Re: Cygwin 3.4.3 and 3.5.0... hangs in make, top, procps, ls /proc/PID/...

[Takashi Yano <takashi.yano@nifty.ne.jp>]

On Mon, 16 Jan 2023 11:23:54 +0100
Corinna Vinschen wrote:
> On Jan 16 18:02, Takashi Yano via Cygwin wrote:
> > Hi Corinna,
> > 
> > On Mon, 9 Jan 2023 14:20:56 +0100
> > Corinna Vinschen wrote:
> > > On Jan  2 17:21, Takashi Yano via Cygwin wrote:
> > > > On Mon, 2 Jan 2023 14:38:03 +0900
> > > > Takashi Yano wrote:
> > > > > On Mon, 2 Jan 2023 11:32:01 +0900
> > > > > Takashi Yano wrote:
> > > > > > On Sat, 31 Dec 2022 13:01:29 -0700
> > > > > > Brian Inglis wrote:
> > > > > > > was also getting the messages below locally and still on GitHub scallywag:
> > > > > > > 
> > > > > > > 	cygcheck (6936) child_copy: cygheap read copy failed,
> > > > > > > 
> > > > > > > ../curl/scallywag/1_x86_64  build.log:2022-12-26T00:39:35.6163236Z       0 
> > > > > > > [main] cygcheck (6936) child_copy: cygheap read copy failed, 0x0..0x80003B5F0, 
> > > > > > > done 0, windows pid 6936, Win32 error 299
> > > > > > > [...]
> > > > > I found this issue occurs after the commit 30add3e6b3e3:
> > > > > "Cygwin: exec: don't access cygheap before it's initialized"
> > > > > .
> > > > > 
> > > > > Reverting this commit solves the issue.
> > > 
> > > That would break strace again, but...
> > > 
> > > > I'm not sure if this is the right thing, but the following
> > > > patch seems to fix the issue.
> > > 
> > > This looks pretty good to me and it keeps strace working per the
> > > description in 30add3e6b3e3.  Please push this to master and the
> > > 3.4 branch.
> > 
> > I noticed that the following error occurs even with this patch.
> > If you run:
> > while true; do cygcheck -cd cygwin > /dev/null; done
> > for one day or so, you will find the issue can be reproduced.
> > 
> > Both cygwin-3_4-branch and main (master) branch have this issue,
> > while cygwin 3.3.6 does not.
> > 
> > $ while true; do cygcheck -cd cygwin > /dev/null; done
> >       0 [main] cygcheck (15244) C:\cygwin64\bin\cygcheck.exe: *** fatal error -
> > MapViewOfFileEx 'shared.5'(0x138), Win32 error 487.  Terminating.
> >    3540 [main] cygcheck (15244) cygwin_exception::open_stackdumpfile: Dumping st
> > ack trace to cygcheck.exe.stackdump
> >       0 [main] cygcheck (10844) C:\cygwin64\bin\cygcheck.exe: *** fatal error -
> > MapViewOfFileEx 'cygpid.51742'(0x148), Win32 error 487.  Terminating.
> >       0 [main] cygcheck (1976) C:\cygwin64\bin\cygcheck.exe: *** fatal error - M
> [...]
> > Errors seem to be three types: (null), cygpid.xxx and shared.5.
> > I'm not sure what is happening and why at all, however, this
> > did not seem to happen before the commit 30add3e6b3e3.
> 
> I'll try to reproduce this issue.  But the weird thing is certainly
> this: The affected shared mem regions are apparently not the cygheap.
> Rather, they are the "shared" and "cygpid" shared mem regions, which
> should not at all collide with the cygheap.  I guess we need more
> debug output in the api_fatal call inside open_shared...

I am now trying the test case with reverting the commit 60675f1a7eb2
"Cygwin: decouple shared mem regions from Cygwin DLL", and the issue
does not happen for several hours so far. I guess this most likely
is the direct cause of the problem.

-- 
Takashi Yano <takashi.yano@nifty.ne.jp>