On Fri 2020-02-21 08:42:41, Greg Kroah-Hartman wrote:
From: Steve French stfrench@microsoft.com
[ Upstream commit d6fd41905ec577851734623fb905b1763801f5ef ]
We ran into a confusing problem where an application wasn't checking return code on close and so user didn't realize that the application ran out of disk space. log a warning message (once) in these cases. For example:
[ 8407.391909] Out of space writing to \oleg-server\small-share
Out of space can happen on any filesystem, and yes, it can be confusing. But why is cifs so special that we warn here (and not elsewhere) and why was this marked for stable?
Best regards, Pavel
+++ b/fs/cifs/smb2pdu.c @@ -3425,6 +3425,9 @@ smb2_writev_callback(struct mid_q_entry *mid) wdata->cfile->fid.persistent_fid, tcon->tid, tcon->ses->Suid, wdata->offset, wdata->bytes, wdata->result);
if (wdata->result == -ENOSPC)
printk_once(KERN_WARNING "Out of space writing to %s\n",
} else trace_smb3_write_done(0 /* no xid */, wdata->cfile->fid.persistent_fid,tcon->treeName);