| Message ID | 20190309153923.22806-1-beagleboard@davidjohnsummers.uk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v3,1/3] arm: dts: rk3288-tinker.dtsi: Fix SD card detection | expand |
Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: > The Problem: > > On ASUS Tinker Board S, when booting from the eMMC, and there is card > in the sd slot, there are constant errors. > > Also when warm reboot, uboot can not access the sd slot > > Cause: > > Identified by Robin Murphy @ ARM. The Card Detect on rk3288 > devices is pulled up by vccio-sd; so when the regulator powers this > off, card detect gives spurious errors. A second problem, is during > power down, vccio-sd apprears to be powered down. This causes a > problem when warm rebooting from the sd card. This was identified by > Jonas Karlman. > > History: > > A common fault on these rk3288 board, which impliment the reference > design. > > When this arose before: > > http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153.html > > And Ulf and Jaehoon clearly said this was a broken card detect design, > which should be solved via polling > > Solution: > > Hence broken-cd is set as a property. This cures the errors. The > powering down of vccio-sd during reboot is cured by adding > regulator-boot-on. > > This solutions has been fairly widely reviewed and tested. > > Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> > Reviewed by: Robin Murphy <robin.murphy@arm.com> > Reviewed by: Jonas Karlman <jonas@kwiboo.se> > Test by: TheSaint @ ArchLinux Arm please follow patch submission guidelines, here it would be Tested-by: ... with an actual Name + mail address Also I did try to go through the way big mail thread but didn't find an actual "Reviewed-by" from Robin Murphy, similar to what he pointed out in patch2 Other than that and including the past discussions, this looks good. Heiko > --- > arch/arm/boot/dts/rk3288-tinker.dtsi | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/arch/arm/boot/dts/rk3288-tinker.dtsi b/arch/arm/boot/dts/rk3288-tinker.dtsi > index aa107ee41b8b..ef653c3209bc 100644 > --- a/arch/arm/boot/dts/rk3288-tinker.dtsi > +++ b/arch/arm/boot/dts/rk3288-tinker.dtsi > @@ -254,6 +254,7 @@ > }; > > vccio_sd: LDO_REG5 { > + regulator-boot-on; > regulator-min-microvolt = <1800000>; > regulator-max-microvolt = <3300000>; > regulator-name = "vccio_sd"; > @@ -430,7 +431,7 @@ > bus-width = <4>; > cap-mmc-highspeed; > cap-sd-highspeed; > - card-detect-delay = <200>; > + broken-cd; > disable-wp; /* wp not hooked up */ > pinctrl-names = "default"; > pinctrl-0 = <&sdmmc_clk &sdmmc_cmd &sdmmc_cd &sdmmc_bus4>; >
On 2019-03-12 2:08 pm, Heiko Stuebner wrote: > Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: >> The Problem: >> >> On ASUS Tinker Board S, when booting from the eMMC, and there is card >> in the sd slot, there are constant errors. >> >> Also when warm reboot, uboot can not access the sd slot >> >> Cause: >> >> Identified by Robin Murphy @ ARM. The Card Detect on rk3288 >> devices is pulled up by vccio-sd; so when the regulator powers this >> off, card detect gives spurious errors. A second problem, is during >> power down, vccio-sd apprears to be powered down. This causes a >> problem when warm rebooting from the sd card. This was identified by >> Jonas Karlman. >> >> History: >> >> A common fault on these rk3288 board, which impliment the reference >> design. >> >> When this arose before: >> >> http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153.html >> >> And Ulf and Jaehoon clearly said this was a broken card detect design, >> which should be solved via polling >> >> Solution: >> >> Hence broken-cd is set as a property. This cures the errors. The >> powering down of vccio-sd during reboot is cured by adding >> regulator-boot-on. >> >> This solutions has been fairly widely reviewed and tested. >> >> Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> >> Reviewed by: Robin Murphy <robin.murphy@arm.com> >> Reviewed by: Jonas Karlman <jonas@kwiboo.se> >> Test by: TheSaint @ ArchLinux Arm > > please follow patch submission guidelines, here it would be > Tested-by: ... with an actual Name + mail address > > Also I did try to go through the way big mail thread but didn't > find an actual "Reviewed-by" from Robin Murphy, similar to what > he pointed out in patch2 Ah, I didn't get a direct CC of this patch so I hadn't spotted that - FWIW