SysV IPC: provide in-kernel helpers to obtain ipcs(8)-like information

[kib]

Summary

PR: 278949

[rbranco_suse.com]

The native FreeBSD ipcs command also uses these sysctl's.

[kib]

In D45175#1030481, @rbranco_suse.com wrote:

The native FreeBSD ipcs command also uses these sysctl's.

So what is the problem with native?

[rbranco_suse.com]

In D45175#1030482, @kib wrote:

In D45175#1030481, @rbranco_suse.com wrote:

The native FreeBSD ipcs command also uses these sysctl's.

So what is the problem with native?

I couldn't statically compile the 32-bits version to try. But I don't know if it would hit the same error.

[imp]

sys/kern/sysv_msg.c

1442

Is linprocfs only called from linux binaries?

[markj]

sys/kern/sysv_msg.c

1442

Oh, right, good point. We'd still want cat(1) etc. to work, so I think this diff isn't the right solution.

[imp]

A 32bit freebsd cat reading /compat/linux/mumble

[rbranco_suse.com]

In D45175#1030482, @kib wrote:

In D45175#1030481, @rbranco_suse.com wrote:

The native FreeBSD ipcs command also uses these sysctl's.

So what is the problem with native?

Forget it. I just tried a 32-bits ipcs on 64-bits and it worked.

[markj]

sys/kern/sysv_msg.c

1509

I would just allocate the array with calloc() and not bother with all of this zeroing.

[kib]

sys/kern/sysv_msg.c

1509

These members are copied by the assignments like *pmsqk = msgids[i] above. Pre-zeroing the array would not help.

[kib]

Move SHMSEG symbols.

[rbranco_suse.com]

I can confirm the patch works as intended.