kind of errors was found. For the details, please refer to the file linux/fs/ocfs2/filecheck.h.

2. If you determine to fix this inode, do

   # echo "FIX <inode>" > /sys/fs/ocfs2/<devname>/filecheck
   # cat /sys/fs/ocfs2/<devname>/filecheck

The output is like this:

   INO		TYPE		DONE		ERROR

39502 1 1 SUCCESS

This time, the <ERROR> column indicates whether this fix is successful or not.

3. The record cache is used to store the history of check/fix result. Its defalut size is 10, and can be adjust between the range of 10 ~ 100. You can adjust the size like this:

   # echo "SET <size>" > /sys/fs/ocfs2/<devname>/filecheck

Fixing stuff
============
On receivng the inode, the filesystem would read the inode and the file metadata. In case of errors, the filesystem would fix the errors and report the problems it fixed in the kernel log. As a precautionary measure, the inode must first be checked for errors before performing a final fix.

The inode and the fix history will be maintained temporarily in a small linked list buffer which would contain the last (N) inodes fixed/checked, along with the logs of what errors were reported/fixed.

Thanks
Gang

Quoted 45 lines

Hi Gang,

thank you for implementing this. I would like to understand this better on where and how it helps ... would you mind sharing couple examples(real scenarios).

Thanks,
--Srini

Quoted 14 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Compare with first version, I use strncasecmp instead of double strncmp functions. Second, update the source file contribution vendor.

Quoted 8 lines

Gang He (4):
ocfs2: export ocfs2_kset for online file check
ocfs2: sysfile interfaces for online file check
ocfs2: create/remove sysfile for online file check
ocfs2: check/fix inode block for online file check

   fs/ocfs2/Makefile      |   3 +-
   fs/ocfs2/filecheck.c   | 566

+++++++++++++++++++++++++++++++++++++++++++++++++

Quoted 11 lines
   fs/ocfs2/filecheck.h   |  48 +++++
   fs/ocfs2/inode.c       | 196 ++++++++++++++++-
   fs/ocfs2/inode.h       |   3 +
   fs/ocfs2/ocfs2_trace.h |   2 +
   fs/ocfs2/stackglue.c   |   3 +-
   fs/ocfs2/stackglue.h   |   2 +
   fs/ocfs2/super.c       |   5 +
   9 files changed, 820 insertions(+), 8 deletions(-)

create mode 100644 fs/ocfs2/filecheck.c
create mode 100644 fs/ocfs2/filecheck.h

author
#32 lore raw

Hello Srini,

The real cases are that we try to fix some independent issues without turning the file system off-line (error=continue was introduced).
You know, the online file check feature is used for fixing some independent or light meta-data block corruption, e.g. inode block, file extent block, dir entry block, etc. These corruptions are usually like checksum error, blk number inconsistency, etc.

Thanks
Gang

Signature · 1 line
<The HTML signature 'New Signature' does not contain any text>
Quoted 178 lines
Quoted 1 line
Quoted 1 line

Hi Gang,

thanks for pointing to explanation of the feature.

What I am curious about is ... what were the real cases that you came across prompted this change and how this change would help in that case.

Thanks,
--Srini

Quoted 7 lines · Gang He

Hello Srini,

There is a doc about ocfs2 online file check.

OCFS2 online file check
-----------------------

Quoted 6 lines

This document will describe OCFS2 online file check feature.

Introduction
============
OCFS2 is often used in high-availaibility systems. However, OCFS2 usually converts the filesystem to read-only on errors. This may not be necessary,

since

Quoted 1 line

turning the filesystem read-only would affect other running processes as

well,

Quoted 1 line

decreasing availability. Then, a mount option (errors=continue) was

introduced,

Quoted 2 lines

which would return the EIO to the calling process and terminate furhter processing so that the filesystem is not corrupted further. So,the

filesystem is

Quoted 1 line

not converted to read-only, and the problematic file's inode number is

reported

Quoted 6 lines

in the kernel log so that the user can try to check/fix this file via online filecheck feature.

