| Message ID | 1449614642-8161-1-git-send-email-paul.gortmaker@windriver.com |
|---|---|
| State | Not Applicable, archived |
| Delegated to: | David Miller |
| Headers | show |
On Tue, 2015-12-08 at 17:44 -0500, Paul Gortmaker wrote: > This file was originally cloned off of the MPC8641D-HPCN reference > platform, which actually had a PHY IRQ line connected. However > this board does not. The bogus entry was largely inert and went > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > ("net: phy: Use interrupts when available in NOLINK state") was > added to the tree. > > With the above commit, the board fails to NFS boot since it sits > waiting for a PHY IRQ event that of course never arrives. Removing > the bogus entries from the DTS file fixes the issue. Commit 321beec5047a ("net: phy: Use interrupts when available in NOLINK state") went into 4.4-rc2, so this is a fix for 4.4 right? cheers -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[Re: [PATCH] sbc8641: drop bogus PHY IRQ entries from DTS file] On 09/12/2015 (Wed 12:10) Michael Ellerman wrote: > On Tue, 2015-12-08 at 17:44 -0500, Paul Gortmaker wrote: > > > This file was originally cloned off of the MPC8641D-HPCN reference > > platform, which actually had a PHY IRQ line connected. However > > this board does not. The bogus entry was largely inert and went > > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > > ("net: phy: Use interrupts when available in NOLINK state") was > > added to the tree. > > > > With the above commit, the board fails to NFS boot since it sits > > waiting for a PHY IRQ event that of course never arrives. Removing > > the bogus entries from the DTS file fixes the issue. > > Commit 321beec5047a ("net: phy: Use interrupts when available in NOLINK state") > went into 4.4-rc2, so this is a fix for 4.4 right? Correct, but I'm guessing there are not thousands of users of this board out there, so it probably doesn't matter a whole lot if it goes in the same release where the regression happened or one release later... P. -- > > cheers > -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
From: Paul Gortmaker <paul.gortmaker@windriver.com> Date: Tue, 8 Dec 2015 17:44:02 -0500 > This file was originally cloned off of the MPC8641D-HPCN reference > platform, which actually had a PHY IRQ line connected. However > this board does not. The bogus entry was largely inert and went > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > ("net: phy: Use interrupts when available in NOLINK state") was > added to the tree. > > With the above commit, the board fails to NFS boot since it sits > waiting for a PHY IRQ event that of course never arrives. Removing > the bogus entries from the DTS file fixes the issue. > > Cc: Andrew Lunn <andrew@lunn.ch> > Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> I'm assuming this will go via the powerpc tree, not mine. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, 2015-12-08 at 21:40 -0500, David Miller wrote: > From: Paul Gortmaker <paul.gortmaker@windriver.com> > Date: Tue, 8 Dec 2015 17:44:02 -0500 > > This file was originally cloned off of the MPC8641D-HPCN reference > > platform, which actually had a PHY IRQ line connected. However > > this board does not. The bogus entry was largely inert and went > > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > > ("net: phy: Use interrupts when available in NOLINK state") was > > added to the tree. > > > > With the above commit, the board fails to NFS boot since it sits > > waiting for a PHY IRQ event that of course never arrives. Removing > > the bogus entries from the DTS file fixes the issue. > > > > Cc: Andrew Lunn <andrew@lunn.ch> > > Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> > > I'm assuming this will go via the powerpc tree, not mine. Yep I'll take it. cheers -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, 2015-12-08 at 21:04 -0500, Paul Gortmaker wrote: > [Re: [PATCH] sbc8641: drop bogus PHY IRQ entries from DTS file] On 09/12/2015 (Wed 12:10) Michael Ellerman wrote: > > On Tue, 2015-12-08 at 17:44 -0500, Paul Gortmaker wrote: > > > This file was originally cloned off of the MPC8641D-HPCN reference > > > platform, which actually had a PHY IRQ line connected. However > > > this board does not. The bogus entry was largely inert and went > > > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > > > ("net: phy: Use interrupts when available in NOLINK state") was > > > added to the tree. > > > > > > With the above commit, the board fails to NFS boot since it sits > > > waiting for a PHY IRQ event that of course never arrives. Removing > > > the bogus entries from the DTS file fixes the issue. > > > > Commit 321beec5047a ("net: phy: Use interrupts when available in NOLINK state") > > went into 4.4-rc2, so this is a fix for 4.4 right? > > Correct, but I'm guessing there are not thousands of users of this board > out there, so it probably doesn't matter a whole lot if it goes in the > same release where the regression happened or one release later... OK sure. As it happens I'm just prepping a fixes branch, so I'll drop it in. cheers -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Dec 08, 2015 at 05:44:02PM -0500, Paul Gortmaker wrote: > This file was originally cloned off of the MPC8641D-HPCN reference > platform, which actually had a PHY IRQ line connected. However > this board does not. The bogus entry was largely inert and went > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > ("net: phy: Use interrupts when available in NOLINK state") was > added to the tree. > > With the above commit, the board fails to NFS boot since it sits > waiting for a PHY IRQ event that of course never arrives. Removing > the bogus entries from the DTS file fixes the issue. Hi Paul Originally the interrupt is used for detecting the link has gone down. That would of also been bogus before. Have you tried this? If that is also broken, maybe you need to add a fixes: tag so that it gets back ported? Andrew -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[Re: [PATCH] sbc8641: drop bogus PHY IRQ entries from DTS file] On 09/12/2015 (Wed 04:07) Andrew Lunn wrote: > On Tue, Dec 08, 2015 at 05:44:02PM -0500, Paul Gortmaker wrote: > > This file was originally cloned off of the MPC8641D-HPCN reference > > platform, which actually had a PHY IRQ line connected. However > > this board does not. The bogus entry was largely inert and went > > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > > ("net: phy: Use interrupts when available in NOLINK state") was > > added to the tree. > > > > With the above commit, the board fails to NFS boot since it sits > > waiting for a PHY IRQ event that of course never arrives. Removing > > the bogus entries from the DTS file fixes the issue. > > Hi Paul > > Originally the interrupt is used for detecting the link has gone > down. That would of also been bogus before. Have you tried this? If Haven't tried it, but chances are you are right. > that is also broken, maybe you need to add a fixes: tag so that it > gets back ported? Nobody noticed in the past ~7 years or so, but I guess I can point Greg at it once it is present in mainline. The main reason I'd Cc'd netdev is just to possibly save anyone else the investigation if they ran into the same issue on a different board -which seems highly probable IMHO. Paul. -- > > Andrew -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
> > Originally the interrupt is used for detecting the link has gone > > down. That would of also been bogus before. Have you tried this? If > > Haven't tried it, but chances are you are right. > > > that is also broken, maybe you need to add a fixes: tag so that it > > gets back ported? > > Nobody noticed in the past ~7 years or so, but I guess I can point Greg > at it once it is present in mainline. The main reason I'd Cc'd netdev > is just to possibly save anyone else the investigation if they ran into > the same issue on a different board -which seems highly probable IMHO. Yes, you are probably right about other boards. So this change might be considered as causing a regression. But if they are already broken, because link down does not work, there might be less demand to get the change reverted.... Andrew -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, 2015-08-12 at 22:44:02 UTC, Paul Gortmaker wrote: > This file was originally cloned off of the MPC8641D-HPCN reference > platform, which actually had a PHY IRQ line connected. However > this board does not. The bogus entry was largely inert and went > undetected until commit 321beec5047af83db90c88114b7e664b156f49fe > ("net: phy: Use interrupts when available in NOLINK state") was > added to the tree. > > With the above commit, the board fails to NFS boot since it sits > waiting for a PHY IRQ event that of course never arrives. Removing > the bogus entries from the DTS file fixes the issue. > > Cc: Andrew Lunn <andrew@lunn.ch> > Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Applied to powerpc fixes, thanks. https://git.kernel.org/powerpc/c/5b01310cfc8d2302dcca1d8d cheers -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/powerpc/boot/dts/sbc8641d.dts b/arch/powerpc/boot/dts/sbc8641d.dts index 631ede72e226..68f0ed7626bd 100644 --- a/arch/powerpc/boot/dts/sbc8641d.dts +++ b/arch/powerpc/boot/dts/sbc8641d.dts @@ -227,23 +227,15 @@ reg = <0x520 0x20>; phy0: ethernet-phy@1f { - interrupt-parent = <&mpic>; - interrupts = <10 1>; reg = <0x1f>; }; phy1: ethernet-phy@0 { - interrupt-parent = <&mpic>; - interrupts = <10 1>; reg = <0>; }; phy2: ethernet-phy@1 { - interrupt-parent = <&mpic>; - interrupts = <10 1>; reg = <1>; }; phy3: ethernet-phy@2 { - interrupt-parent = <&mpic>; - interrupts = <10 1>; reg = <2>; }; tbi0: tbi-phy@11 {
This file was originally cloned off of the MPC8641D-HPCN reference platform, which actually had a PHY IRQ line connected. However this board does not. The bogus entry was largely inert and went undetected until commit 321beec5047af83db90c88114b7e664b156f49fe ("net: phy: Use interrupts when available in NOLINK state") was added to the tree. With the above commit, the board fails to NFS boot since it sits waiting for a PHY IRQ event that of course never arrives. Removing the bogus entries from the DTS file fixes the issue. Cc: Andrew Lunn <andrew@lunn.ch> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> --- arch/powerpc/boot/dts/sbc8641d.dts | 8 -------- 1 file changed, 8 deletions(-)