aboutsummaryrefslogtreecommitdiffstats
path: root/fs
diff options
context:
space:
mode:
authorJan Harkes <jaharkes@cs.cmu.edu>2007-07-21 07:37:26 -0400
committerLinus Torvalds <torvalds@woody.linux-foundation.org>2007-07-21 20:49:14 -0400
commitd3fec424b23c47686efcf3f2004c3f1c1cee4d9c (patch)
treeffad3cf369d35e85fca9a340347eea09caa53ff3 /fs
parentb50731732f926d6c49fd0724616a7344c31cd5cf (diff)
coda: remove CODA_STORE/CODA_RELEASE upcalls
This is an variation on the patch sent by Christoph Hellwig which kills file_count abuse by the Coda kernel module by moving the coda_flush functionality into coda_release. However part of reason we were using the coda_flush callback was to allow Coda to pass errors that occur during writeback from the userspace cache manager back to close(). As Al Viro explained on linux-fsdevel, it is impossible to guarantee that such errors can in fact be returned back to the caller. There are many cases where the last reference to a file is not released by the close system call and it is also impossible to pick some close as a 'last-close' and delay it until all other references have been destroyed. The CODA_STORE/CODA_RELEASE upcall combination is clearly a broken design, and it is better to remove it completely. Signed-off-by: Jan Harkes <jaharkes@cs.cmu.edu> Cc: Christoph Hellwig <hch@lst.de> Cc: Al Viro <viro@ftp.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'fs')
-rw-r--r--fs/coda/dir.c1
-rw-r--r--fs/coda/file.c65
-rw-r--r--fs/coda/upcall.c49
3 files changed, 7 insertions, 108 deletions
diff --git a/fs/coda/dir.c b/fs/coda/dir.c
index 8e61236abf4a..f89ff083079b 100644
--- a/fs/coda/dir.c
+++ b/fs/coda/dir.c
@@ -86,7 +86,6 @@ const struct file_operations coda_dir_operations = {
86 .read = generic_read_dir, 86 .read = generic_read_dir,
87 .readdir = coda_readdir, 87 .readdir = coda_readdir,
88 .open = coda_open, 88 .open = coda_open,
89 .flush = coda_flush,
90 .release = coda_release, 89 .release = coda_release,
91 .fsync = coda_fsync, 90 .fsync = coda_fsync,
92}; 91};
diff --git a/fs/coda/file.c b/fs/coda/file.c
index 7594962604c2..29137ff3ca67 100644
--- a/fs/coda/file.c
+++ b/fs/coda/file.c
@@ -25,10 +25,6 @@
25 25
26#include "coda_int.h" 26#include "coda_int.h"
27 27
28/* if CODA_STORE fails with EOPNOTSUPP, venus clearly doesn't support
29 * CODA_STORE/CODA_RELEASE and we fall back on using the CODA_CLOSE upcall */
30static int use_coda_close;
31
32static ssize_t 28static ssize_t
33coda_file_read(struct file *coda_file, char __user *buf, size_t count, loff_t *ppos) 29coda_file_read(struct file *coda_file, char __user *buf, size_t count, loff_t *ppos)
34{ 30{
@@ -163,47 +159,6 @@ int coda_open(struct inode *coda_inode, struct file *coda_file)
163 return 0; 159 return 0;
164} 160}
165 161
166int coda_flush(struct file *coda_file, fl_owner_t id)
167{
168 unsigned short flags = coda_file->f_flags & ~O_EXCL;
169 unsigned short coda_flags = coda_flags_to_cflags(flags);
170 struct coda_file_info *cfi;
171 struct inode *coda_inode;
172 int err = 0, fcnt;
173
174 lock_kernel();
175
176 /* last close semantics */
177 fcnt = file_count(coda_file);
178 if (fcnt > 1)
179 goto out;
180
181 /* No need to make an upcall when we have not made any modifications
182 * to the file */
183 if ((coda_file->f_flags & O_ACCMODE) == O_RDONLY)
184 goto out;
185
186 if (use_coda_close)
187 goto out;
188
189 cfi = CODA_FTOC(coda_file);
190 BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC);
191
192 coda_inode = coda_file->f_path.dentry->d_inode;
193
194 err = venus_store(coda_inode->i_sb, coda_i2f(coda_inode), coda_flags,
195 coda_file->f_uid);
196
197 if (err == -EOPNOTSUPP) {
198 use_coda_close = 1;
199 err = 0;
200 }
201
202out:
203 unlock_kernel();
204 return err;
205}
206