Scope
=====
This effort is to check/fix small issues which may hinder day-to-day

operations

Quoted 2 lines

of a cluster filesystem by turning the filesystem read-only. The scope of checking/fixing is at the file level, initially for regular files and

eventually

Quoted 6 lines

to all files (including system files) of the filesystem.

In case of directory to file links is incorrect, the directory inode is reported as erroneous.

This feature is not suited for extravagant checks which involve dependency

of

Quoted 1 line

other components of the filesystem, such as but not limited to, checking if

the

Quoted 1 line

bits for file blocks in the allocation has been set. In case of such an

error,

Quoted 3 lines

the offline fsck should/would be recommended.

Finally, such an operation/feature should not be automated lest the

filesystem

Quoted 9 lines

may end up with more damage than before the repair attempt. So, this has to be performed using user interaction and consent.

User interface
==============
When there are errors in the OCFS2 filesystem, they are usually accompanied by the inode number which caused the error. This inode number would be the input to check/fix the file.

Quoted 5 lines

There is a sysfs file for each OCFS2 file system mounting:

/sys/fs/ocfs2/<devname>/filecheck

Here, <devname> indicates the name of OCFS2 volumn device which has been

already

Quoted 2 lines

mounted. The file above would accept inode numbers. This could be used to communicate with kernel space, tell which file(inode number) will be checked

or

Quoted 4 lines

fixed. Currently, three operations are supported, which includes checking inode, fixing inode and setting the size of result record history.

1. If you want to know what error exactly happened to <inode> before fixing,

do

Quoted 10 lines
   # echo "CHECK <inode>" > /sys/fs/ocfs2/<devname>/filecheck
   # cat /sys/fs/ocfs2/<devname>/filecheck

The output is like this:

   INO		TYPE		DONE		ERROR

39502 0 1 GENERATION

<INO> lists the inode numbers.
<TYPE> is what kind of operation you've done, 0 for inode check,1 for inode

fix.

Quoted 2 lines

<DONE> indicates whether the operation has been finished.
<ERROR> says what kind of errors was found. For the details, please refer to

the

Quoted 17 lines

file linux/fs/ocfs2/filecheck.h.

2. If you determine to fix this inode, do

   # echo "FIX <inode>" > /sys/fs/ocfs2/<devname>/filecheck
   # cat /sys/fs/ocfs2/<devname>/filecheck

The output is like this:

   INO		TYPE		DONE		ERROR

39502 1 1 SUCCESS

This time, the <ERROR> column indicates whether this fix is successful or not.

3. The record cache is used to store the history of check/fix result. Its defalut size is 10, and can be adjust between the range of 10 ~ 100. You can adjust the size like this:

Quoted 7 lines
   # echo "SET <size>" > /sys/fs/ocfs2/<devname>/filecheck

Fixing stuff
============
On receivng the inode, the filesystem would read the inode and the file metadata. In case of errors, the filesystem would fix the errors and report the problems it fixed in the kernel log. As a precautionary

measure,

Quoted 55 lines

the inode must first be checked for errors before performing a final fix.

The inode and the fix history will be maintained temporarily in a small linked list buffer which would contain the last (N) inodes fixed/checked, along with the logs of what errors were reported/fixed.

Thanks
Gang

Quoted 45 lines

Hi Gang,

thank you for implementing this. I would like to understand this better on where and how it helps ... would you mind sharing couple examples(real scenarios).

Thanks,
--Srini

Quoted 14 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Compare with first version, I use strncasecmp instead of double strncmp functions. Second, update the source file contribution vendor.

