net/ncsi: allow to customize BMC MAC Address offset
Subsystems (3)
PATCH 1/1 net/ncsi: allow to customize BMC MAC Address offset
Currently BMC's MAC address is calculated by adding 1 to NCSI NIC's base MAC address when CONFIG_NCSI_OEM_CMD_GET_MAC option is enabled. The logic doesn't work for platforms with different BMC MAC offset: for example, Facebook Yamp BMC's MAC address is calculated by adding 2 to NIC's base MAC address ("BaseMAC + 1" is reserved for Host use).
This patch adds NET_NCSI_MC_MAC_OFFSET config option to customize offset between NIC's Base MAC address and BMC's MAC address. Its default value is set to 1 to avoid breaking existing users.
--- net/ncsi/Kconfig | 8 ++++++++ net/ncsi/ncsi-rsp.c | 15 +++++++++++++-- 2 files changed, 21 insertions(+), 2 deletions(-)
net/ncsi/Kconfig +8 -0
@@ -17,3 +17,11 @@ config NCSI_OEM_CMD_GET_MAC ---help--- This allows to get MAC address from NCSI firmware and set them back to controller. +config NET_NCSI_MC_MAC_OFFSET + int + prompt "Offset of Management Controller's MAC Address" + depends on NCSI_OEM_CMD_GET_MAC + default 1 + help + This defines the offset between Network Controller's (base) MAC + address and Management Controller's MAC address.
net/ncsi/ncsi-rsp.c +13 -2
@@ -656,6 +656,11 @@ static int ncsi_rsp_handler_oem_bcm_gma(struct ncsi_request *nr)struct ncsi_rsp_oem_pkt *rsp;struct sockaddr saddr;int ret =0 ; +#ifdef CONFIG_NET_NCSI_MC_MAC_OFFSET +int mac_offset =CONFIG_NET_NCSI_MC_MAC_OFFSET ; +#else +int mac_offset =1 ; +#endif /* Get the response header */ rsp = (struct ncsi_rsp_oem_pkt *)skb_network_header (nr->rsp);
@@ -663,8 +668,14 @@ static int ncsi_rsp_handler_oem_bcm_gma(struct ncsi_request *nr) saddr.sa_family = ndev->type; ndev->priv_flags |=IFF_LIVE_ADDR_CHANGE ;memcpy (saddr.sa_data, &rsp->data[BCM_MAC_ADDR_OFFSET ],ETH_ALEN ); -/* Increase mac address by 1 for BMC's address */ -eth_addr_inc ((u8 *)saddr.sa_data); + +/* Management Controller's MAC address is calculated by adding +* the offset to Network Controller's (base) MAC address. +* Note: negative offset is "ignored", and BMC will use the Base +* MAC address in this case. +*/ +while (mac_offset-- >0 ) +eth_addr_inc ((u8 *)saddr.sa_data);if (!is_valid_ether_addr ((const u8 *)saddr.sa_data))return -ENXIO ;
Signature · 1 line
2.17.1
Quoted 11 lines · Tao Ren
Currently BMC's MAC address is calculated by adding 1 to NCSI NIC's base MAC address when
CONFIG_NCSI_OEM_CMD_GET_MACoption is enabled. The logic doesn't work for platforms with different BMC MAC offset: for example, Facebook Yamp BMC's MAC address is calculated by adding 2 to NIC's base MAC address ("BaseMAC + 1" is reserved for Host use).This patch adds
NET_NCSI_MC_MAC_OFFSETconfig option to customize offset between NIC's Base MAC address and BMC's MAC address. Its default value is set to 1 to avoid breaking existing users.Signed-off-by Tao Ren <[email protected]>
Maybe someone more knowledgeable like Andrew has an opinion here, but to me it seems a bit strange to encode what seems to be platfrom information in the kernel config :(
Quoted 49 lines
net/ncsi/Kconfig +8
-0@@ -17,3 +17,11 @@ config NCSI_OEM_CMD_GET_MAC ---help--- This allows to get MAC address from NCSI firmware and set them back to controller. +config NET_NCSI_MC_MAC_OFFSET + int + prompt "Offset of Management Controller's MAC Address" + depends on NCSI_OEM_CMD_GET_MAC + default 1 + help + This defines the offset between Network Controller's (base) MAC + address and Management Controller's MAC address.net/ncsi/ncsi-rsp.c +13
