Re: stampede hd0*

[Thomas Bushnell, BSG]

Roland McGrath <address@hidden> writes:

> > > I am concerned about the handling of S_ITRANS bits in other places.  I 
> > > have
> > > not checked thoroughly, but I suspect more inconsistencies.
> > > e.g. diskfs_create_node calls do not mask out the S_ITRANS bits and so 
> > > they
> > > can wind up in dn_stat.st_mode when diskfs_read_node was never called.
> > 
> > Yeah, there may be some problems like that.
> 
> You're a helpful bugger, aren't you?  Care to comment on which bits ought
> to be possibly set in a diskfs_create_node call?

Seems to me like diskfs_create_node should not set the S_ITRANS or
S_ISPARE bits, but of course, it should set the S_IFMT and permissions
bits.

> Also, the sticky bit is cleared in dir-lookup/dir-mkfile but not in 
> dir-mkdir.  Should every place that masks off S_IFMT|S_ISPARE be consistent?

The sticky bit means something quite different between dir-mkfile and
dir-mkdir; on ordinary files, it's privileged.  (Or once was, in BSD,
when it really meant something.)