[Qemu-devel] [Bug 1290370] [NEW] FreeBSD 9.2 shell crashes when run with -smp 4 option
Public bug reported: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git configured with target list i386-softmmu and then make OS: FreeBSD 9.2 Text Install ISO (ftp://ftp.freebsd.org/pub/FreeBSD/releases/i386/i386/ISO-IMAGES/9.2/FreeBSD-9.2-RELEASE-i386-disc1.iso) Installed it to a qcow2 format image. ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458 The crash happens randomly. It is not just related with filebench. Here are a few scenarios: * When i run fileserver workload of filebench * After i issue the shutdown -h now shutdown -r now commands * Issuing mount -t linprocfs proc /proc Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). Basically after running some commands and getting the CPU and the kernel worked up i think. NOTE: This bug does not occur in qemu-kvm ** Affects: qemu Importance: Undecided Status: New ** Tags: freebsd ** Description changed: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git configured with target list i386-softmmu - and then + and then make - + OS: FreeBSD 9.2 Text Install ISO Installed it to a qcow2 format image. ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] - Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 + Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458 - The crash happens randomly. It is not just related with filebench. + The crash happens randomly. It is not just related with filebench. Here are a few scenarios: * When i run fileserver workload of filebench * After i issue the shutdown -h now shutdown -r now commands * Issuing mount -t linprocfs proc /proc - Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). + Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). Basically after running some commands and getting the CPU and the kernel worked up i think. + + NOTE: This bug does not occur in qemu-kvm ** Description changed: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git configured with target list i386-softmmu and then make - OS: FreeBSD 9.2 Text Install ISO + OS: FreeBSD 9.2 Text Install ISO (ftp://ftp.freebsd.org/pub/FreeBSD/releases/i386/i386/ISO-IMAGES/9.2/FreeBSD-9.2-RELEASE-i386-disc1.iso) Installed it to a qcow2 format image. ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458 The crash happens randomly. It is not just related with filebench. Here are a few scenarios: * When i run fileserver workload of filebench * After i issue the shutdown -h now shutdown -r now commands * Issuing mount -t linprocfs proc /proc Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). Basically after running some commands and ge
[Qemu-devel] [Bug 1290370] Re: FreeBSD 9.2 shell crashes when run with -smp 4 option
** Description changed: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git configured with target list i386-softmmu and then make OS: FreeBSD 9.2 Text Install ISO (ftp://ftp.freebsd.org/pub/FreeBSD/releases/i386/i386/ISO-IMAGES/9.2/FreeBSD-9.2-RELEASE-i386-disc1.iso) Installed it to a qcow2 format image. ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458 The crash happens randomly. It is not just related with filebench. Here are a few scenarios: * When i run fileserver workload of filebench * After i issue the shutdown -h now shutdown -r now commands * Issuing mount -t linprocfs proc /proc Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). Basically after running some commands and getting the CPU and the kernel worked up i think. - NOTE: This bug does not occur in qemu-kvm + NOTE: This bug does not occur in qemu-kvm [X] + EDIT: Confirmed. This happens on qemu-kvm as well. ** Description changed: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git - configured with target list i386-softmmu - and then - make + qemu 1.7.50 configured with target list i386-softmmu. + OS: FreeBSD 9.2 Text Install ISO (ftp://ftp.freebsd.org/pub/FreeBSD/releases/i386/i386/ISO-IMAGES/9.2/FreeBSD-9.2-RELEASE-i386-disc1.iso) Installed it to a qcow2 format image. - ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic - -net user -monitor stdio + Command to run QEMU:: + ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458 The crash happens randomly. It is not just related with filebench. Here are a few scenarios: * When i run fileserver workload of filebench * After i issue the shutdown -h now shutdown -r now commands * Issuing mount -t linprocfs proc /proc Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). - Basically after running some commands and getting the CPU and the kernel worked up i think. + The bug manifests after running some commands and getting the CPU and the kernel worked up. NOTE: This bug does not occur in qemu-kvm [X] EDIT: Confirmed. This happens on qemu-kvm as well. -- You received this bug notification because you are a member of qemu- devel-ml, which is subscribed to QEMU. https://bugs.launchpad.net/bugs/1290370 Title: FreeBSD 9.2 shell crashes when run with -smp 4 option Status in QEMU: New Bug description: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git qemu 1.7.50 configured with target list i386-softmmu. OS: FreeBSD 9.2 Text Install ISO (ftp://ftp.freebsd.org/pub/FreeBSD/releases/i386/i386/ISO-IMAGES/9.2/FreeBSD-9.2-RELEASE-i386-disc1.iso) Installed it to a qcow2 format image. Command to run QEMU:: ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458
[Qemu-devel] [Bug 1292037] [NEW] Solaris 10 x86 guest crashes qemu with -icount 1 option
*** This bug is a security vulnerability *** Public security bug reported: Commit: f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz ** Affects: qemu Importance: Undecided Status: New ** Tags: qemu solaris ** Summary changed: - Solaris crashes with -icount 1 + Solaris 10 x86 guest crashes with -icount 1 ** Summary changed: - Solaris 10 x86 guest crashes with -icount 1 + Solaris 10 x86 guest crashes with -icount 1 option ** Summary changed: - Solaris 10 x86 guest crashes with -icount 1 option + Solaris 10 x86 guest crashes qemu with -icount 1 option ** Information type changed from Public to Public Security ** Information type changed from Public Security to Private Security ** Information type changed from Private Security to Public Security ** Tags added: qemu -- You received this bug notification because you are a member of qemu- devel-ml, which is subscribed to QEMU. https://bugs.launchpad.net/bugs/1292037 Title: Solaris 10 x86 guest crashes qemu with -icount 1 option Status in QEMU: New Bug description: Commit: f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz To manage notifications about this bug go to: https://bugs.launchpad.net/qemu/+bug/1292037/+subscriptions
[Qemu-devel] [Bug 1292037] Re: Solaris 10 x86 guest crashes qemu with -icount 1 option
** Information type changed from Public Security to Public ** Description changed: ** Information type changed from Public to Public Security ** Information type changed from Public Security to Public -- You received this bug notification because you are a member of qemu- devel-ml, which is subscribed to QEMU. https://bugs.launchpad.net/bugs/1292037 Title: Solaris 10 x86 guest crashes qemu with -icount 1 option Status in QEMU: New Bug description: Commit: f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz To manage notifications about this bug go to: https://bugs.launchpad.net/qemu/+bug/1292037/+subscriptions
[Qemu-devel] [Bug 1290370] Re: FreeBSD 9.2 shell crashes when run with -smp 4 option