though, I'm happy with Doug's reasoning for "broken-cd" being the most robust solution for now, so feel free to keep Reviewed-by: Robin Murphy <robin.murphy@arm.com> for this one. Cheers, Robin. > > Other than that and including the past discussions, this looks good. > > > Heiko > >> --- >> arch/arm/boot/dts/rk3288-tinker.dtsi | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> diff --git a/arch/arm/boot/dts/rk3288-tinker.dtsi b/arch/arm/boot/dts/rk3288-tinker.dtsi >> index aa107ee41b8b..ef653c3209bc 100644 >> --- a/arch/arm/boot/dts/rk3288-tinker.dtsi >> +++ b/arch/arm/boot/dts/rk3288-tinker.dtsi >> @@ -254,6 +254,7 @@ >> }; >> >> vccio_sd: LDO_REG5 { >> + regulator-boot-on; >> regulator-min-microvolt = <1800000>; >> regulator-max-microvolt = <3300000>; >> regulator-name = "vccio_sd"; >> @@ -430,7 +431,7 @@ >> bus-width = <4>; >> cap-mmc-highspeed; >> cap-sd-highspeed; >> - card-detect-delay = <200>; >> + broken-cd; >> disable-wp; /* wp not hooked up */ >> pinctrl-names = "default"; >> pinctrl-0 = <&sdmmc_clk &sdmmc_cmd &sdmmc_cd &sdmmc_bus4>; >> > > > >
Hi Jonas, Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: > The Problem: > > On ASUS Tinker Board S, when booting from the eMMC, and there is card > in the sd slot, there are constant errors. > > Also when warm reboot, uboot can not access the sd slot > > Cause: > > Identified by Robin Murphy @ ARM. The Card Detect on rk3288 > devices is pulled up by vccio-sd; so when the regulator powers this > off, card detect gives spurious errors. A second problem, is during > power down, vccio-sd apprears to be powered down. This causes a > problem when warm rebooting from the sd card. This was identified by > Jonas Karlman. > > History: > > A common fault on these rk3288 board, which impliment the reference > design. > > When this arose before: > > http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153.html > > And Ulf and Jaehoon clearly said this was a broken card detect design, > which should be solved via polling > > Solution: > > Hence broken-cd is set as a property. This cures the errors. The > powering down of vccio-sd during reboot is cured by adding > regulator-boot-on. > > This solutions has been fairly widely reviewed and tested. > > Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> > Reviewed by: Robin Murphy <robin.murphy@arm.com> > Reviewed by: Jonas Karlman <jonas@kwiboo.se> in v2, you gave a Tested-by: Jonas Karlman <jonas@kwiboo.se> it seems, which somehow transformed into a reviewed-by in v3. V3 did change a bit in its approach - for the better, but do you approve of adding that Reviewed-by above (and maybe also providing a Tested-by for the new approach)? Thanks Heiko
On 12/03/2019 14:22, Heiko Stuebner wrote: > Hi Jonas, > > Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: >> The Problem: >> >> On ASUS Tinker Board S, when booting from the eMMC, and there is card >> in the sd slot, there are constant errors. >> >> Also when warm reboot, uboot can not access the sd slot >> >> Cause: >> >> Identified by Robin Murphy @ ARM. The Card Detect on rk3288 >> devices is pulled up by vccio-sd; so when the regulator powers this >> off, card detect gives spurious errors. A second problem, is during >> power down, vccio-sd apprears to be powered down. This causes a >> problem when warm rebooting from the sd card. This was identified by >> Jonas Karlman. >> >> History: >> >> A common fault on these rk3288 board, which impliment the reference >> design. >> >> When this arose before: >> >> http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153.html >> >> And Ulf and Jaehoon clearly said this was a broken card detect design, >> which should be solved via polling >> >> Solution: >> >> Hence broken-cd is set as a property. This cures the errors. The >> powering down of vccio-sd during reboot is cured by adding >> regulator-boot-on. >> >> This solutions has been fairly widely reviewed and tested. >> >> Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> >> Reviewed by: Robin Murphy <robin.murphy@arm.com> >> Reviewed by: Jonas Karlman <jonas@kwiboo.se> > in v2, you gave a > Tested-by: Jonas Karlman <jonas@kwiboo.se> > it seems, which somehow transformed into a reviewed-by in v3. > > V3 did change a bit in its approach - for the better, but do you approve > of adding that Reviewed-by above (and maybe also providing a Tested-by > for the new approach)? > > Thanks > Heiko > > Hi all, Yes apologies for these signed-off lines etc wrong. Its still a learning experiment for me! Both Robin and Jonas added a lot to Patch, as explained in the write up. Robin came up with the cause of the fault, and Jonas noticed a second fault. So the question is to give these two their due. I think what they both added was important. Thanks Robin for the official "Reviewed-by". Jonas tested the second version, which kept power on vccio-sd all the time, whereas v1 and v3 cured the problem with broken-cd. So although both fixed the problem, they way performed is different. So don't think I can say Jonas Tested-by v3. As for the tested by "TheSaint @ ArchLinux Arm", he actually tested just above every single version of the patch, and a whole lot more before it was sent in - he is our ASUS Tinker Board user on arch. He doesn't really want his email in the open, but if push comes to shove he'll allow it. Question is do you need this for the Patch to pass? Heiko, I'm happy to change the sign off etc, and resent the patch (as v4). But would be the same patch. Think the question is what is best for you. Don't want to send unneeded email ... anyway let me know (private email?) and I'll do whatever is best for you. Regards, David. P.S. Eventually I'll get better at these patches , alas the day job keeps me too busy - and I forget the rules :(
Hi David, Am Dienstag, 12. März 2019, 21:34:00 CET schrieb David Summers: > On 12/03/2019 14:22, Heiko Stuebner wrote: > > Hi Jonas, > > > > Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: > >> The Problem: > >> > >> On ASUS Tinker Board S, when booting from the eMMC, and there is card > >> in the sd slot, there are constant errors. > >> > >> Also when warm reboot, uboot can not access the sd slot > >> > >> Cause: > >> > >> Identified by Robin Murphy @ ARM. The Card Detect on rk3288 > >> devices is pulled up by vccio-sd; so when the regulator powers this > >> off, card detect gives spurious errors. A second problem, is during > >> power down, vccio-sd apprears to be powered down. This causes a > >> problem when warm rebooting from the sd card. This was identified by > >> Jonas Karlman. > >> > >> History: > >> > >> A common fault on these rk3288 board, which impliment the reference > >> design. > >> > >> When this arose before: > >> > >> http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153. > >> html > >> > >> And Ulf and Jaehoon clearly said this was a broken card detect design, > >> which should be solved via polling > >> > >> Solution: > >> > >> Hence broken-cd is set as a property. This cures the errors. The > >> powering down of vccio-sd during reboot is cured by adding > >> regulator-boot-on. > >> > >> This solutions has been fairly widely reviewed and tested. > >> > >> Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> > >> Reviewed by: Robin Murphy <robin.murphy@arm.com> > >> Reviewed by: Jonas Karlman <jonas@kwiboo.se> > > > > in v2, you gave a > > Tested-by: Jonas Karlman <jonas@kwiboo.se> > > it seems, which somehow transformed into a reviewed-by in v3. > > > > V3 did change a bit in its approach - for the better, but do you approve > > of adding that Reviewed-by above (and maybe also providing a Tested-by > > for the new approach)? > > > > Thanks > > Heiko > > Hi all, > > Yes apologies for these signed-off lines etc wrong. Its still a learning > experiment for me! > > Both Robin and Jonas added a lot to Patch, as explained in the write up. > Robin came up with the cause of the fault, and Jonas noticed a second > fault. So the question is to give these two their due. I think what they > both added was important. > > Thanks Robin for the official "Reviewed-by". > > Jonas tested the second version, which kept power on vccio-sd all the > time, whereas v1 and v3 cured the problem with broken-cd. So although > both fixed the problem, they way performed is different. So don't think > I can say Jonas Tested-by v3. We'll just give Jonas a bit more time to reply then :-) . > As for the tested by "TheSaint @ ArchLinux Arm", he actually tested just > above every single version of the patch, and a whole lot more before it > was sent in - he is our ASUS Tinker Board user on arch. He doesn't > really want his email in the open, but if push comes to shove he'll > allow it. Question is do you need this for the Patch to pass? No, that is not necessary. I'll just drop