207int coda_release(struct inode *coda_inode, struct file *coda_file) 162int coda_release(struct inode *coda_inode, struct file *coda_file)
208{ 163{
209 unsigned short flags = (coda_file->f_flags) & (~O_EXCL); 164 unsigned short flags = (coda_file->f_flags) & (~O_EXCL);
@@ -215,21 +170,11 @@ int coda_release(struct inode *coda_inode, struct file *coda_file)
215 170
216 lock_kernel(); 171 lock_kernel();
217 172
218 if (!use_coda_close) {
219 err = venus_release(coda_inode->i_sb, coda_i2f(coda_inode),
220 coda_flags);
221 if (err == -EOPNOTSUPP) {
222 use_coda_close = 1;
223 err = 0;
224 }
225 }
226
227 cfi = CODA_FTOC(coda_file); 173 cfi = CODA_FTOC(coda_file);
228 BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC); 174 BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC);
229 175
230 if (use_coda_close) 176 err = venus_close(coda_inode->i_sb, coda_i2f(coda_inode),
231 err = venus_close(coda_inode->i_sb, coda_i2f(coda_inode), 177 coda_flags, coda_file->f_uid);
232 coda_flags, coda_file->f_uid);
233 178
234 host_inode = cfi->cfi_container->f_path.dentry->d_inode; 179 host_inode = cfi->cfi_container->f_path.dentry->d_inode;
235 cii = ITOC(coda_inode); 180 cii = ITOC(coda_inode);
@@ -246,7 +191,10 @@ int coda_release(struct inode *coda_inode, struct file *coda_file)
246 coda_file->private_data = NULL; 191 coda_file->private_data = NULL;
247 192
248 unlock_kernel(); 193 unlock_kernel();
249 return err; 194
195 /* VFS fput ignores the return value from file_operations->release, so
196 * there is no use returning an error here */
197 return 0;
250} 198}
251 199
252int coda_fsync(struct file *coda_file, struct dentry *coda_dentry, int datasync) 200int coda_fsync(struct file *coda_file, struct dentry *coda_dentry, int datasync)
@@ -288,7 +236,6 @@ const struct file_operations coda_file_operations = {
288 .write = coda_file_write, 236 .write = coda_file_write,
289 .mmap = coda_file_mmap, 237 .mmap = coda_file_mmap,
290 .open = coda_open, 238 .open = coda_open,
291 .flush = coda_flush,
292 .release = coda_release, 239 .release = coda_release,
293 .fsync = coda_fsync, 240 .fsync = coda_fsync,
294 .splice_read = coda_file_splice_read, 241 .splice_read = coda_file_splice_read,
diff --git a/fs/coda/upcall.c b/fs/coda/upcall.c
index cd561d2e90b0..cdb4c07a7870 100644
--- a/fs/coda/upcall.c
+++ b/fs/coda/upcall.c
@@ -160,55 +160,8 @@ int venus_lookup(struct super_block *sb, struct CodaFid *fid,
160 return error; 160 return error;
161} 161}
162 162
163int venus_store(struct super_block *sb, struct CodaFid *fid, int flags,
164 vuid_t uid)
165{
166 union inputArgs *inp;
167 union outputArgs *outp;
168 int insize, outsize, error;
169#ifdef CONFIG_CODA_FS_OLD_API
170 struct coda_cred cred = { 0, };
171 cred.cr_fsuid = uid;
172#endif
173
174 insize = SIZE(store);
175 UPARG(CODA_STORE);
176
177#ifdef CONFIG_CODA_FS_OLD_API
178 memcpy(&(inp->ih.cred), &cred, sizeof(cred));
179#else
180 inp->ih.uid = uid;
181#endif
182
183 inp->coda_store.VFid = *fid;
184 inp->coda_store.flags = flags;
185
186 error = coda_upcall(coda_vcp(sb), insize, &outsize, inp);
187
188 CODA_FREE(inp, insize);
189 return error;
190}
191
192int venus_release(struct super_block *sb, struct CodaFid *fid, int flags)
193{
194 union inputArgs *inp;
195 union outputArgs *outp;
196 int insize, outsize, error;
197
198 insize = SIZE(release);
199 UPARG(CODA_RELEASE);
200
201 inp->coda_release.VFid = *fid;
202 inp->coda_release.flags = flags;
203
204 error = coda_upcall(coda_vcp(sb), insize, &outsize, inp);
205
206 CODA_FREE(inp, insize);
207 return error;
208}
209
210int venus_close(struct super_block *sb, struct CodaFid *fid, int flags, 163int venus_close(struct super_block *sb, struct CodaFid *fid, int flags,
211 vuid_t uid) 164 vuid_t uid)
212{ 165{
213 union inputArgs *inp; 166 union inputArgs *inp;
214 union outputArgs *outp; 167 union outputArgs *outp;