I tested on the commit f53f3d0a00b6df39ce8df -- You received this bug notification because you are a member of qemu- devel-ml, which is subscribed to QEMU. https://bugs.launchpad.net/bugs/1290370 Title: FreeBSD 9.2 shell crashes when run with -smp 4 option Status in QEMU: New Bug description: This is a bug that i have noticed in qemu 1.7.50 as well as 1.1.50. It was the latter that forced me to clone the repository to check if this is the case with the resent version as well . The latest commit on which the bug is found is f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git qemu 1.7.50 configured with target list i386-softmmu. OS: FreeBSD 9.2 Text Install ISO (ftp://ftp.freebsd.org/pub/FreeBSD/releases/i386/i386/ISO-IMAGES/9.2/FreeBSD-9.2-RELEASE-i386-disc1.iso) Installed it to a qcow2 format image. Command to run QEMU:: ./i386-softmmu/qemu-system-i386 -hda -m 2G -smp 4 -net nic -net user -monitor stdio (boot into multi-user mode)->(login to root account) I have the filebench benchmark installed on the image and when i run it the default root shell (csh) crashes with the error. [pid (csh) sigreturn eflag = 0x] Here is the piece of kernel code that is getting executed (i think) http://svnweb.freebsd.org/base/release/9.2.0/sys/i386/i386/machdep.c?view=markup#l1095 Here is a related bug https://www.virtualbox.org/ticket/458 The crash happens randomly. It is not just related with filebench. Here are a few scenarios: * When i run fileserver workload of filebench * After i issue the shutdown -h now shutdown -r now commands * Issuing mount -t linprocfs proc /proc Moreover it is not guaranteed that the above scenarios will reproduce it (reliably). The bug manifests after running some commands and getting the CPU and the kernel worked up. NOTE: This bug does not occur in qemu-kvm [X] EDIT: Confirmed. This happens on qemu-kvm as well. To manage notifications about this bug go to: https://bugs.launchpad.net/qemu/+bug/1290370/+subscriptions
[Qemu-devel] [Bug 1292037] Re: Solaris 10 x86 guest crashes qemu with -icount 1 option
** Description changed: Commit: f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz + + + UPDATE: + http://lists.gnu.org/archive/html/qemu-devel/2014-05/msg06365.html + + Workaround: Rename the kvmvapic.bin file under the pc-bios directory to + something different. ** Description changed: - Commit: f53f3d0a00b6df39ce8dfca942608e5b6a9a4f71 on qemu.git - Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz UPDATE: + Tested on QEMU v2.0.50 + Also affects OpenIndiana http://lists.gnu.org/archive/html/qemu-devel/2014-05/msg06365.html Workaround: Rename the kvmvapic.bin file under the pc-bios directory to something different. ** Description changed: Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz - UPDATE: Tested on QEMU v2.0.50 - Also affects OpenIndiana + Also affects OpenIndiana (151a8 - Server Build 32bit) http://lists.gnu.org/archive/html/qemu-devel/2014-05/msg06365.html Workaround: Rename the kvmvapic.bin file under the pc-bios directory to something different. -- You received this bug notification because you are a member of qemu- devel-ml, which is subscribed to QEMU. https://bugs.launchpad.net/bugs/1292037 Title: Solaris 10 x86 guest crashes qemu with -icount 1 option Status in QEMU: New Bug description: Solaris image: Solaris 10 x86 (32 bit) command: ./i386-softmmu/qemu-system-i386 -hda -m 2G -icount 1 -monitor stdio Crashes saying: qemu: Fatal: Raised interrupt while not in I/O function Host: ubuntu x86_64 3.2.0-56 generic intel xeon E5649 @ 2.53GHz UPDATE: Tested on QEMU v2.0.50 Also affects OpenIndiana (151a8 - Server Build 32bit) http://lists.gnu.org/archive/html/qemu-devel/2014-05/msg06365.html Workaround: Rename the kvmvapic.bin file under the pc-bios directory to something different. To manage notifications about this bug go to: https://bugs.launchpad.net/qemu/+bug/1292037/+subscriptions
[Qemu-devel] Build error /oslib-posix.c:153: undefined reference to `trace_qemu_anon_ram_free'
Hi, I am getting this error if i try to build qemu. libqemuutil.a(oslib-posix.o): In function `qemu_anon_ram_alloc': /util/oslib-posix.c:141: undefined reference to `trace_qemu_anon_ram_alloc' libqemuutil.a(oslib-posix.o): In function `qemu_anon_ram_free': /util/oslib-posix.c:153: undefined reference to `trace_qemu_anon_ram_free' collect2: ld returned 1 exit status This occurred when i performed bisection on current master (7721a3044234c46cd6f5f899e7467dc9351f3c8d) and an earlier commit. Now when i checkout master i keep getting this. Any idea what is wrong? Build works if i pull the whole repo again.