the line when applying. In general for a tag to mean something we need the Name, but any tags other than Signed-off-by are optional anyway. > Heiko, I'm happy to change the sign off etc, and resent the patch (as > v4). But would be the same patch. Think the question is what is best for > you. Don't want to send unneeded email ... anyway let me know (private > email?) and I'll do whatever is best for you. No, re-sending just to correct tags is not necessary in _my_ case as I generally just fix up the relevant lines myself when applying. Just keep in mind, other maintainers may handle this differently :-) . > P.S. Eventually I'll get better at these patches , alas the day job > keeps me too busy - and I forget the rules :( Really no problem, it's always a learning experience. So don't worry too much :-) . Heiko
On 2019-03-12 21:52, Heiko Stübner wrote: > Hi David, > > Am Dienstag, 12. März 2019, 21:34:00 CET schrieb David Summers: >> On 12/03/2019 14:22, Heiko Stuebner wrote: >>> Hi Jonas, >>> >>> Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: >>>> The Problem: >>>> >>>> On ASUS Tinker Board S, when booting from the eMMC, and there is card >>>> in the sd slot, there are constant errors. >>>> >>>> Also when warm reboot, uboot can not access the sd slot >>>> >>>> Cause: >>>> >>>> Identified by Robin Murphy @ ARM. The Card Detect on rk3288 >>>> devices is pulled up by vccio-sd; so when the regulator powers this >>>> off, card detect gives spurious errors. A second problem, is during >>>> power down, vccio-sd apprears to be powered down. This causes a >>>> problem when warm rebooting from the sd card. This was identified by >>>> Jonas Karlman. >>>> >>>> History: >>>> >>>> A common fault on these rk3288 board, which impliment the reference >>>> design. >>>> >>>> When this arose before: >>>> >>>> http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153. >>>> html >>>> >>>> And Ulf and Jaehoon clearly said this was a broken card detect design, >>>> which should be solved via polling >>>> >>>> Solution: >>>> >>>> Hence broken-cd is set as a property. This cures the errors. The >>>> powering down of vccio-sd during reboot is cured by adding >>>> regulator-boot-on. >>>> >>>> This solutions has been fairly widely reviewed and tested. >>>> >>>> Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> >>>> Reviewed by: Robin Murphy <robin.murphy@arm.com> >>>> Reviewed by: Jonas Karlman <jonas@kwiboo.se> >>> in v2, you gave a >>> Tested-by: Jonas Karlman <jonas@kwiboo.se> >>> it seems, which somehow transformed into a reviewed-by in v3. >>> >>> V3 did change a bit in its approach - for the better, but do you approve >>> of adding that Reviewed-by above (and maybe also providing a Tested-by >>> for the new approach)? >>> >>> Thanks >>> Heiko >> Hi all, >> >> Yes apologies for these signed-off lines etc wrong. Its still a learning >> experiment for me! >> >> Both Robin and Jonas added a lot to Patch, as explained in the write up. >> Robin came up with the cause of the fault, and Jonas noticed a second >> fault. So the question is to give these two their due. I think what they >> both added was important. >> >> Thanks Robin for the official "Reviewed-by". >> >> Jonas tested the second version, which kept power on vccio-sd all the >> time, whereas v1 and v3 cured the problem with broken-cd. So although >> both fixed the problem, they way performed is different. So don't think >> I can say Jonas Tested-by v3. > We'll just give Jonas a bit more time to reply then :-) . I have now tested v3 on a Tinker Board and a Tinker Board S using u-boot 2019.01 on sd-card, both devices can now soft reboot when running bootloader from sd-card thanks to the "regulator-boot-on". I am happy to add both tags :-) Reviewed-by: Jonas Karlman <jonas@kwiboo.se> Tested-by: Jonas Karlman <jonas@kwiboo.se> Regards, Jonas > >> As for the tested by "TheSaint @ ArchLinux Arm", he actually tested just >> above every single version of the patch, and a whole lot more before it >> was sent in - he is our ASUS Tinker Board user on arch. He doesn't >> really want his email in the open, but if push comes to shove he'll >> allow it. Question is do you need this for the Patch to pass? > No, that is not necessary. I'll just drop the line when applying. > In general for a tag to mean something we need the Name, but any > tags other than Signed-off-by are optional anyway. > > >> Heiko, I'm