64bit BAR bug in pci.c

• Next message: Bill Maidment: "Re: mount misbehaves in 2.4.0-test6 and test7"
• Previous message: Alexander Viro: "Re: SCO: "thread creation is about a thousand times faster than on"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

   I believe I've come across a small piece of broken code for 64bit BAR
support in drivers/pci/pci.c. In the function pci_read_bases(), after
we determine we have a 64bit BAR we read the upper 32bits into variable
l.
We then get the size mask and save it into variable tmp. However, we
then go back and use l to calculate the new size of the range. I
propose
the diff below to fix this. I'm unable to test this code, but if anyone
can, I'd appreciate some feedback. I'm not on the linux-kernel list,
so please be sure I'm cc'd on any follow up. Thanks,

        Alex Williamson

--
Alex Williamson
awilliam@fc.hp.com
--- linux/drivers/pci/pci.c	Mon Jun 19 14:42:38 2000
+++ linux-pcifix/drivers/pci/pci.c	Mon Aug 28 16:52:37 2000
@@ -518,8 +518,8 @@
 			pci_write_config_dword(dev, reg+4, ~0);
 			pci_read_config_dword(dev, reg+4, &tmp);
 			pci_write_config_dword(dev, reg+4, l);
-			if (l)
-				res->end = res->start + (((unsigned long) ~l) << 32);
+			if (tmp)
+				res->end = res->start + (((unsigned long) ~tmp) << 32);
 #else
 			if (l) {
 				printk(KERN_ERR "PCI: Unable to handle 64-bit address for device
%s\n", dev->slot_name);
--
Alex Williamson
awilliam@fc.hp.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

• Next message: Bill Maidment: "Re: mount misbehaves in 2.4.0-test6 and test7"
• Previous message: Alexander Viro: "Re: SCO: "thread creation is about a thousand times faster than on"
• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

This archive was generated by hypermail 2b29 : Thu Aug 31 2000 - 21:00:22 EST