Re: [Qemu-devel] Build error /oslib-posix.c:153: undefined reference to `trace_qemu_anon_ram_free'
Thanks. Out-of-tree build worked (with make distclean before checking out new commit). On Tue, Jun 10, 2014 at 5:53 PM, Peter Maydell wrote: > On 10 June 2014 13:03, Sai Prajeeth wrote: > > Hi, > > I am getting this error if i try to build qemu. > > > > libqemuutil.a(oslib-posix.o): In function `qemu_anon_ram_alloc': > > /util/oslib-posix.c:141: undefined reference to > `trace_qemu_anon_ram_alloc' > > libqemuutil.a(oslib-posix.o): In function `qemu_anon_ram_free': > > /util/oslib-posix.c:153: undefined reference to > `trace_qemu_anon_ram_free' > > collect2: ld returned 1 exit status > > > > This occurred when i performed bisection on current master > > (7721a3044234c46cd6f5f899e7467dc9351f3c8d) and an earlier commit. Now > when i > > checkout master i keep getting this. Any idea what is wrong? > > > > Build works if i pull the whole repo again. > > Yeah, this kind of thing can happen if you bisect between > an old commit and a new one -- make clean and make distclean > don't know how to remove files which were created by the > build in the old commit but which don't get built in the new > one, and then the stale files can occasionally cause > problems with the build. > > You may be able to determine what the stale file is and > remove it, but usually the easiest thing to do in this > situation is simply to delete the build tree and do a > complete rebuild from scratch. (Being able to do that is > a good reason for doing your builds out-of-tree, > incidentally.) > > You might also try 'check out old commit; make distclean; > check out new commit; build'. > > thanks > -- PMM >
[Qemu-devel] ATA timeouts on Solaris/OpenIndiana/Opensolaris
Hi I am booting an OpenIndiana image on qemu by using the -smp 4 option. I seem to be running into this known bug http://docs.oracle.com/cd/E19253-01/820-5245/ggmsj/index.html Can someone tell me how can i go about doing workaround 2 on qemu?? Workaround 3,4 are not an option for me. I already tried workaround 5 and it does not work. Can someone help me ? Thanks.
Re: [Qemu-devel] ATA timeouts on Solaris/OpenIndiana/Opensolaris
Andreas script worked for me. Thank you. But i seem to be running into an issue with when i set multiple cores. (-smp cores=4). Multiple CPUs works fine (-smp 4) During boot i get these messages: NOTICE: System detected 4 cpus, but only 1 cpu(s) were enabled during boot NOTICE: Use "boot-ncpus" parameter to enable more CPU(s).See eeprom(1M) I tried eeprom but to no avail.. It seems there is an issue in the BIOS. Have any of you encountered such an issue before? On Wed, Apr 2, 2014 at 10:23 PM, Ben Taylor wrote: > > > > On Wed, Apr 2, 2014 at 11:51 AM, Andreas Färber wrote: > >> Hi, >> >> Am 02.04.2014 09:32, schrieb Sai Prajeeth: >> > I am booting an OpenIndiana image on qemu by using the -smp 4 option. I >> > seem to be running into this known >> > bug http://docs.oracle.com/cd/E19253-01/820-5245/ggmsj/index.html >> > >> > Can someone tell me how can i go about doing workaround 2 on qemu?? >> >> Here's my script for running OpenIndiana using AHCI (SATA): >> >> qemu-system-x86_64 -enable-kvm -smp cores=1,threads=1,sockets=1 -m 2G \ >> -netdev user -device e1000,macaddr=... \ >> -device ahci,id=ahci \ >> -device ide-drive,bus=ahci.0,drive=ahcihd \ >> -drive file=/.../openindiana-151a.img,format=raw,if=none,id=ahcihd \ >> >> Note that live migration doesn't work with AHCI ATM. > > > If you could start qemu with both ata and ahci, add a new lun to the ahci > controller, and use beadm to migrate to the achi disk, then disconnect the > ata disk/controller and bump the cores back to 4. > > I think that workaround 1 is not technically correct. however, you'd need > a really good understanding of Solaris/opensolaris/OI to force the system > to reenumerate the device path from a failsafe boot. But this is not for > the faint of heart. > > If it's a new install (ie, not much customization/data), why not just > start a new QEMU instance that uses the ahci controller? > > Ben > > > Ben > >
