Split out dumper allocation from list insertion

[mhorne]

Summary

Add a new function, dumper_create(), to allocate a dumper.
dumper_insert() will call this function and retains the existing
behaviour.

This is desirable for performing live dumps of the system. Here, there
is a need to allocate and configure a dumper structure that is invoked
outside of the typical debugger context. Therefore, it should be
excluded from the list of panic-time dumpers.

free_single_dumper() is made public and renamed to dumper_destroy().

[kib]

sys/kern/kern_shutdown.c

1265

This is very unusual to allocate the item and change the output argument, despite returning an error. The requirement that you must

• set the *dip value to NULL before call
• call dumper_destroy() regardless of the return value

are error-prone IMO.

[mhorne]

Make dumper_create() handle its own cleanup in failure cases.

[mhorne]

sys/kern/kern_shutdown.c

1265

Yeah, I clearly did not look this over carefully enough before posting. Even D34069 did not expect to call dumper_destroy() in the failure case.

The new version should handle things more sensibly, although I did not understand your comment about the requirement that *dip be set to NULL.

[kib]

sys/kern/kern_shutdown.c

1265

To reliably detect the need for cleanup after failure in the old code, you have to set *dip to NULL before calling dumper_create(). Otherwise you cannot distinguish case of error before allocation vs. error after.

[markj]

sys/kern/kern_shutdown.c

1243

I wouldn't bother checking for dip == NULL if it only happens due to programmer error.