Runtime

Timestamp:

Jun 16, 2012 2:44:06 PM (13 years ago)

Author:

vboxsync

Message:

bugref..

Location:

trunk/src/VBox/Runtime

Files:

trunk/src/VBox/Runtime/generic/timerlr-generic.cpp

r40320	r41774
319	319	* If we're more than 60 intervals behind, just skip ahead. We
320	320	* don't want the timer thread running wild just because the
321		* clock changed in an unexpected way. As seen in ~~#3611~~ this
	321	* clock changed in an unexpected way. As seen in @bugref{3611} this
322	322	* does happen during suspend/resume, but it may also happen
323	323	* if we're using a non-monotonic clock as time source.

r40966	r41774
445	445	*
446	446	* The unpinning is done to prevent a deadlock, see s this could lead to a
447		* deadlock (see ~~#4259~~ for the full explanation)
	447	* deadlock (see @bugref{4259} for the full explanation)
448	448	*
449	449	* @param pMtx The mutex to enter.

r39083	r41774
91	91	return RTMpIsCpuPossible(idCpu);
92	92	#else
93		/** @todo proper ring-3 support on darwin, see ~~#3014~~. */
	93	/** @todo proper ring-3 support on darwin, see @bugref{3014}. */
94	94	natural_t nCpus;
95	95	processor_basic_info_t pinfo;

r37211	r41774
32	32	* glibc 2.6 fixed a serious bug in the mutex implementation. We wrote this
33	33	* linux specific event semaphores code in order to work around the bug. As it
34		* turns out, this code seems to have an unresolved issue (~~#2599~~), so we'll
	34	* turns out, this code seems to have an unresolved issue (@bugref{2599}), so we'll
35	35	* fall back on the pthread based implementation if glibc is known to contain
36	36	* the bug fix.

Note: See TracChangeset for help on using the changeset viewer.