mctp-client: add a generic mctp client

[molbear]

Adds a generic mctp client to use from the cli. Currently does not support extended addressing.

Testing

control commands

mctp-client type control eid 8 data 0x80 0x02
0x00 0x02 0x00 0x08 0x01 0x00

corner case testing

mctp-client type control eid 10 data 0x80 0x02
mctp-client: sendto(2): No route to host

mctp-client type control eid 8 data
mctp-client: no data to send

mctp-client type control eid data 0x80 0x02
usage:
        mctp-client [net <net>] eid <eid> type <type> data <data>
net defaults to MCTP_NET_ANY, data is space delimited hexadecimal. data must be the last parameter
possible types:
        control: control mctp messages per DSP0236
        pldm: platform level data model per DSP0241
        nc-si: NC-SI traffic over MCTP per DSP0261
        ethernet: Ethernet traffic over MCTP per DSP0261
        nvme: nvme over mctp per DSP0235
        spdm: spdm over mctp per DSP0275
        secured: secured messages using spdm over mctp per DSP0276
        pci: vdm using a pci based vendor id per DSP0236
        iana: vdm using an iana based vendor id per DSP0236

mctp-client type control eid data 0x80 0x02
usage:
        mctp-client [net <net>] eid <eid> type <type> data <data>
net defaults to MCTP_NET_ANY, data is space delimited hexadecimal. data must be the last parameter
possible types:
        control: control mctp messages per DSP0236
        pldm: platform level data model per DSP0241
        nc-si: NC-SI traffic over MCTP per DSP0261
        ethernet: Ethernet traffic over MCTP per DSP0261
        nvme: nvme over mctp per DSP0235
        spdm: spdm over mctp per DSP0275
        secured: secured messages using spdm over mctp per DSP0276
        pci: vdm using a pci based vendor id per DSP0236
        iana: vdm using an iana based vendor id per DSP0236

echo $?
1

mctp-client eid 8 data 0x80 0x02
usage:
        mctp-client [net <net>] eid <eid> type <type> data <data>
net defaults to MCTP_NET_ANY, data is space delimited hexadecimal. data must be the last parameter
possible types:
        control: control mctp messages per DSP0236
        pldm: platform level data model per DSP0241
        nc-si: NC-SI traffic over MCTP per DSP0261
        ethernet: Ethernet traffic over MCTP per DSP0261
        nvme: nvme over mctp per DSP0235
        spdm: spdm over mctp per DSP0275
        secured: secured messages using spdm over mctp per DSP0276
        pci: vdm using a pci based vendor id per DSP0236
        iana: vdm using an iana based vendor id per DSP0236

mctp-client type eid 8 data 0x80 0x02
mctp-client: invalid type: eid

[jk-ozlabs]

Looks great! Just a couple of minor things:

Firstly: could you add a Signed-off-by line, to indicate an agreement to the DCO?

Secondly: your usage text mentions:

data is space delimited hexadecimal.

But you're parsing with:

tmp = strtoul(val, &endp, 0);

- with the 0 there, you're accepting decimal by default, requiring the 0x for hex.

Not a huge issue, but I'd suggest we could make things more succinct with:

tmp = strtoul(val, &endp, 16);

and printing with

printf("%02x", recv_buffer[ctr]);

With the fixed two-char format (ie., from the 02 prefix), it should be obvious that the output is not decimal.

[jk-ozlabs]

But you're parsing with:

tmp = strtoul(val, &endp, 0);

no wait, I'm looking at the wrong parsing there. I see you have 16 for the actual data parsing. All good on that side!

Still, want to make the output the same format (ie, without the 0x prefix?)

[molbear]

Looks great! Just a couple of minor things:

Firstly: could you add a Signed-off-by line, to indicate an agreement to the DCO?

Secondly: your usage text mentions:

data is space delimited hexadecimal.

But you're parsing with:

tmp = strtoul(val, &endp, 0);

• with the 0 there, you're accepting decimal by default, requiring the 0x for hex.

Not a huge issue, but I'd suggest we could make things more succinct with:

tmp = strtoul(val, &endp, 16);

and printing with

printf("%02x", recv_buffer[ctr]);

With the fixed two-char format (ie., from the 02 prefix), it should be obvious that the output is not decimal.

ack to all, i'll follow up tomorrow u.s. time

[jk-ozlabs]

Looks great, thanks!