On Jan 31,  7:56am, enami@but-b.or.jp (enami tsugutomo) wrote:
-- Subject: Re: CVS commit: src/sys/kern

| > Modified Files:
| > 	src/sys/kern: exec_ecoff.c
| > 
| > Log Message:
| > make sure that bss size > 0 before we add a vmcmd for it.
| 
| So, why this test needs to be done in the caller?
| 
| When UVM is first introduced, vmcmd_map_zero had a test for the length
| like vmcmd_map_{page,read}vn, and it is removed later but the log
| message only mentions translation of return value (KERN_* -> E*).  And
| until recently (the last summer), mapping of size is allowed.  So, I
| guess the removal of the test was not intentional.
| 
| Nowadays, NEW_VMCMD{,2} is always expanded to the call of new_vmcmd(),
| probably we can move the test to there.

Yes, we should. What is done now is pointless. I did this so we could
pullup the change to 2.0 without much hassle.

christos