Sorry I should have been clearer. What I actually meant was something
like this:
--- a/fs/xfs/xfsidbg.c 2008-05-01 11:34:55.000000000 +1000
+++ b/fs/xfs/xfsidbg.c 2008-05-01 11:35:18.000000000 +1000
@@ -5845,9 +5845,11 @@
kdb_printf("curr_cycle: %d prev_cycle: %d curr_block: %d prev_block:
%d\n",
log->l_curr_cycle, log->l_prev_cycle, log->l_curr_block,
log->l_prev_block);
- kdb_printf("iclog_bak: 0x%p iclog_size: 0x%x (%d) num iclogs: %d\n",
- log->l_iclog_bak, log->l_iclog_size, log->l_iclog_size,
- log->l_iclog_bufs);
+#ifdef DEBUG
+ kdb_printf("iclog_bak: 0x%p\n", log->l_iclog_bak);
+#endif
+ kdb_printf("iclog_size: 0x%x (%d) num iclogs: %d\n",
+ log->l_iclog_size, log->l_iclog_size, log->l_iclog_bufs);
kdb_printf("l_iclog_hsize %d l_iclog_heads %d\n",
log->l_iclog_hsize, log->l_iclog_heads);
kdb_printf("l_sectbb_log %u l_sectbb_mask %u\n",
This way there is no code duplicated between the debug and non-debug cases.
But I'm fine with your approach too.
Lachlan
David Chinner wrote:
On Wed, Apr 30, 2008 at 05:17:24PM +1000, Lachlan McIlroy wrote:
Dave, would the code look cleaner if iclog_bak was printed on a line by
itself?
That way we wouldn't need the #else and we wouldn't be duplicating code.
Yup, probably a good idea. I've changed it to match how this is handled
in other places in xfsidbg.c:
@@ -5846,8 +5846,12 @@
log->l_curr_cycle, log->l_prev_cycle, log->l_curr_block,
log->l_prev_block);
kdb_printf("iclog_bak: 0x%p iclog_size: 0x%x (%d) num iclogs: %d\n",
- log->l_iclog_bak, log->l_iclog_size, log->l_iclog_size,
- log->l_iclog_bufs);
+#ifdef DEBUG
+ log->l_iclog_bak,
+#else
+ NULL,
+#endif
+ log->l_iclog_size, log->l_iclog_size, log->l_iclog_bufs);
kdb_printf("l_iclog_hsize %d l_iclog_heads %d\n",
log->l_iclog_hsize, log->l_iclog_heads);
kdb_printf("l_sectbb_log %u l_sectbb_mask %u\n",
Cheers,
Dave.
|