Databases
StackOverflow2013
Postsdbo
POForcing WaitForDebugEvent to execute a certain thread
 

Note that there are some explanatory texts on larger screens.

plurals
  1. POForcing WaitForDebugEvent to execute a certain thread
    primarykey
    data
    text
    <p>I am writing a windows debugger, so I have the usual ContinueDebugEvent/WaitForDebugEvent loop:</p> <pre><code>while (true) { ContinueDebugEvent(hTargetProcessId, hTargetThreadId, lastDebugResult); WaitForDebugEvent(&amp;evt, INFINITE); ... hTargetThreadId = evt.dwThreadId; } </code></pre> <p>This works fine, but I'm in the process of adding a feature which requires the ability to execute a specific thread mid-debugging. To do this, I suspend all threads I'm not interested in, like this (I've omitted a load of code for brevity - hope this is clear enough!):</p> <pre><code>while (true) { if (onlyACertainThread) { for (int n=0; n&lt;threadcount; n++) if(threads[n] != certainThread) SuspendThread(threads[n]); } ContinueDebugEvent(hTargetProcessId, hTargetThreadId, lastDebugResult); WaitForDebugEvent(&amp;evt, INFINITE); if (onlyACertainThread) { for (int n=0; n&lt;threadcount; n++) if(threads[n] != certainThread) ResumeThread(threads[n]); } ... (code to add to 'threads', update threadcount, decide if we want to restrict to a certain thread, etc) ... hTargetThreadId = evt.dwThreadId; } </code></pre> <p>However, I often see debug events from the wrong thread. For example:</p> <ul> <li>Thread A is running</li> <li>Decide we want to observe thread B; suspend thread A, call continue/waitfordebugevent(.. A.pid..)</li> <li>Get debug notifications (breakpoints, exceptions) from thread A.</li> </ul> <p>What I want is for the final step to be 'get notifications from B'. I've tried calling WaitForDebugEvent(..B.pid..) which doesn't work.</p> <p>Is this expected behaviour? </p> <p>My suspicion is that thread suspend/resume counters are only observed by the kernel when at thread is being scheduled; ie, at the end of a timeslice. This would cause thread A to continue running until its current timeslice is up. Can anyone confirm/deny this idea? If so.. how would I work around it? My best idea at the moment is to 'inject' an artificial call to Sleep(0) after I suspend a thread, but I really don't want to do that unless I have to!</p> <p>If it's any difference, the machine I run on is a single-threaded XP box.</p>
    singulars
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. PTQuestion
    1. This table or related slice is empty.
    1. USrandomdude
    plurals
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. This table or related slice is empty.
    1. VO
      singulars
      1. POForcing WaitForDebugEvent to execute a certain thread
      1. USLuke
      1. VTFavorite
    2. VO
      singulars
      1. POForcing WaitForDebugEvent to execute a certain thread
      1. This table or related slice is empty.
      1. VTUpMod
    1. COMSDN:'If the function succeeds, execution of the specified thread is suspended and the thread's suspend count is incremented'. I would be very, very surprised if SuspendThread() returns before the target thread has been removed from the set of running/ready threads. There is no mention of waiting until the next, unrelated interrupt that exits via the scheduler. On one core, you're guaranteed that the target thread is not running, if the target thread is running on another core, SuspendThread() would cause the core to be hardware-interrupted to suspend the thread.
      singulars
      1. POForcing WaitForDebugEvent to execute a certain thread
      1. USMartin James
 

Querying!

 
Guidance
A row detail

Detail views are divided into sections. All the information in the data section comes from columns in the selected row. The other sections display data from other, related rows.

Related data can be related in a to-one or a to-many fashion. Captions of data related in a to-many fashion link to a list view showing a filtered view of the table.

Try moving around until you find a non-empty to-many entry and click on the label to get to one. You can move back to the root by clicking on the database name in the header.

SQuiL has stopped working due to an internal error.

If you are curious you may find further information in the browser console, which is accessible through the devtools (F12).

Reload