Skip to contentSkip to conversation

Typo in urcu_qsbr.h ?

DISCUSSION Mathieu Desnoyers 1 participant 1 message Oct 8, 2013 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

----- Original Message -----

Quoted 39 lines

From: "Dmitri Shubin" <[email protected]>
To: [email protected]
Sent: Tuesday, October 8, 2013 4:09:37 AM
Subject: [lttng-dev] Typo in urcu_qsbr.h ?

Hi,

In urcu_qsbr.h I can see the following

85 /*
86 * QSBR read lock/unlock are guaranteed to be no-ops.
Therefore, we expose them
87 * in the LGPL header for any code to use. However, the debug
version is not
88 * nops and may contain sanity checks. To activate it,
applications must be
89 * recompiled with -DRCU_DEBUG (even non-LGPL/GPL
applications). This is the

     90   * best trade-off between license/performance/code triviality and
     91   * library debugging & tracing features we could come up with.
     92   */
     93
     94  #if (!defined(BUILD_QSBR_LIB) && defined(RCU_DEBUG))
     95
     96  static inline void rcu_read_lock(void)
     97  {
     98  }
     99
    100  static inline void rcu_read_unlock(void)
    101  {
    102  }
    103
    104  #else /* !RCU_DEBUG */
    105
    106  extern void rcu_read_lock(void);
    107  extern void rcu_read_unlock(void);
    108
    109  #endif /* !RCU_DEBUG */
Quoted 6 lines

Shouldn't the condition in line 94 actually be

     94  #if (!defined(BUILD_QSBR_LIB) && !defined(RCU_DEBUG))

So when RCU_DEBUG is _not_ defined we get static inlines for rcu_read_{,un}lock() ?

Indeed! Commit pushed:

commit d47b05cb05b91bb9679d127c51bc4beac12b899d

Author: Mathieu Desnoyers <[email protected]>
Date:   Tue Oct 8 09:23:10 2013 -0400

Fix: urcu-qsbr: reversed logic on RCU_DEBUG

* Dmitri Shubin <[email protected]> wrote:
> Shouldn't the condition in line 94 actually be
>
> 94 #if (!defined(BUILD_QSBR_LIB) && !defined(RCU_DEBUG))
>
> So when RCU_DEBUG is _not_ defined we get static inlines for > rcu_read_{,un}lock() ?

Indeed!

Thanks for reporting!

Mathieu

Quoted 9 lines

Thanks!

_______________________________________________
lttng-dev mailing list
[email protected]
http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

Signature · 3 lines
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com