net: fec: generate warning when using deprecated phy reset
Subsystems (3)
PATCH 1/1 net: fec: generate warning when using deprecated phy reset
Allowing the fec to reset its PHY via the phy-reset-gpios devicetree property is deprecated. To improve developer awareness, generate a warning whenever the deprecated property is used.
--- drivers/net/ethernet/freescale/fec_main.c | 6 ++++++ 1 file changed, 6 insertions(+)
drivers/net/ethernet/freescale/fec_main.c +6 -0
@@ -3244,6 +3244,12 @@ static int fec_reset_phy(struct platform_device *pdev)else if (!gpio_is_valid (phy_reset))return 0 ; +/* Recommended way to provide a PHY reset: +* - create a phy devicetree node, and link it to its fec (phy-handle) +* - add your reset gpio to the phy devicetree node +*/ +dev_warn (&pdev->dev,"devicetree: phy-reset-gpios is deprecated\n" ); + err =of_property_read_u32 (np,"phy-reset-post-delay" , &phy_post_delay);/* valid reset duration should be less than 1s */ if (!err && phy_post_delay >1000 )
Signature · 1 line
2.17.1
Am Donnerstag, den 18.07.2019, 10:34 -0400 schrieb Sven Van Asbroeck:
Quoted 4 lines
Allowing the fec to reset its PHY via the phy-reset-gpios devicetree property is deprecated. To improve developer awareness, generate a warning whenever the deprecated property is used.
Not really a fan of this. This will cause existing DTs, which are provided by the firmware in an ideal world and may not change at the same rate as the kernel, to generate a warning with new kernels. Not really helpful from the user experience point of view.
Regards,
Lucas
Quoted 22 lines
Signed-off-by Sven Van Asbroeck <[email protected]>---
drivers/net/ethernet/freescale/fec_main.c | 6 ++++++1 file changed, 6 insertions(+)drivers/net/ethernet/freescale/fec_main.c +0
-0@@ -3244,6 +3244,12 @@ static int fec_reset_phy(struct platform_device *pdev) > else if (!gpio_is_valid(phy_reset)) > return 0;drivers/net/ethernet/freescale/fec_main.c static int fec_reset_phy(struct platform_device *pdev) · line 3251
Quoted 5 lines
drivers/net/ethernet/freescale/fec_main.c +5
-0+/* Recommended way to provide a PHY reset: +* - create a phy devicetree node, and link it to its fec (phy-handle) +* - add your reset gpio to the phy devicetree node +*/ +dev_warn (&pdev->dev,"devicetree: phy-reset-gpios is deprecated\n" );+
Quoted 3 lines
err =
of_property_read_u32(np, "phy-reset-post-delay", &phy_post_delay); /* valid reset duration should be less than 1s */
if (!err &&phy_post_delay> 1000)
Quoted 11 lines · Lucas Stach
Am Donnerstag, den 18.07.2019, 10:34 -0400 schrieb Sven Van Asbroeck:
Quoted 4 lines
Allowing the fec to reset its PHY via the phy-reset-gpios devicetree property is deprecated. To improve developer awareness, generate a warning whenever the deprecated property is used.
Not really a fan of this. This will cause existing DTs, which are provided by the firmware in an ideal world and may not change at the same rate as the kernel, to generate a warning with new kernels. Not really helpful from the user experience point of view.
I agree. I don't think this warning is useful.
Lucas, Fabio,
Quoted 7 lines · Fabio Estevam
Quoted 5 lines
Not really a fan of this. This will cause existing DTs, which are provided by the firmware in an ideal world and may not change at the same rate as the kernel, to generate a warning with new kernels. Not really helpful from the user experience point of view.
I agree. I don't think this warning is useful.
Few users watch the dmesg log, But I totally see your point.
The problem I'm trying to address here is this: when I want to reset the fec phy, I go look at existing devicetrees. Nearly all of them use phy-reset-gpios, so that's what I'll use. But, when I try to upstream the patch, the maintainer will tell me:
"no, that is deprecated, use this other method".
Is there a good solution you can think of which would point the unwary developer to the correct phy reset method?
See my previous thread here:
https://lkml.org/lkml/2019/7/15/1764
Quoted 13 lines · Sven Van Asbroeck
Lucas, Fabio,
Quoted 7 lines · Fabio Estevam
Quoted 5 lines
Not really a fan of this. This will cause existing DTs, which are provided by the firmware in an ideal world and may not change at the same rate as the kernel, to generate a warning with new kernels. Not really helpful from the user experience point of view.
I agree. I don't think this warning is useful.
Few users watch the dmesg log, But I totally see your point.
Quoted 8 lines
The problem I'm trying to address here is this: when I want to reset the fec phy, I go look at existing devicetrees. Nearly all of them use phy-reset-gpios, so that's what I'll use. But, when I try to upstream the patch, the maintainer will tell me:
"no, that is deprecated, use this other method".Is there a good solution you can think of which would point the unwary developer to the correct phy reset method?
Hi Sven
One option would be to submit a patch or a patchset changing all existing device tree files to make use of the core method. Anybody cut/pasting will then automatically use the correct core way of doing it.
There is also a move towards using YAML to verify the correctness of DT files. It should be possible to mark the old property as deprecated, so there will be a build time warning, not a boot time warning.
Andrew
Hi Andrew,
Quoted 8 lines · Andrew Lunn
Hi Sven
One option would be to submit a patch or a patchset changing all existing device tree files to make use of the core method. Anybody cut/pasting will then automatically use the correct core way of doing it.
Quoted 5 lines
There is also a move towards using YAML to verify the correctness of DT files. It should be possible to mark the old property as deprecated, so there will be a build time warning, not a boot time warning.
Thanks for the helpful suggestions, that makes sense.
What I keep forgetting in my little arm-imx6 world, is that devicetrees aren't in-kernel apis, but that they have out-of-kernel dependencies. It makes more sense to to see them as userspace apis, albeit directed at firmware/bootloaders, right?
So if bootloaders were as varied/uncontrollable as userspace, then deprecated properties would have to be supported forever...
Quoted 4 lines
What I keep forgetting in my little arm-imx6 world, is that devicetrees aren't in-kernel apis, but that they have out-of-kernel dependencies. It makes more sense to to see them as userspace apis, albeit directed at
firmware/bootloaders, right?
It is an ongoing debate, but generally they should be considered ABI and follow the ABI rules about not breaking backwards compatibility.
However, there is also an argument that something like a NAS box running Debian is going to use the DT blob which came with the kernel, so deprecated DT properties and the code to support them could be removed after a period of time.
Andrew
David Miller
maintainerFrom: Sven Van Asbroeck <[email protected]>
Date: Thu, 18 Jul 2019 10:34:28 -0400
Quoted 6 lines
Allowing the fec to reset its PHY via the phy-reset-gpios devicetree property is deprecated. To improve developer awareness, generate a warning whenever the deprecated property is used.
Signed-off-by Sven Van Asbroeck <[email protected]>
The device tree documentation in the kernel tree is where information like this belongs. Yelling at the user in the kernel logs is not.
Quoted 3 lines · David Miller
The device tree documentation in the kernel tree is where information like this belongs. Yelling at the user in the kernel logs is not.
Good point, thank you. I'll post a patch which will do exactly that.