From: Tao Ren <redacted>
The patch series introuces 3 new BMC platforms for the Facebook's next
generation switch platforms: Cloudripper, Elbert and Fuji.
Patch #1 adds dtsi file to describe the fixed flash layout of 128MB mtd
device. The flash layout is used by all the 3 AST2600 BMC as well as
Wedge400 (AST2500).
Patch #2 simplies Wedge400 dts by including the flash layout dtsi.
Patch #3 adds common dtsi which is included by all the new Facebook
AST2600 Network BMC platforms.
Patch #4 adds the device tree for Facebook Cloudripper (AST2600) BMC.
Patch #5 adds the device tree for Facebook Elbert (AST2600) BMC.
Patch #6 adds the device tree for Facebook Fuji (AST2600) BMC.
Tao Ren (6):
ARM: dts: Add Facebook BMC 128MB flash layout
ARM: dts: aspeed: wedge400: Use common flash layout
ARM: dts: aspeed: Common dtsi for Facebook AST2600 Network BMCs
ARM: dts: aspeed: Add Facebook Cloudripper (AST2600) BMC
ARM: dts: aspeed: Add Facebook Elbert (AST2600) BMC
ARM: dts: aspeed: Add Facebook Fuji (AST2600) BMC
arch/arm/boot/dts/Makefile | 3 +
.../dts/aspeed-bmc-facebook-cloudripper.dts | 539 +++++++
.../boot/dts/aspeed-bmc-facebook-elbert.dts | 185 +++
.../arm/boot/dts/aspeed-bmc-facebook-fuji.dts | 1251 +++++++++++++++++
.../boot/dts/aspeed-bmc-facebook-wedge400.dts | 48 +-
.../dts/ast2600-facebook-netbmc-common.dtsi | 169 +++
.../dts/facebook-bmc-flash-layout-128.dtsi | 60 +
7 files changed, 2208 insertions(+), 47 deletions(-)
create mode 100644 arch/arm/boot/dts/aspeed-bmc-facebook-cloudripper.dts
create mode 100644 arch/arm/boot/dts/aspeed-bmc-facebook-elbert.dts
create mode 100644 arch/arm/boot/dts/aspeed-bmc-facebook-fuji.dts
create mode 100644 arch/arm/boot/dts/ast2600-facebook-netbmc-common.dtsi
create mode 100644 arch/arm/boot/dts/facebook-bmc-flash-layout-128.dtsi
--
2.17.1
From: Tao Ren <redacted>
This is the layout used by Facebook BMC systems. It describes the fixed
flash layout of a 128MB mtd device.
Signed-off-by: Tao Ren <redacted>
---
Changes in v2:
- None.
.../dts/facebook-bmc-flash-layout-128.dtsi | 60 +++++++++++++++++++
1 file changed, 60 insertions(+)
create mode 100644 arch/arm/boot/dts/facebook-bmc-flash-layout-128.dtsi
From: Tao Ren <redacted>
Simplify wedge400 flash layout by using the common layout defined in
"facebook-bmc-flash-layout-128.dtsi".
Signed-off-by: Tao Ren <redacted>
---
Changes in v2:
- None.
.../boot/dts/aspeed-bmc-facebook-wedge400.dts | 48 +------------------
1 file changed, 1 insertion(+), 47 deletions(-)
From: Tao Ren <redacted>
This common descirption is included by all Facebook AST2600 Network BMC
platforms to minimize duplicated device entries across Facebook Network
BMC device trees.
Signed-off-by: Tao Ren <redacted>
---
Changes in v2:
- "mac3" is removed from the dtsi (moved to platform dts files).
.../dts/ast2600-facebook-netbmc-common.dtsi | 169 ++++++++++++++++++
1 file changed, 169 insertions(+)
create mode 100644 arch/arm/boot/dts/ast2600-facebook-netbmc-common.dtsi
From: Tao Ren <redacted>
Add initial version of device tree for Facebook Cloudripper (AST2600) BMC.
Cloudripper is Facebook's next generation switch platform with an AST2600
BMC integrated for health monitoring purpose.
Signed-off-by: Tao Ren <redacted>
---
Changes in v2:
- added more notes about the new platform in patch description.
- "spi1" was removed as the driver is not ready.
