Jeff Garzik <[EMAIL PROTECTED]> writes: > Arnaud Patard wrote: >> Hi, >> I was testing the sis190 driver and got an oops. Tests show that the >> error >> path is the following : sis190_init_one() >> -> register_netdev >> * Userspace sees the new interface and run ethtool ethX. >> * ethtool calls ethtools_get_settings >> * ethtools_get_settings try to access to the mii datas >> * oops occurs >> -> sis190_mii_probe The fix I made is to replace the sequence by >> the following one : sis190_init_one() >> -> sis190_mii_probe >> -> register_netdev >> Regards, >> Arnaud Patard >> Signed-off-by: Arnaud Patard <[EMAIL PROTECTED]> >> ------------------------------------------------------------------------ >> diff -rpuN a/drivers/net/sis190.c b/drivers/net/sis190.c >> --- a/drivers/net/sis190.c 2005-08-25 15:36:24.000000000 +0200 >> +++ b/drivers/net/sis190.c 2005-08-25 15:36:34.000000000 +0200 >> @@ -1779,14 +1779,18 @@ static int __devinit sis190_init_one(str >> dev->irq = pdev->irq; >> dev->base_addr = (unsigned long) 0xdead; >> + pci_set_drvdata(pdev, dev); >> + >> spin_lock_init(&tp->lock); >> + rc = sis190_mii_probe(dev); >> + if (rc < 0) >> + goto err_release_board; >> + >> rc = register_netdev(dev); >> if (rc < 0) >> goto err_release_board; >> - pci_set_drvdata(pdev, dev); >> - rc = sis190_mii_probe(dev); > > yes mii probe should be moved, but pci_set_drvdata() should be the > very last thing to occur.
Hi, Francois told me that. I should admitt that it was a quick and dirty fix. A proper fix has already been sent to you (commit 3690b6c124fbc7259634f3b80d92a6d9fe51ec79 on your libata). Regards, Arnaud - To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