Quoted 8 lines

Gang He (4):
ocfs2: export ocfs2_kset for online file check
ocfs2: sysfile interfaces for online file check
ocfs2: create/remove sysfile for online file check
ocfs2: check/fix inode block for online file check

   fs/ocfs2/Makefile      |   3 +-
   fs/ocfs2/filecheck.c   | 566

+++++++++++++++++++++++++++++++++++++++++++++++++

Quoted 11 lines
   fs/ocfs2/filecheck.h   |  48 +++++
   fs/ocfs2/inode.c       | 196 ++++++++++++++++-
   fs/ocfs2/inode.h       |   3 +
   fs/ocfs2/ocfs2_trace.h |   2 +
   fs/ocfs2/stackglue.c   |   3 +-
   fs/ocfs2/stackglue.h   |   2 +
   fs/ocfs2/super.c       |   5 +
   9 files changed, 820 insertions(+), 8 deletions(-)

create mode 100644 fs/ocfs2/filecheck.c
create mode 100644 fs/ocfs2/filecheck.h

#33 lore raw
In reply to Gang He
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

...or returning back results.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
author
#34 lore raw

Hello Pavel,

Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Quoted 3 lines

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

Usually, the corrupted file (inode) should be rarely encountered for OCFS2 file system, then lots of commands are executed via this interface with high performance is not expected by us. Second, after online file check is added, we also plan to add a mount option "error=fix", that means the file system can fix these errors automatically without a manual command triggering.

Thanks
Gang

Quoted 8 lines

...or returning back results.
Pavel

Signature · 4 lines
(english) http://www.livejournal.com/~pavelmachek 
(cesky, pictures) 
http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html 
.
#35 lore raw
Quoted 18 lines · Gang He

Hello Pavel,

Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Quoted 7 lines
Quoted 3 lines

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

Usually, the corrupted file (inode) should be rarely encountered for OCFS2 file system, then lots of commands are executed via this interface with high performance is not expected by us. Second, after online file check is added, we also plan to add a mount option "error=fix", that means the file system can fix these errors automatically without a manual command triggering.

It's not a "performance" issue, it's a "sysfs files only have one value" type thing. Have two files, "inode_fix" and "inode_check" and then just write the inode into them, no need to have a "verb <inode>" type parser.

thanks,

greg k-h

author
#36 lore raw

Hi Greg,

Quoted 27 lines
Quoted 1 line
Quoted 1 line
Quoted 18 lines · Gang He

Hello Pavel,

Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Quoted 4 lines
Quoted 3 lines

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

Usually, the corrupted file (inode) should be rarely encountered for OCFS2

file system, then

Quoted 1 line

lots of commands are executed via this interface with high performance is

not expected by us.

Quoted 8 lines
Quoted 1 line

Second, after online file check is added, we also plan to add a mount option

"error=fix", that means

Quoted 1 line

the file system can fix these errors automatically without a manual command

triggering.

It's not a "performance" issue, it's a "sysfs files only have one value" type thing. Have two files, "inode_fix" and "inode_check" and then just write the inode into them, no need to have a "verb <inode>" type parser.

Current, we have three functional items "check, fix and set", in the future, maybe we can add more item. Then, for each functional item, we need to create a sys file and add related code (actual some code is duplicated), I prefer to one sys file to handle multiple sub-commands.

Thanks
Gang

Quoted 4 lines

thanks,

greg k-h

#37 lore raw
Quoted 33 lines · Gang He

Hi Greg,

Quoted 30 lines
Quoted 1 line
Quoted 1 line
Quoted 18 lines · Gang He

Hello Pavel,

Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Quoted 4 lines
Quoted 3 lines

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

Usually, the corrupted file (inode) should be rarely encountered for OCFS2

file system, then

Quoted 1 line

lots of commands are executed via this interface with high performance is

not expected by us.

Quoted 1 line

Second, after online file check is added, we also plan to add a mount option

"error=fix", that means

Quoted 1 line

the file system can fix these errors automatically without a manual command

Quoted 8 lines
Quoted 5 lines

triggering.

It's not a "performance" issue, it's a "sysfs files only have one value" type thing. Have two files, "inode_fix" and "inode_check" and then just write the inode into them, no need to have a "verb <inode>" type parser.

Current, we have three functional items "check, fix and set", in the future, maybe we can add more item. Then, for each functional item, we need to create a sys file and add related code (actual some code is duplicated), I prefer to one sys file to handle multiple sub-commands.

