we should release new version?

GoogleCodeExporter commented 9 years ago

code in current trunk seem stable and dont have the problem about performance.

we should release current trunk as v0.7 ?

Original issue reported on code.google.com by kazuh...@gmail.com on 24 Feb 2007 at 5:37

GoogleCodeExporter commented 9 years ago

check the forum

there's a bug that causes a kernel panic when downloading a large file - panic 
at 160M

when this bug is fixed we should post a new version

Original comment by jalav...@gmail.com on 24 Feb 2007 at 7:20

GoogleCodeExporter commented 9 years ago

jalavoui, 

you register the bug in issues as New Issue.

Original comment by kazuh...@gmail.com on 25 Feb 2007 at 1:32

GoogleCodeExporter commented 9 years ago

this issue is fine for this

do you also get this bug?

this could be because of several things:

- threads that aren't release
- need iolocks in some functions
- io80211 family - can you try to use ioethernetcontroller instead? - this 
could help fixing this bug

i liked you last changes in code - but i'm annoyed because i've to rewrite my 
changes in svn again

Original comment by jalav...@gmail.com on 25 Feb 2007 at 1:58

GoogleCodeExporter commented 9 years ago

i find out that if you use a iolock in ipw_rx() you avoid the kernel panic when 
downloading at 170M - but mac os 
freezes. any idea?

Original comment by jalav...@gmail.com on 25 Feb 2007 at 2:00

GoogleCodeExporter commented 9 years ago

>this could be because of several things:

> - threads that aren't release

this thread is thread of queue_td and queue_te ?

> - need iolocks in some functions

locking is in rx and tx ?

we should write checking if race conditions cause or not in all part.

locking is not required in rx.
the reason 
  interrupt service routine is not  work parallely and  no-interrupt.
  this is written in apple doccument.

locling is required in tx. 
  outputPacket is called in multi thread. this' routine is thread safe.

I know if requirement or not about other part.
we must write checking code(race conditions happen or not?)  in another part

> - io80211 family - can you try to use ioethernetcontroller instead? - this 
could
help fixing this bug

I wish not to try  ioethernetcontroller base. 
because i dont think that biggest file bug is not caused by  io80211 .

I think that the bug is caused by failing mbuf merging or race condition.

if you think that it is caused by io80211, you should try to test it.

Original comment by kazuh...@gmail.com on 25 Feb 2007 at 4:41

GoogleCodeExporter commented 9 years ago

Original comment by jalav...@gmail.com on 9 Mar 2007 at 10:33

Changed state: Fixed

kemenaran / iwidarwin

we should release new version? #36