|
From: | Vikas Desai |
Subject: | Re: [Qemu-devel] Compiling QEMU x86_64 for windows 64 bit |
Date: | Fri, 27 Sep 2013 12:44:32 +0800 |
Thanks Stefan. Appreciate the quick response.
I tried recompiling with --enable-debug but the binaries still die without any error message. I am using mingw64 + MSYS and SDL. I would probably take the easy route and wait for you to publish the new builds :) Cheers, Vikas > Date: Thu, 26 Sep 2013 23:13:58 +0200 > From: address@hidden > To: address@hidden > CC: address@hidden; address@hidden > Subject: Re: [Qemu-devel] Compiling QEMU x86_64 for windows 64 bit > > Am 26.09.2013 21:05, schrieb Stefan Weil: > > Am 26.09.2013 13:23, schrieb Vikas Desai: > >> Hi, > >> > >> After some further testing I found that even the 32 bit binaries from > >> Stefan fail with the same error. I tried the 32 bit binaries from by > >> Eric Lassauge for version 1.6 and they work well. I have tried both 32 > >> and 64 bit binaries from Stefan on 2 different environments, both > >> failing with same errors. > >> > >> When I just run the binaries with no disk image or any other options, > >> I get a proper window with the BIOS going through all drives looking > >> for a bootable device. Only when I have a valid executable image I get > >> the error. Also, in case of the test linux binary I get a kernel panic > >> on linux but qemu does not crash. > >> > >> What should I do further to debug this? > >> > >> Hi Stefan, > >> > >> Could you share what tools you use for the build? Any hints on what > >> more could I try? > >> > >> Thanks, > >> Vikas > > Hi Vikas, > > > > I also get the corouting assertion when I start my precompiled QEMU > > binary with an ISO image (Debian i386 netinstall). > > The error can be reproduced with Wine on Linux, too. > > > > There is no error when QEMU was configured with --enable-debug (which > > disables optimisation), > > nor is there an error when I just run the BIOS code (no disk, no cdrom). > > This explains why I did not > > notice the regression for Windows earlier. > > > > So we have to find the first version which shows that regression, either > > by testing older installers > > or by running git bisect. > > > > Cheers, > > Stefan > > Summary: > > Latest qemu-system-i386 for Windows fails with an assertion > (qemu-coroutine-lock.c:99) > if something more complex than the BIOS is executed. It works when it is > configured with > --enable-debug. This behaviour is identical for 32 bit and 64 bit > executables and can also > be reproduced using Wine. Older versions also fail, but with SIGSEGV > instead of an assertion. > > This is the result of git bisect: > > 402397843e20e35d6cb7c80837c7cfdb19ede591 is the first bad commit > commit 402397843e20e35d6cb7c80837c7cfdb19ede591 > Author: Paolo Bonzini <address@hidden> > Date: Tue Feb 19 11:59:09 2013 +0100 > > coroutine: move pooling to common code > > The coroutine pool code is duplicated between the ucontext and > sigaltstack backends, and absent from the win32 backend. But the > code can be shared easily by moving it to qemu-coroutine.c. > > Signed-off-by: Paolo Bonzini <address@hidden> > Reviewed-by: Stefan Hajnoczi <address@hidden> > Signed-off-by: Kevin Wolf <address@hidden> > > When I configure latest QEMU with --disable-coroutine-pool, it works! > > I'll build new installers with this option until there is a bug fix > available. > > Thanks for your bug report. > > Stefan > |
[Prev in Thread] | Current Thread | [Next in Thread] |