-2@@ -656,6 +656,11 @@ static int ncsi_rsp_handler_oem_bcm_gma(struct ncsi_request *nr)struct ncsi_rsp_oem_pkt *rsp;struct sockaddr saddr;int ret =0 ; +#ifdef CONFIG_NET_NCSI_MC_MAC_OFFSET +int mac_offset =CONFIG_NET_NCSI_MC_MAC_OFFSET ; +#else +int mac_offset =1 ; +#endif /* Get the response header */ rsp = (struct ncsi_rsp_oem_pkt *)skb_network_header (nr->rsp);@@ -663,8 +668,14 @@ static int ncsi_rsp_handler_oem_bcm_gma(struct ncsi_request *nr) saddr.sa_family = ndev->type; ndev->priv_flags |=IFF_LIVE_ADDR_CHANGE ;memcpy (saddr.sa_data, &rsp->data[BCM_MAC_ADDR_OFFSET ],ETH_ALEN ); -/* Increase mac address by 1 for BMC's address */ -eth_addr_inc ((u8 *)saddr.sa_data); + +/* Management Controller's MAC address is calculated by adding +* the offset to Network Controller's (base) MAC address. +* Note: negative offset is "ignored", and BMC will use the Base +* MAC address in this case. +*/ +while (mac_offset-- >0 ) +eth_addr_inc ((u8 *)saddr.sa_data);if (!is_valid_ether_addr ((const u8 *)saddr.sa_data))return -ENXIO ;
Quoted 11 lines · Jakub Kicinski
Quoted 10 lines · Tao Ren
Currently BMC's MAC address is calculated by adding 1 to NCSI NIC's base MAC address when
CONFIG_NCSI_OEM_CMD_GET_MACoption is enabled. The logic doesn't work for platforms with different BMC MAC offset: for example, Facebook Yamp BMC's MAC address is calculated by adding 2 to NIC's base MAC address ("BaseMAC + 1" is reserved for Host use).This patch adds
NET_NCSI_MC_MAC_OFFSETconfig option to customize offset between NIC's Base MAC address and BMC's MAC address. Its default value is set to 1 to avoid breaking existing users.
Quoted 5 lines
Quoted 1 line
Signed-off-by Tao Ren <[email protected]>Maybe someone more knowledgeable like Andrew has an opinion here, but to me it seems a bit strange to encode what seems to be platfrom information in the kernel config :(
Yes, this is not a good idea. It makes it impossible to have a 'BMC distro' kernel which you install on a number of different BMCs.
A device tree property would be better. Ideally it would be the standard local-mac-address, or mac-address.
Andrew
Quoted 18 lines · Andrew Lunn
Quoted 11 lines · Jakub Kicinski
Quoted 10 lines · Tao Ren
Currently BMC's MAC address is calculated by adding 1 to NCSI NIC's base MAC address when
CONFIG_NCSI_OEM_CMD_GET_MACoption is enabled. The logic doesn't work for platforms with different BMC MAC offset: for example, Facebook Yamp BMC's MAC address is calculated by adding 2 to NIC's base MAC address ("BaseMAC + 1" is reserved for Host use).This patch adds
NET_NCSI_MC_MAC_OFFSETconfig option to customize offset between NIC's Base MAC address and BMC's MAC address. Its default value is set to 1 to avoid breaking existing users.Quoted 5 lines
Quoted 1 line
Signed-off-by Tao Ren <[email protected]>Maybe someone more knowledgeable like Andrew has an opinion here, but to me it seems a bit strange to encode what seems to be platfrom information in the kernel config :(
Quoted 5 lines
Yes, this is not a good idea. It makes it impossible to have a 'BMC distro' kernel which you install on a number of different BMCs.
A device tree property would be better. Ideally it would be the standard local-mac-address, or mac-address.
Thank you Andrew and Jakub for the feedback. I picked up kconfig approach mainly because it's an OEM-only extention which is used only when NCSI_OEM_CMD_GET_MAC is enabled.
