hollalinux
/
linux-sg2042
mirror of https://github.com/sophgo/linux-riscv.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Orangefs: fix dir_emit code in pvfs2_readdir. 

Al Viro glanced at readdir and surmised that getdents
would misbehave the way it was written... and sure enough.

Signed-off-by: Mike Marshall <hubcap@omnibond.com>
This commit is contained in:
Mike Marshall 2015-09-23 16:48:40 -04:00
parent 4d1c44043b
commit 88309aae3d
2 changed files with 50 additions and 82 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

131
fs/orangefs/dir.c
View File

@ -95,26 +95,16 @@ static void readdir_handle_dtor(struct pvfs2_bufmap *bufmap,
/* /*
* Read directory entries from an instance of an open directory. * Read directory entries from an instance of an open directory.
*
* \note This routine was converted for the readdir to iterate change
* in "struct file_operations". "converted" mostly amounts to
* changing occurrences of "readdir" and "filldir" in the
* comments to "iterate" and "dir_emit". Also filldir calls
* were changed to dir_emit calls.
*
* \param dir_emit callback function called for each entry read.
*
* \retval 0 when directory has been completely traversed
* \retval >0 if we don't call dir_emit for all entries
*
* \note If the dir_emit call-back returns non-zero, then iterate should
* assume that it has had enough, and should return as well.
*/ */
static int pvfs2_readdir(struct file *file, struct dir_context *ctx) static int pvfs2_readdir(struct file *file, struct dir_context *ctx)
{ {
struct pvfs2_bufmap *bufmap = NULL; struct pvfs2_bufmap *bufmap = NULL;
int ret = 0; int ret = 0;
int buffer_index; int buffer_index;
/*
* ptoken supports Orangefs' distributed directory logic, added
* in 2.9.2.
*/
__u64 *ptoken = file->private_data; __u64 *ptoken = file->private_data;
__u64 pos = 0; __u64 pos = 0;
ino_t ino = 0; ino_t ino = 0;
@ -129,11 +119,11 @@ static int pvfs2_readdir(struct file *file, struct dir_context *ctx)
char *current_entry = NULL; char *current_entry = NULL;
long bytes_decoded; long bytes_decoded;
gossip_ldebug(GOSSIP_DIR_DEBUG, gossip_debug(GOSSIP_DIR_DEBUG,
"%s: ctx->pos:%lld, token = %llu\n", "%s: ctx->pos:%lld, ptoken = %llu\n",
__func__, __func__,
lld(ctx->pos), lld(ctx->pos),
llu(*ptoken)); llu(*ptoken));
pos = (__u64) ctx->pos; pos = (__u64) ctx->pos;
@ -165,16 +155,6 @@ static int pvfs2_readdir(struct file *file, struct dir_context *ctx)
__func__, __func__,
&new_op->upcall.req.readdir.refn.khandle); &new_op->upcall.req.readdir.refn.khandle);
/*
* NOTE: the position we send to the readdir upcall is out of
* sync with ctx->pos since:
* 1. pvfs2 doesn't include the "." and ".." entries that are
* added below.
* 2. the introduction of distributed directory logic makes token no
* longer be related to f_pos and pos. Instead an independent
* variable is used inside the function and stored in the
* private_data of the file structure.
*/
new_op->upcall.req.readdir.token = *ptoken; new_op->upcall.req.readdir.token = *ptoken;
get_new_buffer_index: get_new_buffer_index:
@ -238,13 +218,18 @@ get_new_buffer_index:
} }
if (bytes_decoded != new_op->downcall.trailer_size) { if (bytes_decoded != new_op->downcall.trailer_size) {
gossip_err("pvfs2_readdir: # bytes decoded (%ld) != trailer size (%ld)\n", gossip_err("pvfs2_readdir: # bytes decoded (%ld) "
bytes_decoded, "!= trailer size (%ld)\n",
(long)new_op->downcall.trailer_size); bytes_decoded,
(long)new_op->downcall.trailer_size);
ret = -EINVAL; ret = -EINVAL;
goto out_destroy_handle; goto out_destroy_handle;
} }
/*
* pvfs2 doesn't actually store dot and dot-dot, but
* we need to have them represented.
*/
if (pos == 0) { if (pos == 0) {
ino = get_ino_from_khandle(dentry->d_inode); ino = get_ino_from_khandle(dentry->d_inode);
gossip_debug(GOSSIP_DIR_DEBUG, gossip_debug(GOSSIP_DIR_DEBUG,
@ -252,12 +237,7 @@ get_new_buffer_index:
__func__, __func__,
llu(pos)); llu(pos));
ret = dir_emit(ctx, ".", 1, ino, DT_DIR); ret = dir_emit(ctx, ".", 1, ino, DT_DIR);
ctx->pos++; pos += 1;
gossip_ldebug(GOSSIP_DIR_DEBUG,
"%s: ctx->pos:%lld\n",
__func__,
lld(ctx->pos));
