Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On 31/05/2013 01:17, Douglas Gilbert : On 13-05-30 03:36 PM, Andrew Morton wrote: On Thu, 30 May 2013 09:50:27 +0200 Nicolas Ferre wrote: The review of this patch series was in my TODO list for some time... Today, I magically took time to review it ;-) The patch series is good and I (even if it is too late) here is my: Acked-by: Nicolas Ferre I do not know if the series can be stacked for inclusion in 3.10-rc but the resolution of this bug can help a lot (as Douglas is saying in subsequent email...). We can do that, but looking through the discussion and changelogs I can't seem to find a usable description of what impact the bug (and its fix) have upon end-users. A nicely packaged description of that impact would help grease the wheels, please. How about this: The members of Atmel's at91sam9x5 family (9x5) have a broken RTC interrupt mask register (AT91_RTC_IMR). It does not reflect enabled interrupts but instead always returns zero. The kernel's rtc-at91rm9200 driver handles the RTC for the 9x5 family. Currently when the date/time is set, an interrupt is generated and this driver neglects to handle the interrupt. The kernel complains about the un-handled interrupt and disables it henceforth. This not only breaks the RTC function, but since that interrupt is shared (Atmel's SYS interrupt) then other things break as well (e.g. the debug port no longer accepts characters). Tested on the at91sam9g25. Bug confirmed by Atmel. Absolutely. Thank you Douglas for the detailed description. Edit as you please. Doug Gilbert -- Nicolas Ferre -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On 13-05-30 03:36 PM, Andrew Morton wrote: On Thu, 30 May 2013 09:50:27 +0200 Nicolas Ferre wrote: The review of this patch series was in my TODO list for some time... Today, I magically took time to review it ;-) The patch series is good and I (even if it is too late) here is my: Acked-by: Nicolas Ferre I do not know if the series can be stacked for inclusion in 3.10-rc but the resolution of this bug can help a lot (as Douglas is saying in subsequent email...). We can do that, but looking through the discussion and changelogs I can't seem to find a usable description of what impact the bug (and its fix) have upon end-users. A nicely packaged description of that impact would help grease the wheels, please. How about this: The members of Atmel's at91sam9x5 family (9x5) have a broken RTC interrupt mask register (AT91_RTC_IMR). It does not reflect enabled interrupts but instead always returns zero. The kernel's rtc-at91rm9200 driver handles the RTC for the 9x5 family. Currently when the date/time is set, an interrupt is generated and this driver neglects to handle the interrupt. The kernel complains about the un-handled interrupt and disables it henceforth. This not only breaks the RTC function, but since that interrupt is shared (Atmel's SYS interrupt) then other things break as well (e.g. the debug port no longer accepts characters). Tested on the at91sam9g25. Bug confirmed by Atmel. Edit as you please. Doug Gilbert -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On Thu, 30 May 2013 09:50:27 +0200 Nicolas Ferre wrote: > The review of this patch series was in my TODO list for some time... > > Today, I magically took time to review it ;-) > The patch series is good and I (even if it is too late) here is my: > > Acked-by: Nicolas Ferre > > I do not know if the series can be stacked for inclusion in 3.10-rc but > the resolution of this bug can help a lot (as Douglas is saying in > subsequent email...). We can do that, but looking through the discussion and changelogs I can't seem to find a usable description of what impact the bug (and its fix) have upon end-users. A nicely packaged description of that impact would help grease the wheels, please. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On 30/05/2013 01:22, Douglas Gilbert : On 13-05-29 04:41 PM, Robert Nelson wrote: On Wed, May 29, 2013 at 3:33 PM, Andrew Morton wrote: On Thu, 23 May 2013 10:38:50 +0200 Johan Hovold wrote: This is an update of the shadow-interrupt-mask series against v3.10-rc2. I guess we need Atmel to confirm that all sam9x5 SoCs are indeed affected. If not, then some probing mechanism as the one Doug suggested could be implemented on top of (a subset of) these patches. What do you say, Nicolas? Note that the first patch (adding a missing OF compile guard) could be applied straight away. At this stage it is unclear to me how to proceed with patches 2-5. fyi: A version of these patches had been applied once before: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=0ef1594c017521ea89278e80fe3f80dafb17abde But due to a few issues it was later reverted: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=e24b0bfa2f0446ffaad2661040be23668133aef8 