Let me prepare patch v2 using device tree. I'm not sure if standard "mac-address" fits this situation because all we need is an offset (integer) and BMC MAC is calculated by adding the offset to NIC's MAC address. Anyways, let me work out v2 patch we can discuss more then.
Thanks,
Tao
Quoted 5 lines
Let me prepare patch v2 using device tree. I'm not sure if standard "mac-address" fits this situation because all we need is an offset (integer) and BMC MAC is calculated by adding the offset to NIC's MAC address. Anyways, let me work out v2 patch we can discuss more then.
Hi Tao
I don't know BMC terminology. By NICs MAC address, you are referring to the hosts MAC address? The MAC address the big CPU is using for its interface? Where does this NIC get its MAC address from? If the BMCs bootloader has access to it, it can set the mac-address property in the device tree.
Andrew
Hi Andrew,
Quoted 6 lines · Andrew Lunn
Quoted 5 lines
Let me prepare patch v2 using device tree. I'm not sure if standard "mac-address" fits this situation because all we need is an offset (integer) and BMC MAC is calculated by adding the offset to NIC's MAC address. Anyways, let me work out v2 patch we can discuss more then.
Quoted 7 lines
Hi Tao
I don't know BMC terminology. By NICs MAC address, you are referring to the hosts MAC address? The MAC address the big CPU is using for its interface? Where does this NIC get its MAC address from? If the BMCs bootloader has access to it, it can set the mac-address property in the device tree.
Sorry for the confusion and let me clarify more:
The NIC here refers to the Network controller which provide network connectivity for both BMC (via NC-SI) and Host (for example, via PCIe).
On Facebook Yamp BMC, BMC sends NCSI_OEM_GET_MAC command (as an ethernet packet) to the Network Controller while bringing up eth0, and the (Broadcom) Network Controller replies with the Base MAC Address reserved for the platform. As for Yamp, Base-MAC and Base-MAC+1 are used by Host (big CPU) and Base-MAC+2 are assigned to BMC. In my opinion, Base MAC and MAC address assignments are controlled by Network Controller, which is transparent to both BMC and Host.
I'm not sure if I understand your suggestion correctly: do you mean we should move the logic (GET_MAC from Network Controller, adding offset and configuring BMC MAC) from kernel to boot loader?
Sam posted several ncsi patches for u-boot recently. Sam, do we support the work (implemented in this patch) in uboot? Or are we planning to do so?
Thanks,
Tao
Quoted 25 lines · Tao Ren
Hi Andrew,
Quoted 6 lines · Andrew Lunn
Quoted 5 lines
Let me prepare patch v2 using device tree. I'm not sure if standard "mac-address" fits this situation because all we need is an offset (integer) and BMC MAC is calculated by adding the offset to NIC's MAC address. Anyways, let me work out v2 patch we can discuss more then.
Quoted 7 lines
Hi Tao
I don't know BMC terminology. By NICs MAC address, you are referring to the hosts MAC address? The MAC address the big CPU is using for its interface? Where does this NIC get its MAC address from? If the BMCs bootloader has access to it, it can set the mac-address property in the device tree.
Sorry for the confusion and let me clarify more:
The NIC here refers to the Network controller which provide network connectivity for both BMC (via NC-SI) and Host (for example, via PCIe).
Quoted 8 lines
On Facebook Yamp BMC, BMC sends
NCSI_OEM_GET_MACcommand (as an ethernet packet) to the Network Controller while bringing up eth0, and the (Broadcom) Network Controller replies with the Base MAC Address reserved for the platform. As for Yamp, Base-MAC and Base-MAC+1 are used by Host (big CPU) and Base-MAC+2 are assigned to BMC. In my opinion, Base MAC and MAC address assignments are controlled by Network Controller, which is transparent to both BMC and Host.
Hi Tao
I've not done any work in the BMC field, so thanks for explaining this.
In a typical embedded system, each network interface is assigned a MAC address by the vendor. But here, things are different. The BMC SoC network interface has not been assigned a MAC address, it needs to ask the network controller for its MAC address, and then do some magical transformation on the answer to derive a MAC address for itself. Correct?
