Hi,
I'm currently porting the r8169 driver to 2.6.28 (which builds fine by
now, but not sure if it works properly yet). After a few hours of
debugging I found a problem in the Ethercat master code, which is due to
an API change in the 2.6.27 kernels.
The device_create() call has received another argument in >= 2.6.27.
This is the 2.6.26 version:
extern struct device *device_create(struct class *cls, struct device *parent,
dev_t devt, const char *fmt, ...)
__attribute__((format(printf, 4, 5)));
This is from >= 2.6.27:
extern struct device *device_create(struct class *cls, struct device *parent,
dev_t devt, void *drvdata,
const char *fmt, ...)
__attribute__((format(printf, 5, 6)));
Attached is a patch which should fix the issue.
I'll post more patches and test results of r8169 on 2.6.28 later next
week or so.
HTH, Uwe.
--
http://www.hermann-uwe.de | http://www.holsham-traders.de
http://www.crazy-hacks.org | http://www.unmaintained-free-software.org
diff -Naur ethercat-devel-r1667.orig/master/master.c ethercat-devel-r1667/master/master.c
--- ethercat-devel-r1667.orig/master/master.c 2009-02-24 13:51:39.000000000 +0100
+++ ethercat-devel-r1667/master/master.c 2009-03-06 19:32:57.000000000 +0100
@@ -202,7 +202,11 @@
if (ret)
goto out_clear_fsm;
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 26)
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 27)
+ master->class_device = device_create(class, NULL,
+ MKDEV(MAJOR(device_number), master->index), NULL,
+ "EtherCAT%u", master->index);
+#elif LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 26)
master->class_device = device_create(class, NULL,
MKDEV(MAJOR(device_number), master->index),
"EtherCAT%u", master->index);
_______________________________________________
etherlab-users mailing list
[email protected]
http://lists.etherlab.org/mailman/listinfo/etherlab-users
