Hi Bruce,
On 7/15/20 10:24 PM, Bruce Ashfield wrote:
In message: [linux-yocto] [linux-yocto-dev standard/xlnx-soc][PATCH] net: macb:
drop wol modifications from upstream
on 15/07/2020 [email protected] wrote:
From: Quanyang Wang <[email protected]>
The v5.8-rc5 kernel applies the commit 253fe0943504 ("net: macb: fix
macb_get/set_wol() when moving to phylink") and the commit ced4799d0637
("net: macb: mark device wake capable when magic-packet property present")
on the magic packet WOL feature of macb. But in the SDK commit 88abc38a8b99
("net: macb: Use WOL via ARP"), the magic packet WOL support is removed and
using ARP WOL instead.
I wondered about this. But leaving the functions in place shouldn't
be a problem if they aren't actually called. It leaves a route back
to a more standard implementation in the future. Or are they actually
causing a build failure ?
Yes, they are causing a build failure since these functions contains
undeclared macros.
/home/wrsadmin/project/repos/linux/drivers/net/ethernet/cadence/macb_main.c:2939:16:
error: ‘MACB_WOL_HAS_MAGIC_PACKET’ undeclared (first use in this
function); did you mean ‘MACB_WOL_MTI_OFFSET’?
if (bp->wol & MACB_WOL_HAS_MAGIC_PACKET) {
^~~~~~~~~~~~~~~~~~~~~~~~~
MACB_WOL_MTI_OFFSET
/home/wrsadmin/project/repos/linux/drivers/net/ethernet/cadence/macb_main.c:2939:16:
note: each undeclared identifier is reported only once for each function
it appears in
/home/wrsadmin/project/repos/linux/drivers/net/ethernet/cadence/macb_main.c:2943:17:
error: ‘MACB_WOL_ENABLED’ undeclared (first use in this function); did
you mean ‘MACB_WOL_SIZE’?
if (bp->wol & MACB_WOL_ENABLED)
^~~~~~~~~~~~~~~~
MACB_WOL_SIZE
/home/wrsadmin/project/repos/linux/drivers/net/ethernet/cadence/macb_main.c:
In function ‘macb_set_wol’:
/home/wrsadmin/project/repos/linux/drivers/net/ethernet/cadence/macb_main.c:2961:18:
error: ‘MACB_WOL_HAS_MAGIC_PACKET’ undeclared (first use in this
function); did you mean ‘MACB_WOL_MTI_OFFSET’?
if (!(bp->wol & MACB_WOL_HAS_MAGIC_PACKET) ||
^~~~~~~~~~~~~~~~~~~~~~~~~
MACB_WOL_MTI_OFFSET
/home/wrsadmin/project/repos/linux/drivers/net/ethernet/cadence/macb_main.c:2966:14:
error: ‘MACB_WOL_ENABLED’ undeclared (first use in this function); did
you mean ‘MACB_WOL_SIZE’?
bp->wol |= MACB_WOL_ENABLED;
^~~~~~~~~~~~~~~~
MACB_WOL_SIZE
Thanks,
Quanyang
Either either way, I've merged the patch .. I was just looking for
more information for future merges when similar conflicts pop up
again.
So drop these 2 upstream patches from xlnx-soc branch.
Signed-off-by: Quanyang Wang <[email protected]>
---
drivers/net/ethernet/cadence/macb_main.c | 50 ------------------------
1 file changed, 50 deletions(-)
diff --git a/drivers/net/ethernet/cadence/macb_main.c
b/drivers/net/ethernet/cadence/macb_main.c
index 2f7a9dcf9060..69c5f267625f 100644
--- a/drivers/net/ethernet/cadence/macb_main.c
+++ b/drivers/net/ethernet/cadence/macb_main.c
@@ -2932,46 +2932,6 @@ static void macb_get_regs(struct net_device *dev, struct
ethtool_regs *regs,
regs_buff[13] = gem_readl(bp, DMACFG);
}
-static void macb_get_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
-{
- struct macb *bp = netdev_priv(netdev);
-
- if (bp->wol & MACB_WOL_HAS_MAGIC_PACKET) {
- phylink_ethtool_get_wol(bp->phylink, wol);
- wol->supported |= WAKE_MAGIC;
-
- if (bp->wol & MACB_WOL_ENABLED)
- wol->wolopts |= WAKE_MAGIC;
- }
-}
-
-static int macb_set_wol(struct net_device *netdev, struct ethtool_wolinfo *wol)
-{
- struct macb *bp = netdev_priv(netdev);
- int ret;
-
- /* Pass the order to phylink layer */
- ret = phylink_ethtool_set_wol(bp->phylink, wol);
- /* Don't manage WoL on MAC if handled by the PHY
- * or if there's a failure in talking to the PHY
- */
- if (!ret || ret != -EOPNOTSUPP)
- return ret;
-
- if (!(bp->wol & MACB_WOL_HAS_MAGIC_PACKET) ||
- (wol->wolopts & ~WAKE_MAGIC))
- return -EOPNOTSUPP;
-
- if (wol->wolopts & WAKE_MAGIC)
- bp->wol |= MACB_WOL_ENABLED;
- else
- bp->wol &= ~MACB_WOL_ENABLED;
-
- device_set_wakeup_enable(&bp->pdev->dev, bp->wol & MACB_WOL_ENABLED);
-
- return 0;
-}
-
static int macb_get_link_ksettings(struct net_device *netdev,
struct ethtool_link_ksettings *kset)
{
@@ -4574,17 +4534,7 @@ static int macb_probe(struct platform_device *pdev)
bp->tx_clk = tx_clk;
bp->rx_clk = rx_clk;
bp->tsu_clk = tsu_clk;
-<<<<<<< HEAD
bp->jumbo_max_len = macb_config->jumbo_max_len;
-=======
- if (macb_config)
- bp->jumbo_max_len = macb_config->jumbo_max_len;
-
- bp->wol = 0;
- if (of_get_property(np, "magic-packet", NULL))
- bp->wol |= MACB_WOL_HAS_MAGIC_PACKET;
- device_set_wakeup_capable(&pdev->dev, bp->wol &
MACB_WOL_HAS_MAGIC_PACKET);
->>>>>>> standard/base
That should have never slipped in. I swear that I fixed it, but I can see
that I didn't actually commit the file.
Thanks for catching it, and fixing it up!
Bruce
spin_lock_init(&bp->lock);
--
2.17.1
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#8840):
https://lists.yoctoproject.org/g/linux-yocto/message/8840
Mute This Topic: https://lists.yoctoproject.org/mt/75513918/21656
Group Owner: [email protected]
Unsubscribe: https://lists.yoctoproject.org/g/linux-yocto/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