It seems like a better design would of been, the BMC sends a NCSI_OEM_GET_BMC_MAC and the answer it gets back is the MAC address the BMC should use. No magic involved. But i guess it is too late to do that now.
Quoted 3 lines
I'm not sure if I understand your suggestion correctly: do you mean we should move the logic (
GET_MACfrom Network Controller, adding offset and configuring BMC MAC) from kernel to boot loader?
In general, the kernel is generic. It probably boots on any ARM system which is has the needed modules for. The bootloader is often much more specific. It might not be fully platform specific, but it will be at least specific to the general family of BMC SoCs. If you consider the combination of the BMC bootloader and the device tree blob, you have something specific to the platform. This magical transformation of adding 2 seems to be very platform specific. So having this magic in the bootloader+DT seems like the best place to put it.
However, how you pass the resulting MAC address to the kernel should be as generic as possible. The DT "mac-address" property is very generic, many MAC drivers understand it. Using it also allows for vendors which actually assign a MAC address to the BMC to pass it to the BMC, avoiding all this NCSI_OEM_GET_MAC handshake. Having an API which just passing '2' is not generic at all.
Andrew
Quoted 40 lines · Andrew Lunn
Quoted 25 lines · Tao Ren
Hi Andrew,
Quoted 6 lines · Andrew Lunn
Quoted 5 lines
Let me prepare patch v2 using device tree. I'm not sure if standard "mac-address" fits this situation because all we need is an offset (integer) and BMC MAC is calculated by adding the offset to NIC's MAC address. Anyways, let me work out v2 patch we can discuss more then.
Quoted 7 lines
Hi Tao
I don't know BMC terminology. By NICs MAC address, you are referring to the hosts MAC address? The MAC address the big CPU is using for its interface? Where does this NIC get its MAC address from? If the BMCs bootloader has access to it, it can set the mac-address property in the device tree.
Sorry for the confusion and let me clarify more:
The NIC here refers to the Network controller which provide network connectivity for both BMC (via NC-SI) and Host (for example, via PCIe).
Quoted 8 lines
On Facebook Yamp BMC, BMC sends
NCSI_OEM_GET_MACcommand (as an ethernet packet) to the Network Controller while bringing up eth0, and the (Broadcom) Network Controller replies with the Base MAC Address reserved for the platform. As for Yamp, Base-MAC and Base-MAC+1 are used by Host (big CPU) and Base-MAC+2 are assigned to BMC. In my opinion, Base MAC and MAC address assignments are controlled by Network Controller, which is transparent to both BMC and Host.Hi Tao
I've not done any work in the BMC field, so thanks for explaining this.
Quoted 6 lines
In a typical embedded system, each network interface is assigned a MAC address by the vendor. But here, things are different. The BMC SoC network interface has not been assigned a MAC address, it needs to ask the network controller for its MAC address, and then do some magical transformation on the answer to derive a MAC address for itself. Correct?
Yes. It's correct.
Quoted 4 lines
It seems like a better design would of been, the BMC sends a
NCSI_OEM_GET_BMC_MACand the answer it gets back is the MAC address the BMC should use. No magic involved. But i guess it is too late to do that now.
Some NCSI Network Controllers support such OEM command (Get Provisioned BMC MAC Address), but unfortunately it's not supported on Yamp.
Quoted 4 lines
Quoted 3 lines
I'm not sure if I understand your suggestion correctly: do you mean we should move the logic (
GET_MACfrom Network Controller, adding offset and configuring BMC MAC) from kernel to boot loader?
Quoted 8 lines
In general, the kernel is generic. It probably boots on any ARM system which is has the needed modules for. The bootloader is often much more specific. It might not be fully platform specific, but it will be at least specific to the general family of BMC SoCs. If you consider the combination of the BMC bootloader and the device tree blob, you have something specific to the platform. This magical transformation of adding 2 seems to be very platform specific. So having this magic in the bootloader+DT seems like the best place to put it.
I understand your concern now. Thank you for the explanation.