[Qemu-devel] Solaris 10 x86 not booting with -icount
Hi list, I am unable to boot the solaris 10 x86 (32-bit) operating system on qemu when i use the -icount 1 option. I get the error "qemu: Fatal: Raised an Interrupt while not in I/O function" I tried different values for icount but still i am not able to get it working. I compiled qemu from sources and the version is 1.1.50 However when i boot the same OS using qemu-system-i386 (with -icount 1 and -no-kvm flags) installed on the system (qemu-kvm 1.0) it runs fine. Can anyone help me out ? Thanks
[Qemu-devel] OpenIndiana x86 with tcg hangs/timeouts
Hi list, Many services timeout in OpenIndiana (151a8 Server Build 32 bit x86) during boot when i use the tcg accelerator. This is pushing the boot time of the OS to more than 45 mins depending on the number of CPUs activated. I did the tests with qemu-system-i386 -smp sockets=4,cores=1,threads=1 and used a drive (HDD) with AHCI controller -device ahci,id=ahci \ -device ide-drive,bus=ahci.0,drive=ahcihd \ -drive file=.../indiana.ahci,format=qcow2,if=none,id=ahcihd \ My host computer information: (uname-a) Linux 3.2.0-56-generic #86-Ubuntu SMP x86_64 x86_64 x86_64 GNU/Linux Distribution: Ubuntu 12.04.3 LTS CPU: Intel(R) Xeon(R) CPU E5649 @ 2.53GHz In general OpenIndiana with tcg is slow(very) as compared to ubuntu which is quite fast (multiple CPUs). Unable to boot on the following commits : 178ac111bca16c08a79b2609ebdc75197bea976a (v2.0.50) ba014af39c6f9ee5af70d39ef495ac9f2bd4a76c (v1.7.1) >From the kernel boot messages, the boot hangs after CPU 0 and 1 are initialized. Services time out on the following commits 4ce5bc2dd1914a706725186c6563e8f92eedfc84 (v1.1.2) 3ffb4001c2e28624cd59a2b3598858784d0718d0 (v1.0.1) Can anyone else confirm this?
[Qemu-devel] Commit 5b6fb069378e breaks Solaris/OpenIndiana support (icount)
Git bisection tells that 5b6fb069378e61c45c577bbec3d7ef60367f7e4c is the first bad commit that breaks support for Solaris/OpenIndiana 32-bit x86 guest with icount 1. Seems like the option rom (kvmvapic.bin) is causing problems. Currently the workaround is to not load the option rom. Command: qemu-system-i386 -hda .../solaris.qcow2 -icount 1 Hope someone can take a look or give me pointers on how to proceed to make a patch. Thanks !
[Qemu-devel] FreeBSD clock_gettime()
Hi list, I have done experiments and it seems that FreeBSD's libc function clock_gettime() results in a system call when running on QEMU whereas on hardware it does not. Does anybody know why? For those interested , you can find the clock_gettime libc call in freebsd source here: https://github.com/lattera/freebsd/blob/master/lib/libc/sys/clock_gettime.c Thanks!
[Qemu-devel] OpenBSD sched_yield() calls increase
Hi list, When running OpenBSD on QEMU without KVM, I see an increase in the number of sched_yield() system calls in certain multi-threaded benchmarks (sysbench). However while using KVM accelerator, the number of this system calls is minimal. Does any have any insight why this is happening? The execution of the benchmark is really slowed when not using KVM which is to be expected but why is there such as massive increase in sched_yield system calls? Moreover when using icount the number of sched_yield calls is further increased. Some Number: With KVM accelerator - #of sched_yields = 500 Using TCG - #of sched_yields = 2600 Using icount - #of sched_yields = 16000 Anyone has an idea why this might be happening? Thank you!