Strange life of a patch. Mine was the original, Johan Hovold objected and had it reverted. Johan then presented his first patch then v2. They got lost in the weeds. No, they were not lost. No patch is ever lost and this thread is the proof. My hardware was still broken and this bug caused collateral damage. My original patch no longer applied to lk 3.10.0-rc1 so I rewrote it, borrowing some of Johan's ideas and doing a probe time check for the broken RTC_IMR. That patch was presented about a week ago: http://marc.info/?l=linux-arm-kernel&m=136917492531478&w=2 The top of that post gives some more background. That prompted Johan to produce v3 of his patch which is the subject of this thread. I was hoping that Nicolas Ferre would comment or ack one of these patches. Still waiting. Sure that all this did not progressed at the speed you expected. I understand that. But even if I did not answered in a timely manner, that does not mean that I didn't considered it and marked it as "things to be done before next merge window"... So, today, too late, I gave my "Acked-by". Sorry for the delay. Let's still monitor the progress of this series upstream. I have a copy of the original, publicly released manual for the at91sam9g25 (a member of the at91sam9x5 family) marked "11032A–ATARM–27-Jul-11". It contains the following: Errata 49.3.1 RTC: Interrupt Mask Register cannot be used Interrupt Mask Register reading always returns 0. Both Rev B and Rev C of that manual drop that particular erratum. My g25 SoC-based subsystems come from an Atmel partner and still have the RTC IMR bug. We already talked about this Douglas. Why are you saying this again. So, to summarize: 1/ each and every at91sam9x5 family SoC have and will probably always have this IMR bug (including 9g25 which is part of the family). 2/ you kindly reported the errata disappearing in the documentation. It is an error with document appearance which you probably noted. I have made the necessary actions to correct this. But here again, you have to be patient waiting for the datasheet's next revision. Best regards, -- Nicolas Ferre -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On 29/05/2013 22:41, Robert Nelson : On Wed, May 29, 2013 at 3:33 PM, Andrew Morton wrote: On Thu, 23 May 2013 10:38:50 +0200 Johan Hovold wrote: This is an update of the shadow-interrupt-mask series against v3.10-rc2. I guess we need Atmel to confirm that all sam9x5 SoCs are indeed affected. If not, then some probing mechanism as the one Doug suggested could be implemented on top of (a subset of) these patches. What do you say, Nicolas? Note that the first patch (adding a missing OF compile guard) could be applied straight away. At this stage it is unclear to me how to proceed with patches 2-5. fyi: A version of these patches had been applied once before: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=0ef1594c017521ea89278e80fe3f80dafb17abde But due to a few issues it was later reverted: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=e24b0bfa2f0446ffaad2661040be23668133aef8 This new revision doesn't have the issue encountered by first version. Andrew, The review of this patch series was in my TODO list for some time... Today, I magically took time to review it ;-) The patch series is good and I (even if it is too late) here is my: Acked-by: Nicolas Ferre I do not know if the series can be stacked for inclusion in 3.10-rc but the resolution of this bug can help a lot (as Douglas is saying in subsequent email...). Best regards, -- Nicolas Ferre -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On 23/05/2013 10:38, Johan Hovold : This is an update of the shadow-interrupt-mask series against v3.10-rc2. I guess we need Atmel to confirm that all sam9x5 SoCs are indeed affected. If not, then some probing mechanism as the one Doug suggested could be implemented on top of (a subset of) these patches. What do you say, Nicolas? No probing mechanism is needed: only sam9x5 are affected. Note that the first patch (adding a missing OF compile guard) could be applied straight away. Thanks, Johan v3: - rebase against v3.10-rc2 - remove some comments Thanks for rebasing this series. Best regards, v2: - rebase on top of DT-support patch by Joachim Eastwood - add missing brace in DT-id table Johan Hovold (5): rtc-at91rm9200: add match-table compile guard rtc-at91rm9200: add configuration support rtc-at91rm9200: refactor interrupt-register handling rtc-at91rm9200: add shadow interrupt mask rtc-at91rm9200: use shadow IMR on at91sam9x5 .../bindings/rtc/atmel,at91rm9200-rtc.txt | 2 +- arch/arm/boot/dts/at91sam9x5.dtsi | 2 +- drivers/rtc/rtc-at91rm9200.c | 131 + 3 files changed, 113 insertions(+), 22 deletions(-) -- Nicolas Ferre -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On 