Quoted 6 lines
However, how you pass the resulting MAC address to the kernel should be as generic as possible. The DT "mac-address" property is very generic, many MAC drivers understand it. Using it also allows for vendors which actually assign a MAC address to the BMC to pass it to the BMC, avoiding all this
NCSI_OEM_GET_MAChandshake. Having an API which just passing '2' is not generic at all.
After giving it more thought, I'm thinking about adding ncsi dt node with following structure (mac/ncsi similar to mac/mdio/phy):
&mac0 {
/* MAC properties... */
use-ncsi;
ncsi {
/* ncsi level properties if any */
package@0 {
/* package level properties if any */
channel@0 {
/* channel level properties if any */
bmc-mac-offset = <2>;
};
channel@1 {
/* channel #1 properties */
};
};
/* package #1 properties start here.. */
};
};The reasons behind this are:
1) mac driver doesn't need to parse "mac-offset" stuff: these ncsi-network-controller specific settings should be parsed in ncsi stack.
2) get_bmc_mac_address command is a channel specific command, and technically people can configure different offset/formula for different channels.
Any concerns or suggestions?
Thanks,
Tao
Quoted 7 lines
After giving it more thought, I'm thinking about adding ncsi dt node with following structure (mac/ncsi similar to mac/mdio/phy):
&mac0 { /* MAC properties... */ use-ncsi;
This property seems to be specific to Faraday FTGMAC100. Are you going to make it more generic?
Quoted 4 lines
ncsi {
/* ncsi level properties if any */
package@0 {You should get Rob Herring involved. This is not really describing hardware, so it might get rejected by the device tree maintainer.
Quoted 7 lines
1) mac driver doesn't need to parse "mac-offset" stuff: these ncsi-network-controller specific settings should be parsed in ncsi stack.
2)
get_bmc_mac_addresscommand is a channel specific command, and technically people can configure different offset/formula for different channels.
Does that mean the NCSA code puts the interface into promiscuous mode? Or at least adds these unicast MAC addresses to the MAC receive filter? Humm, ftgmac100 only seems to support multicast address filtering, not unicast filters, so it must be using promisc mode, if you expect to receive frames using this MAC address.
Andrew
Quoted 10 lines · Andrew Lunn
Quoted 7 lines
After giving it more thought, I'm thinking about adding ncsi dt node with following structure (mac/ncsi similar to mac/mdio/phy):
&mac0 { /* MAC properties... */ use-ncsi;This property seems to be specific to Faraday FTGMAC100. Are you going to make it more generic?
I'm also using ftgmac100 on my platform, and I don't have plan to change this property.
Quoted 7 lines
Quoted 4 lines
ncsi { /* ncsi level properties if any */ package@0 {You should get Rob Herring involved. This is not really describing hardware, so it might get rejected by the device tree maintainer.
Got it. Thank you for the sharing, and let me think it over :-)
Quoted 8 lines
Quoted 7 lines
1) mac driver doesn't need to parse "mac-offset" stuff: these ncsi-network-controller specific settings should be parsed in ncsi stack.
2)
get_bmc_mac_addresscommand is a channel specific command, and technically people can configure different offset/formula for different channels.
Quoted 5 lines
Does that mean the NCSA code puts the interface into promiscuous mode? Or at least adds these unicast MAC addresses to the MAC receive filter? Humm, ftgmac100 only seems to support multicast address filtering, not unicast filters, so it must be using promisc mode, if you expect to receive frames using this MAC address.
Uhh, I actually didn't think too much about this: basically it's how to configure frame filtering when there are multiple packages/channels active: single BMC MAC or multiple BMC MAC is also allowed? I don't have the answer yet, but will talk to NCSI expert and figure it out.
Thanks,
Tao
A message that is not in the archive (sent off-list or not yet received): 10079A1AC4244A41BC7939A794B72C238FCE0E03@fmsmsx104.amr.corp.intel.com
A message that is not in the archive (sent off-list or not yet received): [email protected]
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send NCSI_OEM_GET_MAC perhaps with netlink (TBD) to get the system address without it affecting the BMC address.
I was about to send patches to add support for the Intel adapters when I saw this thread.
