.../bindings/net/ethernet-controller.yaml | 1 + Documentation/networking/phy.rst | 7 ++++ drivers/net/phy/broadcom.c | 39 ++++++++++++++++--- drivers/net/phy/phy-core.c | 1 + drivers/net/phy/phy_caps.c | 4 ++ drivers/net/phy/phylink.c | 1 + include/linux/brcmphy.h | 6 +++ include/linux/phy.h | 4 ++ 8 files changed, 58 insertions(+), 5 deletions(-)
PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for their two-wire PHYs. It can be used with most Ethernet controllers under certain limitations (no half-duplex link modes etc.). PATCH 2 - Add MII-Lite PHY interface type PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x PHYs PATCH 4 - Fix the BCM54811 PHY initialization so that it conforms to the datasheet regarding a reserved bit in the LRE Control register, which must be written to zero after every device reset. Also fix the LRE Status register reading, there is another bit to be ignored on bcm54811. Changes in v2: - Applied reviewers' comments - Divided into more patches (separated common and Broadcom PHY specific code) Changes in v3: - Added MII-Lite documentation Changes in v4: - Added missing Fixes headers Changes in v5: - Removed unused PHY register bits definition Kamil Horák - 2N (4): net: phy: MII-Lite PHY interface mode dt-bindings: ethernet-phy: add MII-Lite phy interface type net: phy: bcm5481x: MII-Lite activation net: phy: bcm54811: Fix the PHY initialization .../bindings/net/ethernet-controller.yaml | 1 + Documentation/networking/phy.rst | 7 ++++ drivers/net/phy/broadcom.c | 39 ++++++++++++++++--- drivers/net/phy/phy-core.c | 1 + drivers/net/phy/phy_caps.c | 4 ++ drivers/net/phy/phylink.c | 1 + include/linux/brcmphy.h | 6 +++ include/linux/phy.h | 4 ++ 8 files changed, 58 insertions(+), 5 deletions(-) -- 2.39.5
On Tue, 1 Jul 2025 09:50:11 +0200 Kamil Horák - 2N wrote: > PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for > their two-wire PHYs. It can be used with most Ethernet controllers > under certain limitations (no half-duplex link modes etc.). > > PATCH 2 - Add MII-Lite PHY interface type > > PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x > PHYs > > PATCH 4 - Fix the BCM54811 PHY initialization so that it conforms > to the datasheet regarding a reserved bit in the LRE Control > register, which must be written to zero after every device reset. > Also fix the LRE Status register reading, there is another bit to > be ignored on bcm54811. I'm a bit lost why the first 3 patches are included in a series for net. My naive reading is we didn't support this extra mode, now we do, which sounds like a new feature.. Patch 4, sure, but the dependency is not obvious.
On 7/2/25 15:02, Jakub Kicinski wrote: > On Tue, 1 Jul 2025 09:50:11 +0200 Kamil Horák - 2N wrote: >> PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for >> their two-wire PHYs. It can be used with most Ethernet controllers >> under certain limitations (no half-duplex link modes etc.). >> >> PATCH 2 - Add MII-Lite PHY interface type >> >> PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x >> PHYs >> >> PATCH 4 - Fix the BCM54811 PHY initialization so that it conforms >> to the datasheet regarding a reserved bit in the LRE Control >> register, which must be written to zero after every device reset. >> Also fix the LRE Status register reading, there is another bit to >> be ignored on bcm54811. > > I'm a bit lost why the first 3 patches are included in a series for net. > My naive reading is we didn't support this extra mode, now we do, > which sounds like a new feature.. Patch 4, sure, but the dependency > is not obvious. I don't see the dependency either, at least not in an explicit way. Kamil, could patch #4 stand on its own and routed through "net" while patches 1-3 are routed through "net-next"? Thanks -- Florian
On 7/3/25 01:46, Florian Fainelli wrote: > On 7/2/25 15:02, Jakub Kicinski wrote: >> On Tue, 1 Jul 2025 09:50:11 +0200 Kamil Horák - 2N wrote: >>> PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for >>> their two-wire PHYs. It can be used with most Ethernet controllers >>> under certain limitations (no half-duplex link modes etc.). >>> >>> PATCH 2 - Add MII-Lite PHY interface type >>> >>> PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x >>> PHYs >>> >>> PATCH 4 - Fix the BCM54811 PHY initialization so that it conforms >>> to the datasheet regarding a reserved bit in the LRE Control >>> register, which must be written to zero after every device reset. >>> Also fix the LRE Status register reading, there is another bit to >>> be ignored on bcm54811. >> >> I'm a bit lost why the first 3 patches are included in a series for net. >> My naive reading is we didn't support this extra mode, now we do, >> which sounds like a new feature.. Patch 4, sure, but the dependency >> is not obvious. > > I don't see the dependency either, at least not in an explicit way. > Kamil, could patch #4 stand on its own and routed through "net" while > patches 1-3 are routed through "net-next"? It can be done this way, however, even the patch #3 is effectively a fix, not new feature, because the bcm54811 PHY in MLP package only has MII-Lite interface available externally. As far I know, there is no BGA casing available for bcm54811 (unlike bcm54810, that one having both MLP and BGA). Thus, it cannot function without being switched to MII-Lite mode. The introduction of MII-Lite itself is clearly a new feature and it is even (theoretically) available for any MII-capable PHY. So if putting it all to net it is really impossible or contrary to the net-next vs. net selection rules, let's divide it.... To get fully functional, bcm54811-based networking, all patches are necessary so any other user out there must wait for both branches to join. > > Thanks Kamil