- "mac3" properties were updated with pinctrl, and etc.
arch/arm/boot/dts/Makefile | 1 +
.../dts/aspeed-bmc-facebook-cloudripper.dts | 539 ++++++++++++++++++
2 files changed, 540 insertions(+)
create mode 100644 arch/arm/boot/dts/aspeed-bmc-facebook-cloudripper.dts
From: Tao Ren <redacted>
Add initial version of device tree for Facebook Elbert (AST2600) BMC.
Elbert is Facebook's next generation switch platform with an AST2600 BMC
integrated for health monitoring purpose.
Signed-off-by: Tao Ren <redacted>
---
Changes in v2:
- added more notes about the new platform in patch description.
- "mac3" was disabled (because it was removed from Facebook AST2600
NetBMC common dtsi). Will add "mac3" back when "fixed-link" is
supported in the MAC driver.
arch/arm/boot/dts/Makefile | 1 +
.../boot/dts/aspeed-bmc-facebook-elbert.dts | 185 ++++++++++++++++++
2 files changed, 186 insertions(+)
create mode 100644 arch/arm/boot/dts/aspeed-bmc-facebook-elbert.dts
From: Tao Ren <redacted>
Add initial version of device tree for Facebook Fuji (AST2600) BMC.
Fuji is Facebook's next generation switch platform with an AST2600 BMC
integrated for health monitoring purpose.
Signed-off-by: Tao Ren <redacted>
---
Changes in v2:
- added more notes about the new platform in patch description.
- "spi1" was removed as the driver is not ready.
- "mac3" properties were updated with pinctrl, and etc.
arch/arm/boot/dts/Makefile | 1 +
.../arm/boot/dts/aspeed-bmc-facebook-fuji.dts | 1251 +++++++++++++++++
2 files changed, 1252 insertions(+)
create mode 100644 arch/arm/boot/dts/aspeed-bmc-facebook-fuji.dts
On Fri, Aug 13, 2021 at 01:29:17AM +0000, Joel Stanley wrote:
On Thu, 5 Aug 2021 at 22:28, [off-list ref] wrote:
quoted
+&mdio1 {
+ status = "okay";
You're enabling this but it looks like it's unused?
Thanks Joel for the careful review. The MDIO controller is not paired
with BMC MAC; instead, it's connected to the MDC/MDIO interface of the
on-board switch (whose ports are connected to BMC, Host and front panel
management port).
In other word, the MDIO controller is used, but we do need some user
space scripts to control the controller. What's your suggestion? For
example, should I add some comments to explain the purpose?
Cheers,
Tao
From: Joel Stanley <joel@jms.id.au> Date: 2021-08-13 04:13:39
On Fri, 13 Aug 2021 at 03:40, Tao Ren [off-list ref] wrote:
On Fri, Aug 13, 2021 at 01:29:17AM +0000, Joel Stanley wrote:
quoted
On Thu, 5 Aug 2021 at 22:28, [off-list ref] wrote:
quoted
+&mdio1 {
+ status = "okay";
You're enabling this but it looks like it's unused?
Thanks Joel for the careful review. The MDIO controller is not paired
with BMC MAC; instead, it's connected to the MDC/MDIO interface of the
on-board switch (whose ports are connected to BMC, Host and front panel
management port).
I see!
In other word, the MDIO controller is used, but we do need some user
space scripts to control the controller. What's your suggestion? For
example, should I add some comments to explain the purpose?
From: Andrew Lunn <andrew@lunn.ch> Date: 2021-08-13 14:46:14
On Thu, Aug 12, 2021 at 08:40:17PM -0700, Tao Ren wrote:
On Fri, Aug 13, 2021 at 01:29:17AM +0000, Joel Stanley wrote:
quoted
On Thu, 5 Aug 2021 at 22:28, [off-list ref] wrote:
quoted
+&mdio1 {
+ status = "okay";
You're enabling this but it looks like it's unused?
Thanks Joel for the careful review. The MDIO controller is not paired
with BMC MAC; instead, it's connected to the MDC/MDIO interface of the
on-board switch (whose ports are connected to BMC, Host and front panel
management port).
What switch is it? Is there a DSA driver for it? drivers/net/dsa/*
Ideally you want Linux to be controlling the switch, in the standard
linux way.