Thanks,
Terry
Quoted 39 lines
Quoted 9 lines
Quoted 7 lines
After giving it more thought, I'm thinking about adding ncsi dt node
with following structure (mac/ncsi similar to mac/mdio/phy):&mac0 { /* MAC properties... */ use-ncsi;This property seems to be specific to Faraday FTGMAC100. Are you going to make it more generic?
I'm also using ftgmac100 on my platform, and I don't have plan to change this property.
Quoted 6 lines
Quoted 4 lines
ncsi { /* ncsi level properties if any */ package@0 {You should get Rob Herring involved. This is not really describing hardware, so it might get rejected by the device tree maintainer.
Got it. Thank you for the sharing, and let me think it over :-)
Quoted 11 lines
Quoted 6 lines
1) mac driver doesn't need to parse "mac-offset" stuff: these ncsi-network-controller specific settings should be parsed in ncsi stack.
2)get_bmc_mac_addresscommand is a channel specific command, and technically people can configure different offset/formula for different channels.Does that mean the NCSA code puts the interface into promiscuous mode? Or at least adds these unicast MAC addresses to the MAC receive filter? Humm, ftgmac100 only seems to support multicast address filtering, not unicast filters, so it must be using promisc mode, if you expect to receive frames using this MAC address.
Uhh, I actually didn't think too much about this: basically it's how to configure frame filtering when there are multiple packages/channels active: single BMC MAC or multiple BMC MAC is also allowed? I don't have the answer yet, but will talk to NCSI expert and figure it out.
Thanks,
Tao
Quoted 7 lines · Terry Duncan
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send
NCSI_OEM_GET_MACperhaps with netlink (TBD) to get the system address without it affecting the BMC address.I was about to send patches to add support for the Intel adapters when I saw this thread.
Thanks,
Terry
Hi Terry,
Sounds like you are planning to configure BMC MAC address from user space via netlink? Ben Wei <[email protected]> started a thread "Out-of-band NIC management" in openbmc community for NCSI management using netlink, and you may follow up with him for details.
I haven't decided what to do in my v2 patch: maybe using device tree, maybe moving the logic to uboot, and I'm also evaluating the netlink option. But it shouldn't impact your patch, because you can disable NCSI_OEM_GET_MAC option from your config file.
Cheers,
Tao
Quoted 5 lines · Terry Duncan
Quoted 4 lines
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send
NCSI_OEM_GET_MACperhaps with netlink (TBD) to get the system address without it affecting the BMC address.I was about to send patches to add support for the Intel adapters when I saw this thread.
Quoted 7 lines
Quoted 3 lines
Thanks,
Terry
Hi Terry,
Sounds like you are planning to configure BMC MAC address from user space via netlink? Ben Wei <[email protected]> started a thread "Out-of-band NIC management" in openbmc community for NCSI management using netlink, and you may follow up with him for details.
Hi Terry,
Do you plan to monitor and configure NIC through use space programs via netlink? I'm curious if you have additional use cases. I had planned to add some daemon in user space to monitor NIC through NC-SI, primarily to log AENs, check link status and retrieve NIC counters. We can collaborate on these if they align with your needs as well.
Also about Intel NIC, do you not plan to derive system address from BMC MAC? Is the BMC MAC independent of system address?
Thanks,
-Ben
Quoted 12 lines · Ben Wei
Quoted 7 lines · Terry Duncan
Quoted 5 lines
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send
NCSI_OEM_GET_MACperhaps with netlink (TBD) to get the system address without it affecting the BMC address.I was about to send patches to add support for the Intel adapters when I saw this thread. Thanks,
TerryHi Terry,
Do you plan to monitor and configure NIC through use space programs via netlink? I'm curious if you have additional use cases. I had planned to add some daemon in user space to monitor NIC through NC-SI, primarily to log AENs, check link status and retrieve NIC counters. We can collaborate on these if they align with your needs as well.
Quoted 5 lines
Also about Intel NIC, do you not plan to derive system address from BMC MAC? Is the BMC MAC independent of system address?
