Re: [patch 24/26] mount options: fix tmpfs

From: Hugh Dickins
Date: Mon Jan 28 2008 - 01:10:18 EST

On Thu, 24 Jan 2008, Miklos Szeredi wrote:
>

Thanks Miklos, that's a welcome enhancement, nicely done. I've only
noticed one thing wrong (MPOL_PREFERRED shown as "default"); but thought
shmem_config didn't add much value - I'd rather avoid those syntactic
changes to unchanged code; and several tmpfs defaults being relative
(e.g. to totalram_pages, or to mounter's fsuid), I ended up preferring
to do real tests in shmem_show_options.

Thus, for example, if memory is hotplugged in or out later, what started
out as an unspecified size option will then get shown as explicit size.
(I did think for a while that I wanted to show explicit size in all
cases; but it looked pretty silly on udev.) I think that's the correct
behaviour, that otherwise would be misleading; but I may be looking at
this the wrong way round, what's your view?

If you agree with the version below, please take it into your collection
and insert your Signed-off-by. I should admit, I've not yet tested how
the NUMA policies look: you'll hear from me again tomorrow morning if
those turn out to wrong.

Hugh

From: Miklos Szeredi <mszeredi@xxxxxxx>

Add .show_options super operation to tmpfs.

Signed-off-by: Hugh Dickins <hugh@xxxxxxxxxxx>
---

include/linux/shmem_fs.h | 5
mm/shmem.c | 189 ++++++++++++++++++++++++-------------
2 files changed, 129 insertions(+), 65 deletions(-)

--- 2.6.24-rc8-mm1/include/linux/shmem_fs.h 2006-11-29 21:57:37.000000000 +0000
+++ linux/include/linux/shmem_fs.h 2008-01-27 22:42:52.000000000 +0000
@@ -30,9 +30,12 @@ struct shmem_sb_info {
unsigned long free_blocks; /* How many are left for allocation */
unsigned long max_inodes; /* How many inodes are allowed */
unsigned long free_inodes; /* How many are left for allocation */
+ spinlock_t stat_lock; /* Serialize shmem_sb_info changes */
+ uid_t uid; /* Mount uid for root directory */
+ gid_t gid; /* Mount gid for root directory */
+ mode_t mode; /* Mount mode for root directory */
int policy; /* Default NUMA memory alloc policy */
nodemask_t policy_nodes; /* nodemask for preferred and bind */
- spinlock_t stat_lock;
};

static inline struct shmem_inode_info *SHMEM_I(struct inode *inode)
--- 2.6.24-rc8-mm1/mm/shmem.c 2008-01-17 16:51:21.000000000 +0000
+++ linux/mm/shmem.c 2008-01-27 23:41:31.000000000 +0000
@@ -49,6 +49,7 @@
#include <linux/ctype.h>
#include <linux/migrate.h>
#include <linux/highmem.h>
+#include <linux/seq_file.h>

#include <asm/uaccess.h>
#include <asm/div64.h>
@@ -1068,7 +1069,8 @@ redirty:
}

#ifdef CONFIG_NUMA
-static inline int shmem_parse_mpol(char *value, int *policy, nodemask_t *policy_nodes)
+#ifdef CONFIG_TMPFS
+static int shmem_parse_mpol(char *value, int *policy, nodemask_t *policy_nodes)
{
char *nodelist = strchr(value, ':');
int err = 1;
@@ -1117,6 +1119,42 @@ out:
return err;
}