Andrew
Hi Andrew,
On Fri, Aug 13, 2021 at 04:45:58PM +0200, Andrew Lunn wrote:
On Thu, Aug 12, 2021 at 08:40:17PM -0700, Tao Ren wrote:
quoted
On Fri, Aug 13, 2021 at 01:29:17AM +0000, Joel Stanley wrote:
quoted
On Thu, 5 Aug 2021 at 22:28, [off-list ref] wrote:
quoted
+&mdio1 {
+ status = "okay";
You're enabling this but it looks like it's unused?
Thanks Joel for the careful review. The MDIO controller is not paired
with BMC MAC; instead, it's connected to the MDC/MDIO interface of the
on-board switch (whose ports are connected to BMC, Host and front panel
management port).
What switch is it? Is there a DSA driver for it? drivers/net/dsa/*
Ideally you want Linux to be controlling the switch, in the standard
linux way.
Andrew
Thanks for jumping in. We are using BCM5389 and the MDIO bus is used to
access BCM5389 MDC/MDIO interface in Pseudo-PHY mode.
I didn't know drivers/net/dsa, but let me check out the drivers and see
if it works in the Cloudripper environment.
Thanks,
Tao
From: Andrew Lunn <andrew@lunn.ch> Date: 2021-08-14 15:27:09
Thanks for jumping in. We are using BCM5389 and the MDIO bus is used to
access BCM5389 MDC/MDIO interface in Pseudo-PHY mode.
I didn't know drivers/net/dsa, but let me check out the drivers and see
if it works in the Cloudripper environment.
The b53 driver knows about this switch, so please make use of that
driver. See Documentation/devicetree/bindings/net/dsa/brcm,b53.yaml
for how you describe it in DT.
Andrew
On Sat, Aug 14, 2021 at 05:26:55PM +0200, Andrew Lunn wrote:
quoted
Thanks for jumping in. We are using BCM5389 and the MDIO bus is used to
access BCM5389 MDC/MDIO interface in Pseudo-PHY mode.
I didn't know drivers/net/dsa, but let me check out the drivers and see
if it works in the Cloudripper environment.
The b53 driver knows about this switch, so please make use of that
driver. See Documentation/devicetree/bindings/net/dsa/brcm,b53.yaml
for how you describe it in DT.
Andrew
Thank you Andrew for looking into it. I will update dts and send out
followup patch after validation (perhaps after v5.15 merge window).
Thanks,
Tao
On Thu, Aug 05, 2021 at 03:28:15PM -0700, rentao.bupt@gmail.com wrote:
This common descirption is included by all Facebook AST2600 Network BMC
platforms to minimize duplicated device entries across Facebook Network
BMC device trees.
Hi Lukas,
On Wed, Dec 20, 2023 at 09:14:02AM +0100, Lukas Wunner wrote:
On Thu, Aug 05, 2021 at 03:28:15PM -0700, rentao.bupt@gmail.com wrote:
quoted
This common descirption is included by all Facebook AST2600 Network BMC
platforms to minimize duplicated device entries across Facebook Network
BMC device trees.
On Tue, Jan 02, 2024 at 04:11:28PM -0800, Tao Ren wrote:
On Wed, Dec 20, 2023 at 09:14:02AM +0100, Lukas Wunner wrote:
quoted
On Thu, Aug 05, 2021 at 03:28:15PM -0700, rentao.bupt@gmail.com wrote:
quoted
This common descirption is included by all Facebook AST2600 Network BMC
platforms to minimize duplicated device entries across Facebook Network
BMC device trees.
What's the chip used on this board? Going forward, the DT schema for TPMs
requires the exact chip name in addition to the generic "tcg,tpm_tis-spi".
Sorry about the late response. It's "infineon,slb9670", and I will
submit a patch for fix it soon.
