Mike Hearn [ARCHIVE] on Nostr: 📅 Original date posted:2011-08-11 🗒️ Summary of this message: Upgrade alerts ...
📅 Original date posted:2011-08-11
🗒️ Summary of this message: Upgrade alerts needed as more 0.3.23 nodes are listening than 0.3.24. Chronic problem with new code causing CRITICAL_SECTION deadlocks. ThreadSanitizer may help.
📝 Original message:> they don't look good (I show about 3x as many 0.3.23 nodes listening as
> 0.3.24
*cough* Upgrade alerts.
> Very much needed, didn't TD say something about a friend who wanted to
> do research in this area?
I don't know if he'll actually do anything. Best assume this
"position" is still open.
>> 2. We've got a chronic problem with new code causing CRITICAL_SECTION
>> deadlocks
I've seen locks that track ordering relative to other locks and assert
when they are locked out of order.
Though it's not inversion related, running ThreadSanitizer might help
find other thread safety issues:
http://code.google.com/p/data-race-test/wiki/ThreadSanitizer
🗒️ Summary of this message: Upgrade alerts needed as more 0.3.23 nodes are listening than 0.3.24. Chronic problem with new code causing CRITICAL_SECTION deadlocks. ThreadSanitizer may help.
📝 Original message:> they don't look good (I show about 3x as many 0.3.23 nodes listening as
> 0.3.24
*cough* Upgrade alerts.
> Very much needed, didn't TD say something about a friend who wanted to
> do research in this area?
I don't know if he'll actually do anything. Best assume this
"position" is still open.
>> 2. We've got a chronic problem with new code causing CRITICAL_SECTION
>> deadlocks
I've seen locks that track ordering relative to other locks and assert
when they are locked out of order.
Though it's not inversion related, running ThreadSanitizer might help
find other thread safety issues:
http://code.google.com/p/data-race-test/wiki/ThreadSanitizer