+static void shmem_show_mpol(struct seq_file *seq, int policy,
+ const nodemask_t policy_nodes)
+{
+ char *policy_string;
+
+ switch (policy) {
+ case MPOL_PREFERRED:
+ policy_string = "prefer";
+ break;
+ case MPOL_BIND:
+ policy_string = "bind";
+ break;
+ case MPOL_INTERLEAVE:
+ policy_string = "interleave";
+ break;
+ default:
+ /* MPOL_DEFAULT */
+ return;
+ }
+
+ seq_printf(seq, ",mpol=%s", policy_string);
+
+ if (policy != MPOL_INTERLEAVE ||
+ !nodes_equal(policy_nodes, node_states[N_HIGH_MEMORY])) {
+ char buffer[64];
+ int len;
+
+ len = nodelist_scnprintf(buffer, sizeof(buffer), policy_nodes);
+ if (len < sizeof(buffer))
+ seq_printf(seq, ":%s", buffer);
+ else
+ seq_printf(seq, ":?");
+ }
+}
+#endif /* CONFIG_TMPFS */
+
static struct page *shmem_swapin(swp_entry_t entry, gfp_t gfp,
struct shmem_inode_info *info, unsigned long idx)
{
@@ -1148,13 +1186,20 @@ static struct page *shmem_alloc_page(gfp
mpol_free(pvma.vm_policy);
return page;
}
-#else
+#else /* !CONFIG_NUMA */
+#ifdef CONFIG_TMPFS
static inline int shmem_parse_mpol(char *value, int *policy,
nodemask_t *policy_nodes)
{
return 1;
}

+static inline void shmem_show_mpol(struct seq_file *seq, int policy,
+ const nodemask_t policy_nodes)
+{
+}
+#endif /* CONFIG_TMPFS */
+
static inline struct page *shmem_swapin(swp_entry_t entry, gfp_t gfp,
struct shmem_inode_info *info, unsigned long idx)
{
@@ -1166,7 +1211,7 @@ static inline struct page *shmem_alloc_p
{
return alloc_page(gfp);
}
-#endif
+#endif /* CONFIG_NUMA */

/*
* shmem_getpage - either get the page from swap or allocate a new one
@@ -2077,9 +2122,8 @@ static const struct export_operations sh
.fh_to_dentry = shmem_fh_to_dentry,
};

-static int shmem_parse_options(char *options, int *mode, uid_t *uid,
- gid_t *gid, unsigned long *blocks, unsigned long *inodes,
- int *policy, nodemask_t *policy_nodes)
+static int shmem_parse_options(char *options, struct shmem_sb_info *sbinfo,
+ bool remount)
{
char *this_char, *value, *rest;

@@ -2122,35 +2166,37 @@ static int shmem_parse_options(char *opt
}
if (*rest)
goto bad_val;
- *blocks = DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
+ sbinfo->max_blocks =
+ DIV_ROUND_UP(size, PAGE_CACHE_SIZE);
} else if (!strcmp(this_char,"nr_blocks")) {
- *blocks = memparse(value,&rest);
+ sbinfo->max_blocks = memparse(value, &rest);
if (*rest)
goto bad_val;
} else if (!strcmp(this_char,"nr_inodes")) {
- *inodes = memparse(value,&rest);
+ sbinfo->max_inodes = memparse(value, &rest);
if (*rest)
goto bad_val;
} else if (!strcmp(this_char,"mode")) {
- if (!mode)
+ if (remount)
continue;
- *mode = simple_strtoul(value,&rest,8);
+ sbinfo->mode = simple_strtoul(value, &rest, 8) & 07777;
if (*rest)
goto bad_val;
} else if (!strcmp(this_char,"uid")) {
- if (!uid)
+ if (remount)
continue;
- *uid = simple_strtoul(value,&rest,0);
+ sbinfo->uid = simple_strtoul(value, &rest, 0);
if (*rest)
goto bad_val;
} else if (!strcmp(this_char,"gid")) {
- if (!gid)
+ if (remount)
continue;
- *gid = simple_strtoul(value,&rest,0);
+ sbinfo->gid = simple_strtoul(value, &rest, 0);
if (*rest)
goto bad_val;
} else if (!strcmp(this_char,"mpol")) {
- if (shmem_parse_mpol(value,policy,policy_nodes))
+ if (shmem_parse_mpol(value, &sbinfo->policy,
+ &sbinfo->policy_nodes))
goto bad_val;
} else {
printk(KERN_ERR "tmpfs: Bad mount option %s\n",
@@ -2170,24 +2216,20 @@ bad_val:
static int shmem_remount_fs(struct super_block *sb, int *flags, char *data)
{
struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
- unsigned long max_blocks = sbinfo->max_blocks;
- unsigned long max_inodes = sbinfo->max_inodes;
- int policy = sbinfo->policy;
- nodemask_t policy_nodes = sbinfo->policy_nodes;
+ struct shmem_sb_info config = *sbinfo;
unsigned long blocks;
unsigned long inodes;
int error = -EINVAL;

