openbmc.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* 12v slot AC Support  in x86-power-control
@ 2020-10-06 16:43 P. Priyatharshan
  2020-10-07  1:51 ` Patrick Williams
  2020-10-08  0:41 ` Vijay Khemka
  0 siblings, 2 replies; 10+ messages in thread
From: P. Priyatharshan @ 2020-10-06 16:43 UTC (permalink / raw)
  To: openbmc, Vijay Khemka, ed; +Cc: Velumani T-ERS,HCLTech

[-- Attachment #1: Type: text/plain, Size: 2634 bytes --]

Hi Vijay,

I am working on a patch to support  12v slot AC cycle in x86-power-control.This can be enabled by configuring slotACPowerName in json config file.Since you are also working on similar kind of feature in x86-power-control, I would like bring up this discussion on mailing list to understand how these two features, and the design you proposed for chassis reset are intended to work together.Kindly go through my design changes below and share your thoughts.

I have modified the existing /xyz/openbmc_project/state/chassis0 object to implement 12V On, Off, and Power Cycle for system which has 12V Slot power support.


If slotACPowerName is disabled and if,

RequestedPowerTransition = "xyz.openbmc_project.State.Chassis.Transition.Off" then poweroff [existing functionality not changed]

RequestedPowerTransition = "xyz.openbmc_project.State.Chassis.Transition.On" then powerOn [existing functionality not changed]

RequestedPowerTransition = "xyz.openbmc_project.State.Chassis.Transition.PowerCycle" then PowerCycle [existing functionality not changed]

events will be called.



If slotACPowerName is enabled and if,

RequestedPowerTransition = "xyz.openbmc_project.State.Chassis.Transition.Off" then 12V Ac Slot Off

RequestedPowerTransition = "xyz.openbmc_project.State.Chassis.Transition.On" then 12V Ac Slot On

RequestedPowerTransition = "xyz.openbmc_project.State.Chassis.Transition.PowerCycle" then 12V Ac Slot PowerCycle

events will be called.

Thanks,
Priyatharshan P
::DISCLAIMER::
________________________________
The contents of this e-mail and any attachment(s) are confidential and intended for the named recipient(s) only. E-mail transmission is not guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or may contain viruses in transmission. The e mail and its contents (with or without referred errors) shall therefore not attach any liability on the originator or HCL or its affiliates. Views or opinions, if any, presented in this email are solely those of the author and may not necessarily reflect the views or opinions of HCL or its affiliates. Any form of reproduction, dissemination, copying, disclosure, modification, distribution and / or publication of this message without the prior written consent of authorized representative of HCL is strictly prohibited. If you have received this email in error please delete it and notify the sender immediately. Before opening any email and/or attachments, please check them for viruses and other defects.
________________________________

[-- Attachment #2: Type: text/html, Size: 3815 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread
* Re: 12v slot AC Support  in x86-power-control
@ 2020-10-09 15:49 P. Priyatharshan
  2020-10-09 19:48 ` Patrick Williams
  0 siblings, 1 reply; 10+ messages in thread
From: P. Priyatharshan @ 2020-10-09 15:49 UTC (permalink / raw)
  To: Patrick Williams; +Cc: openbmc, ed, Velumani T-ERS,HCLTech, Vijay Khemka

[-- Attachment #1: Type: text/plain, Size: 5480 bytes --]

Hi Patrick,

Please find my response below.

>Priyatharshan,

>I'm still very confused by the word "AC" here.  I know your server
>doesn't have a single joule of AC power on any voltage rail.  This
>server plugs into an OCP rack with a 12V DC bus-bar powering all the
>servers in the rack.

Sorry for confusing with the word "Ac" . It is actually 12V On/Off.

>Vijay implemented, for Tiogapass, the 12V system power cycle.  This,
>in hardware, uses the HSC* to remove the server from the bus-bar, which
>power cycles the entire server including the BMC.  This is as if someone
>physically pulled the server from the rack and then pushed it back in
>but implemented via the HSC.

Agree, Vijay Implementation power cycles the entire server including the BMC.In Yv2(multi-host-system) for each slot 12V_STBY is connected via hot swap ADM1172.This ADM1172 power will be controlled by dedicated GPIO.So my functionality will implement power cycles only to the slot.This is as if someone physically pulled host from the slot and then pushed it back.

>On the server you are working on, we should have that same design for
>the "full system power cycle" which interacts with the HSC.  This should
>be done the same way as what Vijay implemented.

Agree.I also used Vijay's implementation and implemented "full system power cycle" for Yv2.
Gerrit Link : https://gerrit.openbmc-project.xyz/c/openbmc/meta-facebook/+/37065


>You also have power control to the multi-host cards which feed 12V to
>each of the compute cards in your chassis.  What I don't understand is
>why does this also need control of the full 12V and why is it separate
>from the normal "power on the compute card" sequence?  This "power on
>the compute card" sequence is what is suppose to be modelled by the
>'chassisN' objects referenced by others.

>Can you expand on what the difference is and what functionality you are
>enabling by having two different slot-level power controls?

The earlier "power on/off/cycle the computer card" sequence will only handle the computer card portion of the host.
But the new feature [Slot on/off/cycle] will power cycles the entire card [ BIC + computer card portion + rest of the hardware in twin lake] for each slot.

Thanks,
Priyatharshan P

________________________________
From: Patrick Williams
Sent: Thursday, October 08, 2020 22:01
To: P. Priyatharshan
Cc: Vijay Khemka; openbmc@lists.ozlabs.org; ed@tanous.net; Velumani T-ERS,HCLTech
Subject: Re: 12v slot AC Support in x86-power-control

On Thu, Oct 08, 2020 at 01:13:49PM +0000, P. Priyatharshan wrote:
> Out platform is multi host system and we  have the host as FRU and it can be plugged in to a slot. The slot has an option to control slot power and this can be used for a AC cycle.

