| Message ID | 20220803075553.884839-2-aaron.ma@canonical.com |
|---|---|
| State | New |
| Headers | show |
| Series | Fix invalid MAC address after hotplug tbt dock | expand |
On Wed, Aug 03, 2022 at 03:55:53PM +0800, Aaron Ma wrote: > BugLink: https://bugs.launchpad.net/bugs/1942999 > > Such as dock hot plug event when runtime, for hardware implementation, > the MAC copy takes less than one second when BIOS enabled MAC passthrough. > After test on Lenovo TBT4 dock, 600ms is enough to update the > MAC address. > Otherwise ethernet fails to work. > > Link: https://lore.kernel.org/lkml/20210702045120.22855-2-aaron.ma@canonical.com/ > Signed-off-by: Aaron Ma <aaron.ma@canonical.com> > --- > drivers/net/ethernet/intel/igc/igc_main.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c > index f99819fc559d9..634cfdc203203 100644 > --- a/drivers/net/ethernet/intel/igc/igc_main.c > +++ b/drivers/net/ethernet/intel/igc/igc_main.c > @@ -6323,6 +6323,9 @@ static int igc_probe(struct pci_dev *pdev, > memcpy(&hw->mac.ops, ei->mac_ops, sizeof(hw->mac.ops)); > memcpy(&hw->phy.ops, ei->phy_ops, sizeof(hw->phy.ops)); > > + if (pci_is_thunderbolt_attached(pdev)) > + msleep(600); > + Unconditional sleeps are always unpredictable and a bit scary... isn't there a more reliable way to detect when the MAC address is updated? Like an interrupt, reading back from a register or somthing similar? Thanks, -Andrea > /* Initialize skew-specific constants */ > err = ei->get_invariants(hw); > if (err) > -- > 2.37.1 > > > -- > kernel-team mailing list > kernel-team@lists.ubuntu.com > https://lists.ubuntu.com/mailman/listinfo/kernel-team
On 8/4/22 14:37, Andrea Righi wrote: > On Wed, Aug 03, 2022 at 03:55:53PM +0800, Aaron Ma wrote: >> BugLink: https://bugs.launchpad.net/bugs/1942999 >> >> Such as dock hot plug event when runtime, for hardware implementation, >> the MAC copy takes less than one second when BIOS enabled MAC passthrough. >> After test on Lenovo TBT4 dock, 600ms is enough to update the >> MAC address. >> Otherwise ethernet fails to work. >> >> Link: https://lore.kernel.org/lkml/20210702045120.22855-2-aaron.ma@canonical.com/ >> Signed-off-by: Aaron Ma <aaron.ma@canonical.com> >> --- >> drivers/net/ethernet/intel/igc/igc_main.c | 3 +++ >> 1 file changed, 3 insertions(+) >> >> diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c >> index f99819fc559d9..634cfdc203203 100644 >> --- a/drivers/net/ethernet/intel/igc/igc_main.c >> +++ b/drivers/net/ethernet/intel/igc/igc_main.c >> @@ -6323,6 +6323,9 @@ static int igc_probe(struct pci_dev *pdev, >> memcpy(&hw->mac.ops, ei->mac_ops, sizeof(hw->mac.ops)); >> memcpy(&hw->phy.ops, ei->phy_ops, sizeof(hw->phy.ops)); >> >> + if (pci_is_thunderbolt_attached(pdev)) >> + msleep(600); >> + > Unconditional sleeps are always unpredictable and a bit scary... isn't > there a more reliable way to detect when the MAC address is updated? > Like an interrupt, reading back from a register or somthing similar? That's the ideal way, but I discussed with Intel, there is no such interrupts. In the datasheet, the I225 Windows driver implementation is to wait for 1 second. Regards, Aaron > Thanks, > -Andrea > >> /* Initialize skew-specific constants */ >> err = ei->get_invariants(hw); >> if (err) >> -- >> 2.37.1 >> >> >> -- >> kernel-team mailing list >> kernel-team@lists.ubuntu.com >> https://lists.ubuntu.com/mailman/listinfo/kernel-team
On 03.08.22 09:55, Aaron Ma wrote: > BugLink: https://bugs.launchpad.net/bugs/1942999 > > Such as dock hot plug event when runtime, for hardware implementation, > the MAC copy takes less than one second when BIOS enabled MAC passthrough. > After test on Lenovo TBT4 dock, 600ms is enough to update the > MAC address. > Otherwise ethernet fails to work. > > Link: https://lore.kernel.org/lkml/20210702045120.22855-2-aaron.ma@canonical.com/ > Signed-off-by: Aaron Ma <aaron.ma@canonical.com> > --- Applied to jammy:linux/master-next. Thanks. -Stefan > drivers/net/ethernet/intel/igc/igc_main.c | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c > index f99819fc559d9..634cfdc203203 100644 > --- a/drivers/net/ethernet/intel/igc/igc_main.c > +++ b/drivers/net/ethernet/intel/igc/igc_main.c > @@ -6323,6 +6323,9 @@ static int igc_probe(struct pci_dev *pdev, > memcpy(&hw->mac.ops, ei->mac_ops, sizeof(hw->mac.ops)); > memcpy(&hw->phy.ops, ei->phy_ops, sizeof(hw->phy.ops)); > > + if (pci_is_thunderbolt_attached(pdev)) > + msleep(600); > + > /* Initialize skew-specific constants */ > err = ei->get_invariants(hw); > if (err)
diff --git a/drivers/net/ethernet/intel/igc/igc_main.c b/drivers/net/ethernet/intel/igc/igc_main.c index f99819fc559d9..634cfdc203203 100644 --- a/drivers/net/ethernet/intel/igc/igc_main.c +++ b/drivers/net/ethernet/intel/igc/igc_main.c @@ -6323,6 +6323,9 @@ static int igc_probe(struct pci_dev *pdev, memcpy(&hw->mac.ops, ei->mac_ops, sizeof(hw->mac.ops)); memcpy(&hw->phy.ops, ei->phy_ops, sizeof(hw->phy.ops)); + if (pci_is_thunderbolt_attached(pdev)) + msleep(600); + /* Initialize skew-specific constants */ err = ei->get_invariants(hw); if (err)
BugLink: https://bugs.launchpad.net/bugs/1942999 Such as dock hot plug event when runtime, for hardware implementation, the MAC copy takes less than one second when BIOS enabled MAC passthrough. After test on Lenovo TBT4 dock, 600ms is enough to update the MAC address. Otherwise ethernet fails to work. Link: https://lore.kernel.org/lkml/20210702045120.22855-2-aaron.ma@canonical.com/ Signed-off-by: Aaron Ma <aaron.ma@canonical.com> --- drivers/net/ethernet/intel/igc/igc_main.c | 3 +++ 1 file changed, 3 insertions(+)