Bug 591494
Summary: | Virtio: Transfer file caused guest in same vlan abnormally quit | |||
---|---|---|---|---|
Product: | Red Hat Enterprise Linux 6 | Reporter: | Amos Kong <akong> | |
Component: | qemu-kvm | Assignee: | Michael S. Tsirkin <mst> | |
Status: | CLOSED CURRENTRELEASE | QA Contact: | Virtualization Bugs <virt-bugs> | |
Severity: | medium | Docs Contact: | ||
Priority: | medium | |||
Version: | 6.0 | CC: | ailan, amit.shah, ddumas, gyue, mjenner, stanislav.polasek, syeghiay, tburke, virt-maint | |
Target Milestone: | rc | |||
Target Release: | --- | |||
Hardware: | All | |||
OS: | Linux | |||
Whiteboard: | ||||
Fixed In Version: | qemu-kvm-0.12.1.2-2.101.el6 | Doc Type: | Bug Fix | |
Doc Text: | Story Points: | --- | ||
Clone Of: | ||||
: | 606238 (view as bug list) | Environment: | ||
Last Closed: | 2010-11-10 21:24:53 UTC | Type: | --- | |
Regression: | --- | Mount Type: | --- | |
Documentation: | --- | CRM: | ||
Verified Versions: | Category: | --- | ||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | ||
Cloudforms Team: | --- | Target Upstream Version: | ||
Embargoed: | ||||
Bug Depends On: | ||||
Bug Blocks: | 606238 |
Description
Amos Kong
2010-05-12 12:17:23 UTC
This request was evaluated by Red Hat Product Management for inclusion in a Red Hat Enterprise Linux major release. Product Management has requested further review of this request by Red Hat Engineering, for potential inclusion in a Red Hat Enterprise Linux Major release. This request is not yet committed for inclusion. This bug was only reproduced for RHEL4.8 guest with virtio NIC. host kernel: 2.6.32-36.el6.x86_64 guest kernel: 2.6.9-89.28.ELsmp # rpm -qa |grep qemu qemu-kvm-0.12.1.2-2.77.el6.x86_64 gpxe-roms-qemu-0.9.7-6.3.el6.noarch qemu-kvm-tools-0.12.1.2-2.77.el6.x86_64 qemu-img-0.12.1.2-2.77.el6.x86_64 qemu-kvm-debuginfo-0.12.1.2-2.77.el6.x86_64 [ The test refult of different qemu option ] 1. Command line with '-net tap' option: Bug can be reproduced. Output: virtio-net truncating packet: offset 1514, size 1518, hdr_len 10 #qemu-kvm -name 'vm1' -monitor tcp:0:6002,server,nowait -drive file=/root/autotest/client/tests/kvm/images/RHEL-4.8-64.qcow2,if=virtio,cache=none,boot=on,aio=native -net nic,vlan=0,model=virtio,macaddr=02:B7:5E:9A:6f:36 -net tap,vlan=0,ifname=virtio_0_6002,downscript=no,vhost=on -m 512 -smp 1 -soundhw ac97 -vnc :1 -spice port=8001,disable-ticketing -serial unix:/tmp/serial-20100623-190659-BTxs,server,nowait 2. Command line with '-netdev tap' option: Guest could not crash, but file transfer always fail, I'll report another bug later with detail info. Output: virtio-net truncating packet: offset 1514, size 1518, hdr_len 10 #qemu -name 'vm1' -monitor unix:'/tmp/monitor-humanmonitor1-20100623-184427-46jt',server,nowait -drive file='/root/autotest/client/tests/kvm/images/RHEL-4.8-64.qcow2',if=virtio,cache=none,boot=on,aio=native -net nic,vlan=0,netdev=pO8A,model=virtio,macaddr='02:B7:5E:9A:93:15' -netdev tap,id=pO8A,ifname=virtio_0_8001,script=/root/autotest/client/tests/kvm/scripts/qemu-ifup-switch,downscript=no,vhost=on -m 512 -smp 1 -vnc :1 -spice port=8001,disable-ticketing -serial unix:/tmp/serial-20100623-184427-46jt,server,nowait Do we only need test with netdev option ? do you try to reproduce with '-net tap' option ? I'm testing with you provided packages, will reply the result later. The state of this package is failed. https://brewweb.devel.redhat.com/taskinfo?taskID=2541095 -- Bug also existed with this packages (using '-net tap' option) http://brewweb.devel.redhat.com/brew/taskinfo?taskID=2541095 yes, we only care about -netdev OK, I went over the code. There are several issues: 1. when offloads are disabled, rhel4.8 posts packets which are too short: #define MAX_PACKET_LEN (ETH_HLEN+ETH_DATA_LEN) this is fixed upstream e918085aaff34086e265f825dd469926b1aec4a4 2. qemu crashes when not enough buffers. should drop packet instead. 3. why are offloads disabled? I will send a debug breaw build to debug this. 4. vhost seems to truncate packet. should drop instead. I will send a debug breaw build to debug this. Tested with https://brewweb.devel.redhat.com/taskinfo?taskID=2545310 qemu of guest2 outputted: (using '-netdev' option, vhost=off) virtio-net truncating packet: i 1 mergeable 0 offset 1524, size 1528, guest hdr 10, host hdr 10 num 2 base 0x7f122f9bc270 len 10 size 1524 size0 10 # gdb -p $guest2_pid ... (gdb) c Continuing. [Thread 0x7f1234c0b710 (LWP 10160) exited] [New Thread 0x7f1234c0b710 (LWP 10166)] [Thread 0x7f1234c0b710 (LWP 10166) exited] [New Thread 0x7f1234c0b710 (LWP 10167)] [Thread 0x7f1234c0b710 (LWP 10167) exited] [New Thread 0x7f1234c0b710 (LWP 10169)] [Thread 0x7f1234c0b710 (LWP 10169) exited] [New Thread 0x7f1234c0b710 (LWP 10171)] [Thread 0x7f1234c0b710 (LWP 10171) exited] [Thread 0x7f120f065710 (LWP 10141) exited] [Thread 0x7f1236810710 (LWP 10129) exited] Program exited with code 01. (gdb) bt No stack. I think that with the original command line (-net with vlan=, no -netdev or id= in either guest 1 or guest 2), it will no longer crash with qemu-kvm-0.12.1.2-2.72.el6 Correct? (In reply to comment #22) > I think that with the original command line (-net with vlan=, no -netdev > or id= in either guest 1 or guest 2), it will no longer crash with > qemu-kvm-0.12.1.2-2.72.el6 > Correct? Incorrect, bug also existed. Please test with http://brewweb.devel.redhat.com/brew/taskinfo?taskID=2550177 this should fix the exit, but file transfer will still fail. (In reply to comment #24) > Please test with > http://brewweb.devel.redhat.com/brew/taskinfo?taskID=2550177 > this should fix the exit, but file transfer will still fail. yes. guest doesn't quit, and transfer file failed. host dmesg exists noise when vhost is enable. "Discarded truncated rx packet: len 1528 > 1524" When disable vhost, no noise. fixes sent. note we can not make vlans work without guest changes but we can make qemu not exit. Could not reproduce bug with qemu-kvm-0.12.1.2-2.104.el6 Move to VERIFIED. Red Hat Enterprise Linux 6.0 is now available and should resolve the problem described in this bug report. This report is therefore being closed with a resolution of CURRENTRELEASE. You may reopen this bug report if the solution does not work for you. |