> On Mon, Jan 22, 2001 at 12:16:38PM -0800, Mike Smith wrote:
> > In the meantime, perhaps we could
> > ask that one of the SMPng rules of engagement mandate that no mutex
> > structures or structure members should ever be exported as part of a
> > userspace interface?
>
> This sounds fine in
On Mon, Jan 22, 2001 at 12:16:38PM -0800, Mike Smith wrote:
> In the meantime, perhaps we could
> ask that one of the SMPng rules of engagement mandate that no mutex
> structures or structure members should ever be exported as part of a
> userspace interface?
This sounds fine in principle, but
I got quite upset about this last time, and I guess it's time to do it
again.
Folks, *please* stop exporting "pure" kernel structures to userland.
Make a sanitisied, versioned structure and just copy your damn args back
and forth. 'struct ucred' should probably never have been exported to
< said:
> I looked at fixing this once, but got scared off because of binary
> compatibility issues. Would 'fixing' mount to use cmsgcred be
> acceptable?
No, it should use a structure appropriately named and designed for its
own purpose. (By preference, it should be binary-compatible with
4.x
In message <[EMAIL PROTECTED]> Alfred Perlstein writes:
: I looked at fixing this once, but got scared off because of binary
: compatibility issues. Would 'fixing' mount to use cmsgcred be
: acceptable?
I think so. Right now we have lots of killer, panic inducing binary
incompatibilities. One
* Garrett Wollman <[EMAIL PROTECTED]> [010122 10:27] wrote:
> < said:
>
> > Somehow there were few problems when `struct mtx' was added to
> > `struct ucred'. The critical args were probably usually 0.
>
> It's a bug that mount(2) uses a bare `struct ucred' and not a
> well-defined, user-export