Thanks,
-Ben
Hi Ben,
Intel has in the past programmed BMC MACs with an offset from the system MAC address of the first shared interface. We have found this approach causes problems and adds complexity when system interfaces / OCP cards are added or removed or disabled in BIOS. Our approach in OpenBMC is to keep this simple - provide means for the BMC MAC addresses to be set in manufacturing and persist them.
We do also have some of the same use cases you mention. Thanks,
Terry
Quoted 7 lines · Tao Ren
Quoted 6 lines · Terry Duncan
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send
NCSI_OEM_GET_MACperhaps with netlink (TBD) to get the system address without it affecting the BMC address.I was about to send patches to add support for the Intel adapters when I saw this thread.
Thanks,
Quoted 7 lines
Quoted 1 line
Terry
Hi Terry,
Sounds like you are planning to configure BMC MAC address from user space via netlink? Ben Wei <[email protected]> started a thread "Out-of-band NIC management" in openbmc community for NCSI management using netlink, and you may follow up with him for details.
I haven't decided what to do in my v2 patch: maybe using device tree, maybe moving the logic to uboot, and I'm also evaluating the netlink option. But it shouldn't impact your patch, because you can disable
NCSI_OEM_GET_MACoption from your config file.
Thanks Tao. I see now that disabling the NCSI_OEM_GET_MAC option will do what I want.
Best,
Terry
Quoted 13 lines · Terry Duncan
Quoted 11 lines · Tao Ren
Quoted 7 lines · Terry Duncan
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send
NCSI_OEM_GET_MACperhaps with netlink (TBD) to get the system address without it affecting the BMC address.I was about to send patches to add support for the Intel adapters when I saw this thread.
Thanks,
Terry
Hi Terry,
Quoted 8 lines
Quoted 3 lines
Sounds like you are planning to configure BMC MAC address from user space via netlink? Ben Wei <[email protected]> started a thread "Out-of-band NIC management" in openbmc community for NCSI management using netlink, and you may follow up with him for details.
I haven't decided what to do in my v2 patch: maybe using device tree, maybe moving the logic to uboot, and I'm also evaluating the netlink option. But it shouldn't impact your patch, because you can disable
NCSI_OEM_GET_MACoption from your config file.Thanks Tao. I see now that disabling the
NCSI_OEM_GET_MACoption will do what I want.Best,
Terry
Hi Tao,
After a second look, it appears that the OEM handlers for Broadcom and Melanox in ncsi-rsp.c will set the MAC regardless of the origin of the request. Even with NCSI_OEM_GET_MAC disabled, sending an OEM command with netlink would result in setting the BMC MAC.
Thanks,
Terry
Quoted 18 lines · Terry Duncan
Quoted 17 lines · Terry Duncan
Quoted 14 lines · Tao Ren
Quoted 7 lines · Terry Duncan
Tao, in your new patch will it be possible to disable the setting of the BMC MAC? I would like to be able to send
NCSI_OEM_GET_MACperhaps with netlink (TBD) to get the system address without it affecting the BMC address.I was about to send patches to add support for the Intel adapters when I saw this thread.
Thanks,
Terry
Hi Terry,
Sounds like you are planning to configure BMC MAC address from user space via netlink? Ben Wei <[email protected]> started a thread "Out-of-band NIC management" in openbmc community for NCSI management using netlink, and you may follow up with him for details.
I haven't decided what to do in my v2 patch: maybe using device tree, maybe moving the logic to uboot, and I'm also evaluating the netlink option. But it shouldn't impact your patch, because you can disable
NCSI_OEM_GET_MACoption from your config file.
Quoted 7 lines
Quoted 4 lines
Thanks Tao. I see now that disabling the
NCSI_OEM_GET_MACoption will do what I want.Best,
TerryHi Tao,
After a second look, it appears that the OEM handlers for Broadcom and Melanox in
ncsi-rsp.cwill set the MAC regardless of the origin of the request. Even withNCSI_OEM_GET_MACdisabled, sending an OEM command with netlink would result in setting the BMC MAC.
I think it's because those oem handler functions are not controlled by NCSI_OEM_GET_MAC. I'm not sure if we did that intentionally, but IMHO it needs to be fixed.
Vijay might be able to share more background when the option was introduced?
- Tao