On Thu, Aug 23, 2007 at 09:38:34PM +0200, Christoph Hellwig wrote:
> struct bhv_vnode is on it's way out, so move the trace buffer to the
> XFS inode. Note that this makes the tracing macros rather misnamed,
> but this kind of fallout will be fixed up incrementally later on.
....
> @@ -197,11 +192,11 @@ vn_hold(
> #ifdef XFS_VNODE_TRACE
>
> #define KTRACE_ENTER(vp, vk, s, line, ra) \
> - ktrace_enter( (vp)->v_trace, \
> + ktrace_enter( (ip)->i_trace, \
> /* 0 */ (void *)(__psint_t)(vk), \
> /* 1 */ (void *)(s), \
> /* 2 */ (void *)(__psint_t) line, \
> -/* 3 */ (void *)(__psint_t)(vn_count(vp)), \
> +/* 3 */ NULL, \
given that vn_count() returns the linux inode i_count, shouldn't
we put that here as well rather than killing it? It is sometimes
useful to see what is happening with the reference count in the
traces....
> +#ifdef XFS_VNODE_TRACE
> + ip->i_trace = ktrace_alloc(VNODE_TRACE_SIZE, KM_SLEEP);
> +#endif
At some point we need to change that define as well. More
cosmetic stuff, though.
Cheers,
Dave.
--
Dave Chinner
Principal Engineer
SGI Australian Software Group
|