LKML Archive on
help / color / mirror / Atom feed
* [PATCH] serial.c 2.4.23 uart_offset fix
@ 2003-11-27  1:28 Ed Vance
  0 siblings, 0 replies; only message in thread
From: Ed Vance @ 2003-11-27  1:28 UTC (permalink / raw)
  To: ''; +Cc: ''

Hi Marcelo,

The serial driver has a default of 8 for the "board uart_offset" config that
is implemented at about line 3954 in a ternary expression.  It is well and
good for boards that use I/O port space, but fails in the ioremap call at
about line 3965 for boards that use memory space. The bug is that when
uart_offset is zero, it does not actually get changed to the default value,
so the ioremap call is passed a zero length.

The following patch replaces the ternary expression with an explicit check
for zero and assignment to the default value. This works for both I/O ports
and memory space.  --And it actually compiles a little smaller than the
existing code on x86! 

Please put this in at your convenience after the rc's are done. 


diff -urN -X dontdiff.txt linux-2.4.23-rc5/drivers/char/serial.c
--- linux-2.4.23-rc5/drivers/char/serial.c	Wed Nov 26 16:23:45 2003
+++ linux-2.4.23-rc5-ml/drivers/char/serial.c	Wed Nov 26 16:38:13 2003
@@ -3950,8 +3950,11 @@
 	port =  pci_resource_start(dev, base_idx) + offset;
+	if (board->uart_offset == 0)
+		board->uart_offset = 8;
 	if ((board->flags & SPCI_FL_BASE_TABLE) == 0)
-		port += idx * (board->uart_offset ? board->uart_offset : 8);
+		port += idx * board->uart_offset;
 	if (IS_PCI_REGION_IOPORT(dev, base_idx)) {
 		req->port = port;

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2003-11-27  1:28 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2003-11-27  1:28 [PATCH] serial.c 2.4.23 uart_offset fix Ed Vance

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).