Priyatharshan,

I'm still very confused by the word "AC" here.  I know your server
doesn't have a single joule of AC power on any voltage rail.  This
server plugs into an OCP rack with a 12V DC bus-bar powering all the
servers in the rack.

Vijay implemented, for Tiogapass, the 12V system power cycle.  This,
in hardware, uses the HSC* to remove the server from the bus-bar, which
power cycles the entire server including the BMC.  This is as if someone
physically pulled the server from the rack and then pushed it back in
but implemented via the HSC.

On the server you are working on, we should have that same design for
the "full system power cycle" which interacts with the HSC.  This should
be done the same way as what Vijay implemented.

You also have power control to the multi-host cards which feed 12V to
each of the compute cards in your chassis.  What I don't understand is
why does this also need control of the full 12V and why is it separate
from the normal "power on the compute card" sequence?  This "power on
the compute card" sequence is what is suppose to be modelled by the
'chassisN' objects referenced by others.

Can you expand on what the difference is and what functionality you are
enabling by having two different slot-level power controls?

> We proposed  a design to add the slot power control in the x86-power-control.The earlier discussion can be found in the below link.
>
> https://lists.ozlabs.org/pipermail/openbmc/2020-August/022481.html

* - HSC = Hot Swap Controller.

--
Patrick Williams
::DISCLAIMER::
________________________________
The contents of this e-mail and any attachment(s) are confidential and intended for the named recipient(s) only. E-mail transmission is not guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or may contain viruses in transmission. The e mail and its contents (with or without referred errors) shall therefore not attach any liability on the originator or HCL or its affiliates. Views or opinions, if any, presented in this email are solely those of the author and may not necessarily reflect the views or opinions of HCL or its affiliates. Any form of reproduction, dissemination, copying, disclosure, modification, distribution and / or publication of this message without the prior written consent of authorized representative of HCL is strictly prohibited. If you have received this email in error please delete it and notify the sender immediately. Before opening any email and/or attachments, please check them for viruses and other defects.
________________________________

[-- Attachment #2: Type: text/html, Size: 8135 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread
* Re: 12v slot AC Support  in x86-power-control
@ 2020-10-11  6:39 P. Priyatharshan
  0 siblings, 0 replies; 10+ messages in thread
From: P. Priyatharshan @ 2020-10-11  6:39 UTC (permalink / raw)
  To: Patrick Williams; +Cc: openbmc, ed, Velumani T-ERS,HCLTech, Vijay Khemka

[-- Attachment #1: Type: text/plain, Size: 2591 bytes --]

Hi Patrick,

Thanks for the clarification.Got things cleared now.

As you mentioned, I will implement chassis_system0 object for full power cycle of the
entire system, and 'chassis_systemN' [N = 1:4 for my case] to do the slot-level full power cycle in x86-power-control.

Thanks,
Priyatharshan P

________________________________
From: Patrick Williams
Sent: Saturday, October 10, 2020 01:18
To: P. Priyatharshan
Cc: Vijay Khemka; openbmc@lists.ozlabs.org; ed@tanous.net; Velumani T-ERS,HCLTech
Subject: Re: 12v slot AC Support in x86-power-control

On Fri, Oct 09, 2020 at 03:49:55PM +0000, P. Priyatharshan wrote:

> The earlier "power on/off/cycle the computer card" sequence will only handle the computer card portion of the host.
> But the new feature [Slot on/off/cycle] will power cycles the entire card [ BIC + computer card portion + rest of the hardware in twin lake] for each slot.

Got it now.

Take a look at [1].  You should implement the 'chassisN' objects (in
x86-power-control?) to handle the normal power sequence to each of your
compute cards.  I think in your case N = 1:4.  You should also implement
the 'chassis_system0' object, which will be a full power cycle of the
entire system.  Finally, you'll want to implement 'chassis_systemN'
(again N = 1:4) to do the slot-level full power cycle.

1. https://github.com/openbmc/phosphor-dbus-interfaces/tree/master/xyz/openbmc_project/State#chassis-system

--
Patrick Williams
::DISCLAIMER::
________________________________
The contents of this e-mail and any attachment(s) are confidential and intended for the named recipient(s) only. E-mail transmission is not guaranteed to be secure or error-free as information could be intercepted, corrupted, lost, destroyed, arrive late or incomplete, or may contain viruses in transmission. The e mail and its contents (with or without referred errors) shall therefore not attach any liability on the originator or HCL or its affiliates. Views or opinions, if any, presented in this email are solely those of the author and may not necessarily reflect the views or opinions of HCL or its affiliates. Any form of reproduction, dissemination, copying, disclosure, modification, distribution and / or publication of this message without the prior written consent of authorized representative of HCL is strictly prohibited. If you have received this email in error please delete it and notify the sender immediately. Before opening any email and/or attachments, please check them for viruses and other defects.
________________________________

[-- Attachment #2: Type: text/html, Size: 4659 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2020-10-11  6:41 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-10-06 16:43 12v slot AC Support in x86-power-control P. Priyatharshan
2020-10-07  1:51 ` Patrick Williams
2020-10-08  0:41 ` Vijay Khemka
2020-10-08 13:13   ` P. Priyatharshan
2020-10-08 13:45     ` Anton Kachalov
2020-10-08 14:55       ` P. Priyatharshan
2020-10-08 16:31     ` Patrick Williams
2020-10-09 15:49 P. Priyatharshan
2020-10-09 19:48 ` Patrick Williams
2020-10-11  6:39 P. Priyatharshan

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).