From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; cs; rv:1.8.0.7) Gecko/20061011 Fedora/1.5.0.7-7.fc6 Firefox/1.5.0.7 Description of problem: I noticed this while reading longer directories (+300 files). The `ls` gives: ls: reading directory .: Invalid argument Accessing files is OK. It seems to be file name length (or sum) dependent. fstab line: //storage/home /home/lukas/storage cifs domain=W2KDOMAIN,user=XXXXX,password=XXXXX,uid=lukas,gid=lukas,file_mode=0644,dir_mode=0755,rw,user,auto In FC5 2.6.18-2200 kernel is same result. In FC5 pre 2200 kernels this was OK. Version-Release number of selected component (if applicable): kernel-2.6.18-1.2798.fc6 samba-client-3.0.23c-2 How reproducible: Always Steps to Reproduce: 1. Mount CIFS share (from Win 2003 server). 2. Make files: for f in `seq -f '%05g' 1 127 `; do touch $f; done (126 files works OK) 3. ls directory Actual Results: ls: reading directory .: Invalid argument Expected Results: List of files Additional info:
Once the mount succeeds, the mount.cifs program is out of the picture. Any subsequent failures are entirely in the kernel. I'm redirecting this bug to the correct component.
This is fixed in CVS, will be in an update soon.
Created attachment 139907 [details] dmesg output
Sorry, forgot to add a comment for attachment 139907 [details]. :-( Just confirming that I'm having the same issue with kernel-2.6.18-1.2798.fc6 and samba-client-3.0.23c-2. Directory listings containing approx 100+ entries return "ls: reading directory .: Invalid argument". This is happening when connecting to various Samba servers of differing versions.
should be fixed in 2.6.18-1.2849.fc6 now in updates
(In reply to comment #5) > should be fixed in 2.6.18-1.2849.fc6 now in updates I'm not finding 2849.fc6. But I did try .2835.fc6 from testing, and bonnie++ run on a CIFS mount still gives intermittent "Drastic I/O" errors, complaining that the directory it is trying to remove is not empty. It creates 16384 files in a directory and then deletes them and removes the directory. Apparently, it thinks it deleted everything but really didn't. This seems like it would be related. Again, though, that is in 2835 which is the latest I have been able to find.
(In reply to comment #6) > (In reply to comment #5) > > should be fixed in 2.6.18-1.2849.fc6 now in updates > > Again, though, that is in 2835 which is the latest I have been able to find. Looks like many of the mirrors are yet to update.. I'm grabbing 2835 from here: http://mirror.linux.duke.edu/pub/fedora/linux/core/updates/6/i386/ right now. Hopefully that't fix the prob.
I'm testing with 2849 now and it looks like it does, indeed, resolve the problem. I'm not sure if it is just luck of the draw or if there has really been a badly needed change, but I am finding that the bugs that I have run into with FC6 are getting developer attention, and are getting resolved, in a way that my FC5 bugs did not. Two of the significant bugs that I have been dealing with were resolved today. And on a Sunday, no less. Kudos!
Seems resolved in kernel-2.6.18-1.2849.fc6 Thanks
verified, now works for me in kernel-2.6.18-1.2849.fc6
*** Bug 214735 has been marked as a duplicate of this bug. ***