- if (shmem_parse_options(data, NULL, NULL, NULL, &max_blocks,
- &max_inodes, &policy, &policy_nodes))
+ if (shmem_parse_options(data, &config, true))
return error;

spin_lock(&sbinfo->stat_lock);
blocks = sbinfo->max_blocks - sbinfo->free_blocks;
inodes = sbinfo->max_inodes - sbinfo->free_inodes;
- if (max_blocks < blocks)
+ if (config.max_blocks < blocks)
goto out;
- if (max_inodes < inodes)
+ if (config.max_inodes < inodes)
goto out;
/*
* Those tests also disallow limited->unlimited while any are in
@@ -2195,23 +2237,43 @@ static int shmem_remount_fs(struct super
* but we must separately disallow unlimited->limited, because
* in that case we have no record of how much is already in use.
*/
- if (max_blocks && !sbinfo->max_blocks)
+ if (config.max_blocks && !sbinfo->max_blocks)
goto out;
- if (max_inodes && !sbinfo->max_inodes)
+ if (config.max_inodes && !sbinfo->max_inodes)
goto out;

error = 0;
- sbinfo->max_blocks = max_blocks;
- sbinfo->free_blocks = max_blocks - blocks;
- sbinfo->max_inodes = max_inodes;
- sbinfo->free_inodes = max_inodes - inodes;
- sbinfo->policy = policy;
- sbinfo->policy_nodes = policy_nodes;
+ sbinfo->max_blocks = config.max_blocks;
+ sbinfo->free_blocks = config.max_blocks - blocks;
+ sbinfo->max_inodes = config.max_inodes;
+ sbinfo->free_inodes = config.max_inodes - inodes;
+ sbinfo->policy = config.policy;
+ sbinfo->policy_nodes = config.policy_nodes;
out:
spin_unlock(&sbinfo->stat_lock);
return error;
}
-#endif
+
+static int shmem_show_options(struct seq_file *seq, struct vfsmount *vfs)
+{
+ struct shmem_sb_info *sbinfo = SHMEM_SB(vfs->mnt_sb);
+
+ if (sbinfo->max_blocks != totalram_pages / 2)
+ seq_printf(seq, ",size=%luk",
+ sbinfo->max_blocks << (PAGE_CACHE_SHIFT - 10));
+ if (sbinfo->max_inodes !=
+ min(totalram_pages - totalhigh_pages, totalram_pages / 2))
+ seq_printf(seq, ",nr_inodes=%lu", sbinfo->max_inodes);
+ if (sbinfo->mode != (S_IRWXUGO | S_ISVTX))
+ seq_printf(seq, ",mode=%03o", sbinfo->mode);
+ if (sbinfo->uid != 0)
+ seq_printf(seq, ",uid=%u", sbinfo->uid);
+ if (sbinfo->gid != 0)
+ seq_printf(seq, ",gid=%u", sbinfo->gid);
+ shmem_show_mpol(seq, sbinfo->policy, sbinfo->policy_nodes);
+ return 0;
+}
+#endif /* CONFIG_TMPFS */

