Corinna Vinschen corinna-cygwin@cygwin.com
Thu Jul 2 12:13:00 GMT 2015

On Jul  1 22:10, Ken Brown wrote:
> On 7/1/2015 4:12 PM, Ken Brown wrote:
> >On 7/1/2015 9:57 AM, Corinna Vinschen wrote:
> >>On Jul  1 12:47, Corinna Vinschen wrote:
> >>>On Jun 30 16:13, Ken Brown wrote:
> >>>>On 6/30/2015 3:55 PM, Corinna Vinschen wrote:
> >>>>>On Jun 27 16:52, Corinna Vinschen wrote:
> >>>>>>On Jun 26 18:28, Ken Brown wrote:
> >>>>>>>On the other hand, emacs doesn't really make a full recovery.  For example,
> >>>>>>>if I try to call a subprocess (e.g., 'C-x d' to list a directory), I get a
> >>>>>>>fork error:
> >>>>>>>
> >>>>>>>Debugger entered--Lisp error: (file-error "Doing vfork" "Resource
> >>>>>>>temporarily unavailable")
> >>>>>>[...]
> >>>>>Just FYI, I don't know yet what happens exactly, but this has nothing
> >>>>>to do with the alternate stack.  The child process fails with a status
> >>>>>code 0xC00000FD, STATUS_STACK_OVERFLOW.  Which is kind of weird, given
> >>>>>that the stack overflow has been averted by calling siglongjmp.
> >>>>>
> >>>>>I have a hunch.  The stack state in the parent is so that TEB::StackLimit
> >>>>>points into the topmost guard area which, when poked into, triggers the
> >>>>>stack overflow exception.  When forking, Cygwin performs exactly this:
> >>>>>It pokes into the stack to push the guard page out of the way, thus
> >>>>>causing the stack memory to be commited, which in turn allows to copy
> >>>>>the stack content from parent to child.
> >>>>>
> >>>>>Ok, I'm not sure if I can debug this soon, but at leats it's not
> >>>>>related to sigaltstack handling nor is it a regression.
> >>>>
> >>>>Thanks for the info, that's good to know.  Just out of curiosity, were you
> >>>>able to modify your testcase for this, or did you test with emacs?
> >>>
> >>>I just added a fork call to my testcase right after the last printf.
> >>
> >>My hunch was correct, apparently.  I changed the way the stack info
> >>is set up for the child so only the actually used part of the stack is
> >>prepared for the stack copy in the child.  This not only avoids the
> >>stack overflow in the child, it should shave a few nanoseconds from
> >>the time a fork takes ;)
> >>
> >>I uploaded new developer snapshots to https://cygwin.com/snapshots/ and
> >>I'm just building and uploading a new test release.
> >>
> >>Please give it another try.
> >
> >That fixes it.  Thanks!
> I may have spoken too soon.  As I repeat the experiment on a different
> computer, with a build from a slightly different snapshot of the emacs
> trunk, emacs crashes when I type 'C-x d' with the following stack dump:
> Stack trace:
> Frame        Function    Args
> 00100A3E240  00180071CC3 (00000829630, 000008296D0, 00000000000, 0000082CE00)
> 00030000002  001800732BE (00000000000, 00000000002, 00100A48C80, 00000000002)
> 00000000000  00000006B40 (00000000002, 00100A48C80, 00000000002, 00100A48768)
> 00000000000  21000000003 (00000000002, 00100A48C80, 00000000002, 00100A48768)
> End of stack trace
> $ addr2line 00180071CC3 -e /usr/lib/debug/usr/bin/cygwin1.dbg
> /usr/src/debug/cygwin-2.1.0-0.3/winsup/cygwin/exception.h:175
> $ addr2line 001800732BE -e /usr/lib/debug/usr/bin/cygwin1.dbg
> /usr/src/debug/cygwin-2.1.0-0.3/winsup/cygwin/exceptions.cc:1639

That points to a crash while setting up the alternate stack.  This is
always a possibility because, in contrast to the kernel signal handler
in a real POSIX system, the Cygwin exception handler is still running on
the stack which triggered the crash up to the point where we call the
signal handler function.  Dependent on how the stack overflow occured,
this additional stack usage may be enough to kill the process for good.

Out of curiosity, can you add this to the init_sigsegv() function:

  #include <windows.h>
  init_sigsegv (void)
    SetThreadStackGuarantee (65536);

And see if that "fixes" the crash?

> The next two days are pretty busy for me, but I'll try to provide further
> information as soon as I have a chance.

Thanks a lot,

Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Maintainer                 cygwin AT cygwin DOT com
Red Hat
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://cygwin.com/pipermail/cygwin/attachments/20150702/a1fca579/attachment.sig>

More information about the Cygwin mailing list