happy to change the sign off etc, and resent the patch (as >> v4). But would be the same patch. Think the question is what is best for >> you. Don't want to send unneeded email ... anyway let me know (private >> email?) and I'll do whatever is best for you. > No, re-sending just to correct tags is not necessary in _my_ case > as I generally just fix up the relevant lines myself when applying. > Just keep in mind, other maintainers may handle this differently :-) . > > >> P.S. Eventually I'll get better at these patches , alas the day job >> keeps me too busy - and I forget the rules :( > Really no problem, it's always a learning experience. > So don't worry too much :-) . > > > Heiko > >
Am Samstag, 9. März 2019, 16:39:21 CET schrieb David Summers: > The Problem: > > On ASUS Tinker Board S, when booting from the eMMC, and there is card > in the sd slot, there are constant errors. > > Also when warm reboot, uboot can not access the sd slot > > Cause: > > Identified by Robin Murphy @ ARM. The Card Detect on rk3288 > devices is pulled up by vccio-sd; so when the regulator powers this > off, card detect gives spurious errors. A second problem, is during > power down, vccio-sd apprears to be powered down. This causes a > problem when warm rebooting from the sd card. This was identified by > Jonas Karlman. > > History: > > A common fault on these rk3288 board, which impliment the reference > design. > > When this arose before: > > http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153.htm > l > > And Ulf and Jaehoon clearly said this was a broken card detect design, > which should be solved via polling > > Solution: > > Hence broken-cd is set as a property. This cures the errors. The > powering down of vccio-sd during reboot is cured by adding > regulator-boot-on. > > This solutions has been fairly widely reviewed and tested. > > Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> > Reviewed by: Robin Murphy <robin.murphy@arm.com> > Reviewed by: Jonas Karlman <jonas@kwiboo.se> > Test by: TheSaint @ ArchLinux Arm applied as fix for 5.1 with the following modifications: - new subject: "ARM: dts: rockchip: Fix SD card detection on rk3288-tinker" to match the subsystem - adapted Reviewed and Tested tags as collected from involved parties - added Fixes and Cc-stable tags (pointing to the dtsi-split only to ease load of stable maintainers, as hopefully most people will use somewhat recent kernels on sbcs] Heiko
diff --git a/arch/arm/boot/dts/rk3288-tinker.dtsi b/arch/arm/boot/dts/rk3288-tinker.dtsi index aa107ee41b8b..ef653c3209bc 100644 --- a/arch/arm/boot/dts/rk3288-tinker.dtsi +++ b/arch/arm/boot/dts/rk3288-tinker.dtsi @@ -254,6 +254,7 @@ }; vccio_sd: LDO_REG5 { + regulator-boot-on; regulator-min-microvolt = <1800000>; regulator-max-microvolt = <3300000>; regulator-name = "vccio_sd"; @@ -430,7 +431,7 @@ bus-width = <4>; cap-mmc-highspeed; cap-sd-highspeed; - card-detect-delay = <200>; + broken-cd; disable-wp; /* wp not hooked up */ pinctrl-names = "default"; pinctrl-0 = <&sdmmc_clk &sdmmc_cmd &sdmmc_cd &sdmmc_bus4>;
The Problem: On ASUS Tinker Board S, when booting from the eMMC, and there is card in the sd slot, there are constant errors. Also when warm reboot, uboot can not access the sd slot Cause: Identified by Robin Murphy @ ARM. The Card Detect on rk3288 devices is pulled up by vccio-sd; so when the regulator powers this off, card detect gives spurious errors. A second problem, is during power down, vccio-sd apprears to be powered down. This causes a problem when warm rebooting from the sd card. This was identified by Jonas Karlman. History: A common fault on these rk3288 board, which impliment the reference design. When this arose before: http://lists.infradead.org/pipermail/linux-arm-kernel/2014-August/281153.html And Ulf and Jaehoon clearly said this was a broken card detect design, which should be solved via polling Solution: Hence broken-cd is set as a property. This cures the errors. The powering down of vccio-sd during reboot is cured by adding regulator-boot-on. This solutions has been fairly widely reviewed and tested. Signed-off-by: David Summers <beagleboard@davidjohnsummers.uk> Reviewed by: Robin Murphy <robin.murphy@arm.com> Reviewed by: Jonas Karlman <jonas@kwiboo.se> Test by: TheSaint @ ArchLinux Arm --- arch/arm/boot/dts/rk3288-tinker.dtsi | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)