And we prefer not to have your code in tree.

Please design some reasonable interface. Abusing sysfs for this is not right.
Pavel

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#38 lore raw
In reply to Gang He
Quoted 33 lines · Gang He

Hi Greg,

Quoted 30 lines
Quoted 1 line
Quoted 1 line
Quoted 18 lines · Gang He

Hello Pavel,

Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Quoted 4 lines
Quoted 3 lines

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

Usually, the corrupted file (inode) should be rarely encountered for OCFS2

file system, then

Quoted 1 line

lots of commands are executed via this interface with high performance is

not expected by us.

Quoted 1 line

Second, after online file check is added, we also plan to add a mount option

"error=fix", that means

Quoted 1 line

the file system can fix these errors automatically without a manual command

Quoted 8 lines
Quoted 5 lines

triggering.

It's not a "performance" issue, it's a "sysfs files only have one value" type thing. Have two files, "inode_fix" and "inode_check" and then just write the inode into them, no need to have a "verb <inode>" type parser.

Current, we have three functional items "check, fix and set", in the future, maybe we can add more item. Then, for each functional item, we need to create a sys file and add related code (actual some code is duplicated), I prefer to one sys file to handle multiple sub-commands.

No, sorry, that is not how sysfs works. Please use individual files, again, sysfs is "one value per file" you should never have to write a "parser" for a sysfs file either reading, or writing to it.

If you need additional things in the future, great, add new sysfs files, that makes it the easiest way for your userspace tools to be able to determine if that feature is present in the kernel or not, it does not have to write a command that it doesn't know if the kernel can handle or not.

thanks,

greg k-h

author
#39 lore raw

Hello Greg and Pavel,

Sorry, there was a misunderstand, I was not aware that there were some design constraints for sysfs interfaces. I will review and modify this portion code.

Thanks
Gang

Quoted 61 lines
Quoted 1 line
Quoted 1 line
Quoted 31 lines · Gang He

Hi Greg,

Quoted 28 lines
Quoted 1 line
Quoted 1 line
Quoted 18 lines · Gang He

Hello Pavel,

Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 11 lines · Gang He

When there are errors in the ocfs2 filesystem,
they are usually accompanied by the inode number which caused the error. This inode number would be the input to fixing the file.
One of these options could be considered:
A file in the sys filesytem which would accept inode numbers. This could be used to communication back what has to be fixed or is fixed. You could write:
$# echo "CHECK <inode>" > /sys/fs/ocfs2/devname/filecheck
or
$# echo "FIX <inode>" > /sys/fs/ocfs2/devname/filecheck

Quoted 4 lines
Quoted 3 lines

Are you sure this is reasonable interface? I mean.... sysfs is supposed to be one value per file. And I don't think its suitable for running commands.

Usually, the corrupted file (inode) should be rarely encountered for OCFS2

file system, then

Quoted 1 line

lots of commands are executed via this interface with high performance is

not expected by us.

Quoted 1 line

Second, after online file check is added, we also plan to add a mount

option

Quoted 9 lines
Quoted 8 lines

"error=fix", that means

Quoted 1 line

the file system can fix these errors automatically without a manual command

triggering.

It's not a "performance" issue, it's a "sysfs files only have one value" type thing. Have two files, "inode_fix" and "inode_check" and then just write the inode into them, no need to have a "verb <inode>" type parser.

Current, we have three functional items "check, fix and set", in the future,

maybe we can add more item.

Quoted 1 line

Then, for each functional item, we need to create a sys file and add related

code (actual some code is duplicated),

Quoted 1 line

I prefer to one sys file to handle multiple sub-commands.

No, sorry, that is not how sysfs works. Please use individual files, again, sysfs is "one value per file" you should never have to write a "parser" for a sysfs file either reading, or writing to it.

If you need additional things in the future, great, add new sysfs files, that makes it the easiest way for your userspace tools to be able to determine if that feature is present in the kernel or not, it does not have to write a command that it doesn't know if the kernel can handle or not.

thanks,

greg k-h