| Message ID | 20230724223057.1208122-1-quic_eberman@quicinc.com (mailing list archive) |
|---|---|
| Headers | show |
| Series | Implement a PSCI SYSTEM_RESET2 reboot-mode driver | expand |
Hello, On 7/24/23 15:30, Elliot Berman wrote: > PSCI implements a restart notifier for architectural defined resets. > The SYSTEM_RESET2 call allows vendor firmware to define additional reset > types which could be mapped to the reboot reason. > > Implement a driver to wire the reboot-mode framework to make vendor > SYSTEM_RESET2 calls on reboot. > > This is a continuation from https://lore.kernel.org/all/4a679542-b48d-7e11-f33a-63535a5c68cb@quicinc.com/ Would appreciate being CC'd on a the non-RFC postings of this patch. FWIW, my use case is better described with this earlier submission: https://lore.kernel.org/lkml/20220122035421.4086618-1-f.fainelli@gmail.com/T/#m74e4243c1af3a8d896e19b573b58f562fa09961d It would be neat if I could leverage your driver in order to implement this custom "reboot powercycle" implementation. Towards that goal, we would likely need to specify the desired reboot "sub" operation alongside its PSCI SYSTEM_RESET2 reboot type argument? Thanks!
On 7/25/2023 12:12 PM, Florian Fainelli wrote: > Hello, > > On 7/24/23 15:30, Elliot Berman wrote: >> PSCI implements a restart notifier for architectural defined resets. >> The SYSTEM_RESET2 call allows vendor firmware to define additional reset >> types which could be mapped to the reboot reason. >> >> Implement a driver to wire the reboot-mode framework to make vendor >> SYSTEM_RESET2 calls on reboot. >> >> This is a continuation from >> https://lore.kernel.org/all/4a679542-b48d-7e11-f33a-63535a5c68cb@quicinc.com/ > > Would appreciate being CC'd on a the non-RFC postings of this patch. > FWIW, my use case is better described with this earlier submission: > > https://lore.kernel.org/lkml/20220122035421.4086618-1-f.fainelli@gmail.com/T/#m74e4243c1af3a8d896e19b573b58f562fa09961d > > It would be neat if I could leverage your driver in order to implement > this custom "reboot powercycle" implementation. Towards that goal, we > would likely need to specify the desired reboot "sub" operation > alongside its PSCI SYSTEM_RESET2 reboot type argument? > > Thanks! I think you you want to describe the PSCI vendor reset under a warm reboot with command "powercycle"? In other words, my series only lets DT describe either reboot_mode (warm) or cmd (powercycle) but not both simultaneously? Please correct me if I got it wrong! Otherwise, I can incorporate way to describe vendor reset type matching both reboot_mode and cmd in the DT. - Elliot
On 7/25/23 13:27, Elliot Berman wrote: > > > On 7/25/2023 12:12 PM, Florian Fainelli wrote: >> Hello, >> >> On 7/24/23 15:30, Elliot Berman wrote: >>> PSCI implements a restart notifier for architectural defined resets. >>> The SYSTEM_RESET2 call allows vendor firmware to define additional reset >>> types which could be mapped to the reboot reason. >>> >>> Implement a driver to wire the reboot-mode framework to make vendor >>> SYSTEM_RESET2 calls on reboot. >>> >>> This is a continuation from >>> https://lore.kernel.org/all/4a679542-b48d-7e11-f33a-63535a5c68cb@quicinc.com/ >> >> Would appreciate being CC'd on a the non-RFC postings of this patch. >> FWIW, my use case is better described with this earlier submission: >> >> https://lore.kernel.org/lkml/20220122035421.4086618-1-f.fainelli@gmail.com/T/#m74e4243c1af3a8d896e19b573b58f562fa09961d >> >> It would be neat if I could leverage your driver in order to implement >> this custom "reboot powercycle" implementation. Towards that goal, we >> would likely need to specify the desired reboot "sub" operation >> alongside its PSCI SYSTEM_RESET2 reboot type argument? >> >> Thanks! > > I think you you want to describe the PSCI vendor reset under a warm > reboot with command "powercycle"? In other words, my series only lets DT > describe either reboot_mode (warm) or cmd (powercycle) but not both > simultaneously? I did not give a lot of thoughts into the different types of reboot (warm, soft, cold) and just went with an extension of whichever reboot type we have to be supplemented by the "powercycle" command. It seems like we should support both the reboot type and command, it would be fine with me if I had to specify reboot_mode + cmd for each reboot_mode.