On 7/3/25 11:03 AM, Kamil Horák (2N) wrote: > On 7/3/25 01:46, Florian Fainelli wrote: >> On 7/2/25 15:02, Jakub Kicinski wrote: >>> On Tue, 1 Jul 2025 09:50:11 +0200 Kamil Horák - 2N wrote: >>>> PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for >>>> their two-wire PHYs. It can be used with most Ethernet controllers >>>> under certain limitations (no half-duplex link modes etc.). >>>> >>>> PATCH 2 - Add MII-Lite PHY interface type >>>> >>>> PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x >>>> PHYs >>>> >>>> PATCH 4 - Fix the BCM54811 PHY initialization so that it conforms >>>> to the datasheet regarding a reserved bit in the LRE Control >>>> register, which must be written to zero after every device reset. >>>> Also fix the LRE Status register reading, there is another bit to >>>> be ignored on bcm54811. >>> >>> I'm a bit lost why the first 3 patches are included in a series for net. >>> My naive reading is we didn't support this extra mode, now we do, >>> which sounds like a new feature.. Patch 4, sure, but the dependency >>> is not obvious. >> >> I don't see the dependency either, at least not in an explicit way. >> Kamil, could patch #4 stand on its own and routed through "net" while >> patches 1-3 are routed through "net-next"? > It can be done this way, however, even the patch #3 is effectively a > fix, not new feature, because the bcm54811 PHY in MLP package only has > MII-Lite interface available externally. As far I know, there is no BGA > casing available for bcm54811 (unlike bcm54810, that one having both MLP > and BGA). Thus, it cannot function without being switched to MII-Lite > mode. The introduction of MII-Lite itself is clearly a new feature and > it is even (theoretically) available for any MII-capable PHY. So if > putting it all to net it is really impossible or contrary to the > net-next vs. net selection rules, let's divide it.... > To get fully functional, bcm54811-based networking, all patches are > necessary so any other user out there must wait for both branches to join. The above makes sense to me, but I think it would be nice to capture some of this info in the cover letter - the fact that many people were confused by the series is an hint the info was indeed missing and required. Please resubmit with an extended cover letter, thanks! (unless Florian jumps in and explicitly asks for something different :-P) Paolo
On 7/3/25 12:03, Paolo Abeni wrote: > On 7/3/25 11:03 AM, Kamil Horák (2N) wrote: >> On 7/3/25 01:46, Florian Fainelli wrote: >>> On 7/2/25 15:02, Jakub Kicinski wrote: >>>> On Tue, 1 Jul 2025 09:50:11 +0200 Kamil Horák - 2N wrote: >>>>> PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for >>>>> their two-wire PHYs. It can be used with most Ethernet controllers >>>>> under certain limitations (no half-duplex link modes etc.). >>>>> >>>>> PATCH 2 - Add MII-Lite PHY interface type >>>>> >>>>> PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x >>>>> PHYs >>>>> >>>>> PATCH 4 - Fix the BCM54811 PHY initialization so that it conforms >>>>> to the datasheet regarding a reserved bit in the LRE Control >>>>> register, which must be written to zero after every device reset. >>>>> Also fix the LRE Status register reading, there is another bit to >>>>> be ignored on bcm54811. >>>> >>>> I'm a bit lost why the first 3 patches are included in a series for net. >>>> My naive reading is we didn't support this extra mode, now we do, >>>> which sounds like a new feature.. Patch 4, sure, but the dependency >>>> is not obvious. >>> >>> I don't see the dependency either, at least not in an explicit way. >>> Kamil, could patch #4 stand on its own and routed through "net" while >>> patches 1-3 are routed through "net-next"? >> It can be done this way, however, even the patch #3 is effectively a >> fix, not new feature, because the bcm54811 PHY in MLP package only has >> MII-Lite interface available externally. As far I know, there is no BGA >> casing available for bcm54811 (unlike bcm54810, that one having both MLP >> and BGA). Thus, it cannot function without being switched to MII-Lite >> mode. The introduction of MII-Lite itself is clearly a new feature and >> it is even (theoretically) available for any MII-capable PHY. So if >> putting it all to net it is really impossible or contrary to the >> net-next vs. net selection rules, let's divide it.... >> To get fully functional, bcm54811-based networking, all patches are >> necessary so any other user out there must wait for both branches to join. > > The above makes sense to me, but I think it would be nice to capture > some of this info in the cover letter - the fact that many people were > confused by the series is an hint the info was indeed missing and required. OK, tried to write more about it > > Please resubmit with an extended cover letter, thanks! > (unless Florian jumps in and explicitly asks for something different :-P) > No such request appeared since yesterday, so submitting an updated text version > Paolo > Kamil
© 2016 - 2025 Red Hat, Inc.