pos++;
} }
if (pos == 1) { if (pos == 1) {
@ -267,62 +247,55 @@ get_new_buffer_index:
__func__, __func__,
llu(pos)); llu(pos));
ret = dir_emit(ctx, "..", 2, ino, DT_DIR); ret = dir_emit(ctx, "..", 2, ino, DT_DIR);
ctx->pos++; pos += 1;
gossip_ldebug(GOSSIP_DIR_DEBUG,
"%s: ctx->pos:%lld\n",
__func__,
lld(ctx->pos));
pos++;
} }
for (i = 0; i < rhandle.readdir_response.pvfs_dirent_outcount; i++) { /*
* we stored PVFS_ITERATE_NEXT in ctx->pos last time around
* to prevent "finding" dot and dot-dot on any iteration
* other than the first.
*/
if (ctx->pos == PVFS_ITERATE_NEXT)
ctx->pos = 0;
for (i = ctx->pos;
i < rhandle.readdir_response.pvfs_dirent_outcount;
i++) {
len = rhandle.readdir_response.dirent_array[i].d_length; len = rhandle.readdir_response.dirent_array[i].d_length;
current_entry = rhandle.readdir_response.dirent_array[i].d_name; current_entry = rhandle.readdir_response.dirent_array[i].d_name;
current_ino = pvfs2_khandle_to_ino( current_ino = pvfs2_khandle_to_ino(
&(rhandle.readdir_response.dirent_array[i].khandle)); &(rhandle.readdir_response.dirent_array[i].khandle));
gossip_debug(GOSSIP_DIR_DEBUG, gossip_debug(GOSSIP_DIR_DEBUG,
"calling dir_emit for %s with len %d, pos %ld\n", "calling dir_emit for %s with len %d"
", ctx->pos %ld\n",
current_entry, current_entry,
len, len,
(unsigned long)pos); (unsigned long)ctx->pos);
/*
* type is unknown. We don't return object type
* in the dirent_array. This leaves getdents
* clueless about type.
*/
ret = ret =
dir_emit(ctx, current_entry, len, current_ino, DT_UNKNOWN); dir_emit(ctx, current_entry, len, current_ino, DT_UNKNOWN);
if (!ret)
break;
ctx->pos++; ctx->pos++;
gossip_ldebug(GOSSIP_DIR_DEBUG, gossip_debug(GOSSIP_DIR_DEBUG,
"%s: ctx->pos:%lld\n", "%s: ctx->pos:%lld\n",
__func__, __func__,
lld(ctx->pos)); lld(ctx->pos));
pos++;
} }
/* this means that all of the dir_emit calls succeeded */ /*
if (i == rhandle.readdir_response.pvfs_dirent_outcount) { * we ran all the way through the last batch, set up for
/* update token */ * getting another batch...
*/
if (ret) {
*ptoken = rhandle.readdir_response.token; *ptoken = rhandle.readdir_response.token;
} else { ctx->pos = PVFS_ITERATE_NEXT;
/* this means a dir_emit call failed */
if (rhandle.readdir_response.token == PVFS_READDIR_END) {
/*
* If PVFS hit end of directory, then there
* is no way to do math on the token that it
* returned. Instead we go by ctx->pos but
* back up to account for the artificial .
* and .. entries.
*/
ctx->pos -= 3;
} else {
/*
* this means a dir_emit call failed. !!! need to set
* back to previous ctx->pos, no middle value allowed
*/
pos -= (i - 1);
ctx->pos -= (i - 1);
}
gossip_debug(GOSSIP_DIR_DEBUG,
"at least one dir_emit call failed. Setting ctx->pos to: %lld\n",
lld(ctx->pos));
} }
/* /*
@ -330,17 +303,11 @@ get_new_buffer_index:
*/ */
if (rhandle.readdir_response.token == PVFS_READDIR_END && if (rhandle.readdir_response.token == PVFS_READDIR_END &&
!buffer_full) { !buffer_full) {
gossip_debug(GOSSIP_DIR_DEBUG, "End of dir detected; setting ctx->pos to PVFS_READDIR_END.\n"); gossip_debug(GOSSIP_DIR_DEBUG,
"End of dir detected; setting ctx->pos to PVFS_READDIR_END.\n");
ctx->pos = PVFS_READDIR_END; ctx->pos = PVFS_READDIR_END;
} }
gossip_debug(GOSSIP_DIR_DEBUG,
"pos = %llu, token = %llu"
", ctx->pos should have been %lld\n",
llu(pos),
llu(*ptoken),
lld(ctx->pos));
out_destroy_handle: out_destroy_handle:
readdir_handle_dtor(bufmap, &rhandle); readdir_handle_dtor(bufmap, &rhandle);
out_free_op: out_free_op:

1
fs/orangefs/protocol.h
View File

@ -384,6 +384,7 @@ DECLARE_ERRNO_MAPPING()
#define INT32_MAX (2147483647) #define INT32_MAX (2147483647)
#define PVFS_ITERATE_START (INT32_MAX - 1) #define PVFS_ITERATE_START (INT32_MAX - 1)
#define PVFS_ITERATE_END (INT32_MAX - 2) #define PVFS_ITERATE_END (INT32_MAX - 2)
#define PVFS_ITERATE_NEXT (INT32_MAX - 3)
#define PVFS_READDIR_START PVFS_ITERATE_START #define PVFS_READDIR_START PVFS_ITERATE_START
#define PVFS_READDIR_END PVFS_ITERATE_END #define PVFS_READDIR_END PVFS_ITERATE_END
#define PVFS_IMMUTABLE_FL FS_IMMUTABLE_FL #define PVFS_IMMUTABLE_FL FS_IMMUTABLE_FL