--==_Exmh_12394438760
Content-Type: text/plain; charset=us-ascii
The attached patch enables the iso9960 filesystem to handle the case where the
directory entry spans two blocks (as it tends to on most Microsoft written CDs)
I don't know why the 2.0 implementation was dropped from the kernel, but if it
was for aesthetic reasons, I hope the new one looks cleaner.
James Bottomley
james@bpgc.com
--==_Exmh_12394438760
Content-Type: text/plain ; name="isofs_50-patch"; charset=us-ascii
Content-Description: isofs_50-patch
Content-Disposition: attachment; filename="isofs_50-patch"
--- 1.1 1997/08/15 14:31:48
+++ fs/isofs/namei.c 1997/08/15 16:58:23
@@ -63,12 +63,14 @@
{
unsigned long bufsize = ISOFS_BUFFER_SIZE(dir);
unsigned char bufbits = ISOFS_BUFFER_BITS(dir);
- unsigned int block, i, f_pos, offset, inode_number;
+ unsigned int block, i, f_pos, offset,
+ inode_number = 0; /* shut gcc up */
struct buffer_head * bh;
unsigned int old_offset;
int dlen, rrflag, match;
char * dpnt;
- struct iso_directory_record * de;
+ struct iso_directory_record * de = NULL; /* shut gcc up */
+ char de_not_in_buf = 0; /* true if de is in kmalloc'd memory */
char c;
*ino = 0;
@@ -84,17 +86,36 @@
if (!block || !(bh = bread(dir->i_dev,block,bufsize))) return NULL;
while (f_pos < dir->i_size) {
- de = (struct iso_directory_record *) (bh->b_data + offset);
- inode_number = (block << bufbits) + (offset & (bufsize - 1));
- /* If byte is zero, this is the end of file, or time to move to
- the next sector. Usually 2048 byte boundaries. */
+ /* if de is in kmalloc'd memory, do not point to the
+ next de, instead we will move to the next sector */
+ if(!de_not_in_buf) {
+ de = (struct iso_directory_record *)
+ (bh->b_data + offset);
+ inode_number = (block << bufbits)
+ + (offset & (bufsize - 1));
+ }
+
+ /* If byte is zero, or we had to fetch this de past
+ the end of the buffer, this is the end of file, or
+ time to move to the next sector. Usually 2048 byte
+ boundaries. */
- if (*((unsigned char *) de) == 0) {
+ if (*((unsigned char *) de) == 0 || de_not_in_buf) {
+ if(de_not_in_buf) {
+ /* james@bpgc.com: Since we slopped
+ past the end of the last buffer, we
+ must start some way into the new
+ one */
+ de_not_in_buf = 0;
+ kfree(de);
+ offset -= bufsize;
+ }
+ else
+ offset = 0;
brelse(bh);
- offset = 0;
f_pos = ((f_pos & ~(ISOFS_BLOCK_SIZE - 1))
- + ISOFS_BLOCK_SIZE);
+ + ISOFS_BLOCK_SIZE) + offset;
if( f_pos >= dir->i_size )
{
@@ -111,11 +132,28 @@
offset += *((unsigned char *) de);
f_pos += *((unsigned char *) de);
- /* Handle case where the directory entry spans two blocks.
- Usually 1024 byte boundaries */
+ /* james@bpgc.com: new code to handle case where the
+ directory entry spans two blocks. Usually 1024
+ byte boundaries */
if (offset >= bufsize) {
- printk("Directory entry extends past end of iso9660 block\n");
- return 0;
+ struct buffer_head *bh_next;
+
+ /* james@bpgc.com: read the next block, and
+ copy the split de into a newly kmalloc'd
+ buffer */
+ block = isofs_bmap(dir,f_pos>>bufbits);
+ if (!block ||
+ !(bh_next = bread(dir->i_dev,block,bufsize)))
+ return NULL;
+
+ de = (struct iso_directory_record *)
+ kmalloc(offset - old_offset, GFP_KERNEL);
+ memcpy((char *)de, bh->b_data + old_offset,
+ bufsize - old_offset);
+ memcpy((char *)de + bufsize - old_offset,
+ bh_next->b_data, offset - bufsize);
+ brelse(bh_next);
+ de_not_in_buf = 1;
}
dlen = de->name_len[0];
@@ -161,11 +199,15 @@
find_rock_ridge_relocation(de,dir));
}
*ino = inode_number;
+ if(de_not_in_buf)
+ kfree(de);
return bh;
}
}
out:
brelse(bh);
+ if(de_not_in_buf)
+ kfree(de);
return NULL;
}
--==_Exmh_12394438760--