2.6.15: via-rhine + link loss + autoneg off == trouble
PATCH 1/1 2.6.15: via-rhine + link loss + autoneg off == trouble
Hi,
2.6.15 still exhibits the via rhine "no tx" syndrome.
Again, the recipe:
* have via-rhine NIC
* unplug network cable
* reboot box
* force HDX (I do in with ethtool -s if autoneg off duplex half)
* plug cable back
* kernel still thinks that carrier is off despite "ethtool if" saying that link is detected.
Why:
...
if (intr_status & IntrLinkChange)
rhine_check_media(dev, 0);
...
static void rhine_check_media(struct net_device *dev, unsigned int init_media) {
struct rhine_private *rp = netdev_priv(dev);
void __iomem *ioaddr = rp->base;
mii_check_media(&rp->mii_if, debug, init_media);
...
unsigned int mii_check_media (struct mii_if_info *mii, unsigned int ok_to_print, unsigned int init_media) {
unsigned int old_carrier, new_carrier; int advertise, lpa, media, duplex;
int lpa2 = 0;
/* if forced media, go no further */
if (mii->force_media) <============================ HERE return 0; /* duplex did not change */
/* check current and old link status */ old_carrier = netif_carrier_ok(mii->dev) ? 1 : 0; new_carrier = (unsigned int) mii_link_ok(mii);
/* if carrier state did not change, this is a "bounce",
* just exit as everything is already set correctly */
if ((!init_media) && (old_carrier == new_carrier)) return 0; /* duplex did not change */
/* no carrier, nothing much to do */
if (!new_carrier) {
netif_carrier_off(mii->dev);
if (ok_to_print)
printk(KERN_INFO "%s: link down\n", mii->dev->name); return 0; /* duplex did not change */
}
/*
* we have carrier, see who's on the other end */
netif_carrier_on(mii->dev);
...
We can never reach netif_carrier_on if mii->force_media is true!
Quoted 10 lines · Jeff Garzik
Quoted 2 lines
As I've explained many times at this point,
force_mediameans that link is always on, andmii_link_ok()should not be called.Sorry, seems like I missed your mails...
Quoted 2 lines
The only bug, it seems, is that the "if (
mii->force_media)" branch should make sure to callnetif_carrier_on()Should also work.
Please find patch attached.
--
vda
drivers/net/mii.c +3 -1
@@ -267,8 +267,10 @@ unsigned int mii_check_media (struct miiint lpa2 =0 ;/* if forced media, go no further */ -if (mii->force_media) +if (mii->force_media) { +netif_carrier_on (mii->dev);return 0 ;/* duplex did not change */ + }/* check current and old link status */ old_carrier =netif_carrier_ok (mii->dev) ?1 :0 ;
Quoted 3 lines · Denis Vlasenko
Hi,
2.6.15 still exhibits the via rhine "no tx" syndrome.
Well you didn't cc: the via-rhine maintainer (Roger Luethi), no wonder the patch didn't get picked up...
Lee
Quoted 7 lines · Lee Revell
Quoted 3 lines · Denis Vlasenko
Hi,
2.6.15 still exhibits the via rhine "no tx" syndrome.
Well you didn't cc: the via-rhine maintainer (Roger Luethi), no wonder the patch didn't get picked up...
As long as you mention via-rhine, I am going to see it (thanks, procmail). I don't remember seeing vda post a patch before. First he posted a problem description, then an analysis (that I commented on), now a patch. Finally, this is not a via-rhine problem. It's a problem in the mii code and affects other drivers as well (those that use the generic mii code).
Roger
Quoted 4 lines · Roger Luethi
Quoted 3 lines · Lee Revell
Quoted 2 lines · Denis Vlasenko
Hi,
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
2.6.15 still exhibits the via rhine "no tx" syndrome.
Well you didn't cc: the via-rhine maintainer (Roger Luethi), no wonder the patch didn't get picked up...
As long as you mention via-rhine, I am going to see it (thanks, procmail). I don't remember seeing vda post a patch before. First he posted a problem description, then an analysis (that I commented on), now a patch. Finally,
It's my second patch for this (really trivial) problem. --
vda
Quoted 6 lines · Denis Vlasenko
Quoted 3 lines · Roger Luethi
As long as you mention via-rhine, I am going to see it (thanks, procmail). I don't remember seeing vda post a patch before. First he posted a problem description, then an analysis (that I commented on), now a patch. Finally,
It's my second patch for this (really trivial) problem.
Duly noted :-). Sorry I missed it the first time around.
Roger