Author: jra Date: 2006-04-10 18:43:41 +0000 (Mon, 10 Apr 2006) New Revision: 15019
WebSVN: http://websvn.samba.org/cgi-bin/viewcvs.cgi?view=rev&root=samba&rev=15019 Log: Fix issue with samba4 netbench torture tester, it sends break replies to "break to none from level2" requests and it shouldn't. Just don't log a debug level zero message. Jeremy. Modified: trunk/source/smbd/reply.c Changeset: Modified: trunk/source/smbd/reply.c =================================================================== --- trunk/source/smbd/reply.c 2006-04-10 15:33:04 UTC (rev 15018) +++ trunk/source/smbd/reply.c 2006-04-10 18:43:41 UTC (rev 15019) @@ -269,10 +269,13 @@ switch(next_mb_char_size(s)) { case 4: *d++ = *s++; + /*fall through*/ case 3: *d++ = *s++; + /*fall through*/ case 2: *d++ = *s++; + /*fall through*/ case 1: *d++ = *s++; break; @@ -374,10 +377,13 @@ switch(next_mb_char_size(s)) { case 4: *d++ = *s++; + /*fall through*/ case 3: *d++ = *s++; + /*fall through*/ case 2: *d++ = *s++; + /*fall through*/ case 1: *d++ = *s++; break; @@ -5241,7 +5247,14 @@ */ if (fsp->oplock_type == 0) { - DEBUG(0,("reply_lockingX: Error : oplock break from " + + /* The Samba4 nbench simulator doesn't understand + the difference between break to level2 and break + to none from level2 - it sends oplock break + replies in both cases. Don't keep logging an error + message here - just ignore it. JRA. */ + + DEBUG(5,("reply_lockingX: Error : oplock break from " "client for fnum = %d (oplock=%d) and no " "oplock granted on this file (%s).\n", fsp->fnum, fsp->oplock_type, fsp->fsp_name));