13-05-29 04:41 PM, Robert Nelson wrote: On Wed, May 29, 2013 at 3:33 PM, Andrew Morton wrote: On Thu, 23 May 2013 10:38:50 +0200 Johan Hovold wrote: This is an update of the shadow-interrupt-mask series against v3.10-rc2. I guess we need Atmel to confirm that all sam9x5 SoCs are indeed affected. If not, then some probing mechanism as the one Doug suggested could be implemented on top of (a subset of) these patches. What do you say, Nicolas? Note that the first patch (adding a missing OF compile guard) could be applied straight away. At this stage it is unclear to me how to proceed with patches 2-5. fyi: A version of these patches had been applied once before: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=0ef1594c017521ea89278e80fe3f80dafb17abde But due to a few issues it was later reverted: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=e24b0bfa2f0446ffaad2661040be23668133aef8 Strange life of a patch. Mine was the original, Johan Hovold objected and had it reverted. Johan then presented his first patch then v2. They got lost in the weeds. My hardware was still broken and this bug caused collateral damage. My original patch no longer applied to lk 3.10.0-rc1 so I rewrote it, borrowing some of Johan's ideas and doing a probe time check for the broken RTC_IMR. That patch was presented about a week ago: http://marc.info/?l=linux-arm-kernel&m=136917492531478&w=2 The top of that post gives some more background. That prompted Johan to produce v3 of his patch which is the subject of this thread. I was hoping that Nicolas Ferre would comment or ack one of these patches. Still waiting. I have a copy of the original, publicly released manual for the at91sam9g25 (a member of the at91sam9x5 family) marked "11032A–ATARM–27-Jul-11". It contains the following: Errata 49.3.1 RTC: Interrupt Mask Register cannot be used Interrupt Mask Register reading always returns 0. Both Rev B and Rev C of that manual drop that particular erratum. My g25 SoC-based subsystems come from an Atmel partner and still have the RTC IMR bug. Doug Gilbert -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On Wed, May 29, 2013 at 3:33 PM, Andrew Morton wrote: > On Thu, 23 May 2013 10:38:50 +0200 Johan Hovold wrote: > >> This is an update of the shadow-interrupt-mask series against v3.10-rc2. >> >> I guess we need Atmel to confirm that all sam9x5 SoCs are indeed >> affected. If not, then some probing mechanism as the one Doug suggested >> could be implemented on top of (a subset of) these patches. What do you >> say, Nicolas? >> >> Note that the first patch (adding a missing OF compile guard) could be >> applied straight away. > > At this stage it is unclear to me how to proceed with patches 2-5. fyi: A version of these patches had been applied once before: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=0ef1594c017521ea89278e80fe3f80dafb17abde But due to a few issues it was later reverted: https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=e24b0bfa2f0446ffaad2661040be23668133aef8 Regards, -- Robert Nelson http://www.rcn-ee.com/ -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
On Thu, 23 May 2013 10:38:50 +0200 Johan Hovold wrote: > This is an update of the shadow-interrupt-mask series against v3.10-rc2. > > I guess we need Atmel to confirm that all sam9x5 SoCs are indeed > affected. If not, then some probing mechanism as the one Doug suggested > could be implemented on top of (a subset of) these patches. What do you > say, Nicolas? > > Note that the first patch (adding a missing OF compile guard) could be > applied straight away. At this stage it is unclear to me how to proceed with patches 2-5. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[PATCH v3 0/5] rtc-at91rm9200: add shadow interrupt mask
This is an update of the shadow-interrupt-mask series against v3.10-rc2. I guess we need Atmel to confirm that all sam9x5 SoCs are indeed affected. If not, then some probing mechanism as the one Doug suggested could be implemented on top of (a subset of) these patches. What do you say, Nicolas? Note that the first patch (adding a missing OF compile guard) could be applied straight away. Thanks, Johan v3: - rebase against v3.10-rc2 - remove some comments v2: - rebase on top of DT-support patch by Joachim Eastwood - add missing brace in DT-id table Johan Hovold (5): rtc-at91rm9200: add match-table compile guard rtc-at91rm9200: add configuration support rtc-at91rm9200: refactor interrupt-register handling rtc-at91rm9200: add shadow interrupt mask rtc-at91rm9200: use shadow IMR on at91sam9x5 .../bindings/rtc/atmel,at91rm9200-rtc.txt | 2 +- arch/arm/boot/dts/at91sam9x5.dtsi | 2 +- drivers/rtc/rtc-at91rm9200.c | 131 + 3 files changed, 113 insertions(+), 22 deletions(-) -- 1.8.2.1 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
