Subject: L68K: lp-driver-patch To: linux-m68k@phil.uni-sb.de (linux-list) Date: Mon, 6 Oct 1997 11:20:13 +0100 (MET) From: "J." Dorchain Sender: owner-linux-m68k@phil.uni-sb.de Reply-To: jdorchain@i-con.de Hi, this it needed to use the lp-driver with modules again. It is a diff against the oldenburg-snapshot (although you see it should apply to any recent kernel). For the cusious, this make insmod minor=?? work again. minor is the minor number the driver should try to get, with -1 to let the kernel choose the next free one. The symptom that made me find this change was a busserror when starting lpd. A call trace showed that it was killed during an open(/dev/lp). A kernel trace showed that it loaded the modules, but didn't initialised them because insmod returned an error (undefined symbol minor) when tryting to insert the module. Short of the story: Is it a kernel bug in the current version to crash when a module can't be initialised?? Joerg P.s.: Sorry that it's a little late, but the message somehow went into the dead.letter file until I found it now. --- lp_mfc.c.orig Thu Sep 25 01:27:42 1997 +++ lp_mfc.c Thu Sep 25 01:28:37 1997 @@ -34,6 +34,7 @@ static volatile int dummy; /* for trigger reads */ static int minor[MAX_LP] = { -1, -1, -1, -1, -1 }; +MODULE_PARM(minor,"1-5i"); #ifdef MODULE static unsigned int board_key[MAX_LP]; #endif --- lp_intern.c.orig Thu Sep 25 01:27:42 1997 +++ lp_intern.c Thu Sep 25 01:28:37 1997 @@ -34,6 *34,7 @@ #include static int minor = -1; +MODULE_PARM(minor,"i"); static void lp_int_out(int, int);