[PATCH] ACPI serial fix

[PATCH] ACPI serial fix

[Bjorn Helgaas]

Here's another little fix for ACPI serial discovery in 2.6.

Intel 870 firmware reports an extra zero-length IO port range, which
is bogus, as far as I can tell.  Ignore it.

===== drivers/serial/8250_acpi.c 1.4 vs edited =====
--- 1.4/drivers/serial/8250_acpi.c	Mon Oct  6 15:57:02 2003
+++ edited/drivers/serial/8250_acpi.c	Wed Oct  8 12:36:31 2003
@@ -38,6 +38,11 @@
 static acpi_status acpi_serial_port(struct serial_struct *req,
 				    struct acpi_resource_io *io)
 {
+	if (!io->range_length) {
+		printk(KERN_ERR "%s: zero-length IO port range?\n", __FUNCTION__);
+		return AE_ERROR;
+	}
+
 	req->port = io->min_base_address;
 	req->io_type = SERIAL_IO_PORT;
 	return AE_OK;

Re: [PATCH] ACPI serial fix

[Bjorn Helgaas]

On Wednesday 08 October 2003 12:41 pm, Bjorn Helgaas wrote:
> Here's another little fix for ACPI serial discovery in 2.6.

Rats.  Ignore the previous patch and use this, which I've actually
tested with the next step (removal of all the stuff in asm-ia64/serial.h):

Intel 870 firmware reports an extra zero-length IO port range, which
is bogus, as far as I can tell.  Ignore it.

===== drivers/serial/8250_acpi.c 1.4 vs edited =====
--- 1.4/drivers/serial/8250_acpi.c	Mon Oct  6 15:57:02 2003
+++ edited/drivers/serial/8250_acpi.c	Wed Oct  8 14:12:46 2003
@@ -38,8 +38,11 @@
 static acpi_status acpi_serial_port(struct serial_struct *req,
 				    struct acpi_resource_io *io)
 {
-	req->port = io->min_base_address;
-	req->io_type = SERIAL_IO_PORT;
+	if (io->range_length) {
+		req->port = io->min_base_address;
+		req->io_type = SERIAL_IO_PORT;
+	} else
+		printk(KERN_ERR "%s: zero-length IO port range?\n", __FUNCTION__);
 	return AE_OK;
 }