Re: a bug of the slab implementation in viengoos?

[email protected] (Ludovic Courtès) Mon, 31 May 2010 18:20:48 +0200

Newsgroups	gmane.os.hurd.l4
Message-ID	<[email protected]>

Hi,

Da Zheng <[email protected]> writes:

> I recently ported the slab implementation in viengoos to Hurd. It seems that the
> refcount field of hurd_slab isn't initialized when it is created in grow(). here
> is the patch to fix it.

Normally ‘allocate_buffer’ creates an anonymous mapping with ‘mmap’,
which is zeroed on GNU systems (info "(libc) Memory-mapped I/O"), so
that shouldn’t be needed.

Thanks,
Ludo’.