static void shmem_put_super(struct super_block *sb)
{
@@ -2224,15 +2286,23 @@ static int shmem_fill_super(struct super
{
struct inode *inode;
struct dentry *root;
- int mode = S_IRWXUGO | S_ISVTX;
- uid_t uid = current->fsuid;
- gid_t gid = current->fsgid;
- int err = -ENOMEM;
struct shmem_sb_info *sbinfo;
- unsigned long blocks = 0;
- unsigned long inodes = 0;
- int policy = MPOL_DEFAULT;
- nodemask_t policy_nodes = node_states[N_HIGH_MEMORY];
+ int err = -ENOMEM;
+
+ /* Round up to L1_CACHE_BYTES to resist false sharing */
+ sbinfo = kmalloc(max((int)sizeof(struct shmem_sb_info),
+ L1_CACHE_BYTES), GFP_KERNEL);
+ if (!sbinfo)
+ return -ENOMEM;
+
+ sbinfo->max_blocks = 0;
+ sbinfo->max_inodes = 0;
+ sbinfo->mode = S_IRWXUGO | S_ISVTX;
+ sbinfo->uid = current->fsuid;
+ sbinfo->gid = current->fsgid;
+ sbinfo->policy = MPOL_DEFAULT;
+ sbinfo->policy_nodes = node_states[N_HIGH_MEMORY];
+ sb->s_fs_info = sbinfo;

#ifdef CONFIG_TMPFS
/*
@@ -2241,34 +2311,24 @@ static int shmem_fill_super(struct super
* but the internal instance is left unlimited.
*/
if (!(sb->s_flags & MS_NOUSER)) {
- blocks = totalram_pages / 2;
- inodes = totalram_pages - totalhigh_pages;
- if (inodes > blocks)
- inodes = blocks;
- if (shmem_parse_options(data, &mode, &uid, &gid, &blocks,
- &inodes, &policy, &policy_nodes))
- return -EINVAL;
+ sbinfo->max_blocks = totalram_pages / 2;
+ sbinfo->max_inodes = totalram_pages - totalhigh_pages;
+ if (sbinfo->max_inodes > sbinfo->max_blocks)
+ sbinfo->max_inodes = sbinfo->max_blocks;
+ if (shmem_parse_options(data, sbinfo, false)) {
+ err = -EINVAL;
+ goto failed;
+ }
}
sb->s_export_op = &shmem_export_ops;
#else
sb->s_flags |= MS_NOUSER;
#endif

- /* Round up to L1_CACHE_BYTES to resist false sharing */
- sbinfo = kmalloc(max((int)sizeof(struct shmem_sb_info),
- L1_CACHE_BYTES), GFP_KERNEL);
- if (!sbinfo)
- return -ENOMEM;
-
spin_lock_init(&sbinfo->stat_lock);
- sbinfo->max_blocks = blocks;
- sbinfo->free_blocks = blocks;
- sbinfo->max_inodes = inodes;
- sbinfo->free_inodes = inodes;
- sbinfo->policy = policy;
- sbinfo->policy_nodes = policy_nodes;
+ sbinfo->free_blocks = sbinfo->max_blocks;
+ sbinfo->free_inodes = sbinfo->max_inodes;

- sb->s_fs_info = sbinfo;
sb->s_maxbytes = SHMEM_MAX_BYTES;
sb->s_blocksize = PAGE_CACHE_SIZE;
sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
@@ -2280,11 +2340,11 @@ static int shmem_fill_super(struct super
sb->s_flags |= MS_POSIXACL;
#endif

- inode = shmem_get_inode(sb, S_IFDIR | mode, 0);
+ inode = shmem_get_inode(sb, S_IFDIR | sbinfo->mode, 0);
if (!inode)
goto failed;
- inode->i_uid = uid;
- inode->i_gid = gid;
+ inode->i_uid = sbinfo->uid;
+ inode->i_gid = sbinfo->gid;
root = d_alloc_root(inode);
if (!root)
goto failed_iput;
@@ -2420,6 +2480,7 @@ static const struct super_operations shm
#ifdef CONFIG_TMPFS
.statfs = shmem_statfs,
.remount_fs = shmem_remount_fs,
+ .show_options = shmem_show_options,
#endif
.delete_inode = shmem_delete_inode,
.drop_inode = generic_delete_inode,
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/