write operation returns incorrect value in case device is out of space

There are at least two separate cases here:

 

1. device is out of space while writing an existing file.

This problem here is actually became worse with revision 1.32 of yaffs_fs.c that changed    

yaffs_commit_write routine to returns -1 (which is translated into EACESS errno) and at least is misleading. The attached patch fixes this problem while keeping desired effect of change 1.32.  

It also fixes a problem with partially written data that actually should return the amount of data written.

 

2. device is out of space while creating new object. In this case current code returns ENOMEM error code which is also misleading. Unfortunatly, I do not see a simple fix here,  the structure of code is such that it is hard to diffirenciate out-of-memory vs. out-of-space situation here but I think ENOSPC would be better here because it is more likely to happen.

 

M.

 

 

 

--- yaffs_fs.c 2005-12-16 18:23:10.000000000 -0800
+++ ./new/yaffs_fs.c 2005-12-21 16:53:31.000000000 -0800
@@ -624,10 +624,9 @@
 
  T(YAFFS_TRACE_OS,
    (KERN_DEBUG "yaffs_commit_write returning %d\n",
-    nWritten == nBytes ? 0 : -1));
-
- return nWritten == nBytes ? 0 : -1;
+    nWritten == nBytes ? 0 : nWritten));
 
+ return nWritten == nBytes ? 0 : nWritten;
 }
 
 static void yaffs_FillInodeFromObject(struct inode *inode, yaffs_Object * obj)
@@ -785,7 +784,7 @@
  }
  yaffs_GrossUnlock(dev);
 
- return nWritten != n ? -ENOSPC : nWritten;
+    return  ( nWritten == n || nWritten > 0 ) ? nWritten : -ENOSPC;
 }
 
 static int yaffs_readdir(struct file *f, void *dirent, filldir_t filldir)
 

_______________________________________________
yaffs mailing list
yaffs@xxxxxxxxxxxxxxxxxxxxxx
http://stoneboat.aleph1.co.uk/cgi-bin/mailman/listinfo/yaffs