Thank you Tao and Patrick for the replies. I've prepared two commits
to fix all violations of the TPM schema on this branch:
https://github.com/l1k/linux/commits/tpm-dt3
The commits are named:
e95bdbc arm64: dts: Fix TPM schema violations
63e5dfd ARM: dts: Fix TPM schema violations
I've now amended the ARM commit to use "infineon,slb9670" for:
arch/arm/boot/dts/aspeed/aspeed-bmc-facebook-bletchley.dts
arch/arm/boot/dts/aspeed/aspeed-bmc-facebook-wedge400.dts
arch/arm/boot/dts/aspeed/ast2600-facebook-netbmc-common.dtsi
I intend to submit the two commits once the TPM schema is applied
either to Jarkko's or Rob's tree:
https://lore.kernel.org/all/20231220160422.GA282877-robh@kernel.org/
Thanks,
Lukas
From: Patrick Williams <hidden> Date: 2024-01-03 15:48:53
On Wed, Jan 03, 2024 at 01:45:02PM +0100, Lukas Wunner wrote:
On Tue, Jan 02, 2024 at 04:11:28PM -0800, Tao Ren wrote:
quoted
On Wed, Dec 20, 2023 at 09:14:02AM +0100, Lukas Wunner wrote:
quoted
On Thu, Aug 05, 2021 at 03:28:15PM -0700, rentao.bupt@gmail.com wrote:
quoted
This common descirption is included by all Facebook AST2600 Network BMC
platforms to minimize duplicated device entries across Facebook Network
BMC device trees.
What's the chip used on this board? Going forward, the DT schema for TPMs
requires the exact chip name in addition to the generic "tcg,tpm_tis-spi".
Sorry about the late response. It's "infineon,slb9670", and I will
submit a patch for fix it soon.
Thank you Tao and Patrick for the replies. I've prepared two commits
to fix all violations of the TPM schema on this branch:
https://github.com/l1k/linux/commits/tpm-dt3
The commits are named:
e95bdbc arm64: dts: Fix TPM schema violations
63e5dfd ARM: dts: Fix TPM schema violations
Feel free to add to "63e5dfd (ARM: dts: Fix TPM schema violations)":
Reviewed-by: Patrick Williams <redacted>
I've now amended the ARM commit to use "infineon,slb9670" for:
arch/arm/boot/dts/aspeed/aspeed-bmc-facebook-bletchley.dts
arch/arm/boot/dts/aspeed/aspeed-bmc-facebook-wedge400.dts
arch/arm/boot/dts/aspeed/ast2600-facebook-netbmc-common.dtsi
I intend to submit the two commits once the TPM schema is applied
either to Jarkko's or Rob's tree:
https://lore.kernel.org/all/20231220160422.GA282877-robh@kernel.org/
Thanks,
Lukas
Hi Lukas,
On Wed, Jan 03, 2024 at 01:45:02PM +0100, Lukas Wunner wrote:
On Tue, Jan 02, 2024 at 04:11:28PM -0800, Tao Ren wrote:
quoted
On Wed, Dec 20, 2023 at 09:14:02AM +0100, Lukas Wunner wrote:
quoted
On Thu, Aug 05, 2021 at 03:28:15PM -0700, rentao.bupt@gmail.com wrote:
quoted
This common descirption is included by all Facebook AST2600 Network BMC
platforms to minimize duplicated device entries across Facebook Network
BMC device trees.
What's the chip used on this board? Going forward, the DT schema for TPMs
requires the exact chip name in addition to the generic "tcg,tpm_tis-spi".
Sorry about the late response. It's "infineon,slb9670", and I will
submit a patch for fix it soon.
Thank you Tao and Patrick for the replies. I've prepared two commits
to fix all violations of the TPM schema on this branch:
https://github.com/l1k/linux/commits/tpm-dt3
The commits are named:
e95bdbc arm64: dts: Fix TPM schema violations
63e5dfd ARM: dts: Fix TPM schema violations
I've now amended the ARM commit to use "infineon,slb9670" for:
arch/arm/boot/dts/aspeed/aspeed-bmc-facebook-bletchley.dts
arch/arm/boot/dts/aspeed/aspeed-bmc-facebook-wedge400.dts
arch/arm/boot/dts/aspeed/ast2600-facebook-netbmc-common.dtsi
I intend to submit the two commits once the TPM schema is applied
either to Jarkko's or Rob's tree:
https://lore.kernel.org/all/20231220160422.GA282877-robh@kernel.org/
Thanks,
Lukas
I've confirmed it's "infineon,slb9670" for both wedge400 and
ast2600-facebook-netbmc-common.
Pleasr free to add to "63e5dfd (ARM: dts: Fix TPM schema violations)":
Reviewed-by: Tao Ren <redacted>
Cheers,
- Tao