[Qemu-devel] [Bug 1639791] [NEW] early virtio console output is lost

[Qemu-devel] [Bug 1639791] [NEW] early virtio console output is lost

[Michal Suchanek]

Public bug reported:

This is broken in git and reportedly in 2.5 through 2.7.

Running a Linux kernel which includes a testsuite in initrd sometimes
produces no output.

Reportedly the console is sometimes not open when the early userspace
tries to log output resulting in either the testsuite terminating early
or not writing the output.

Workaround patch is here:

https://git.zx2c4.com/WireGuard/commit/?id=d2de8b0862a7fbb51a7f2f958d58f0efe4648259

reportedly you would get -EBADF there when no output is generated.

Also this reportedly happens with virtio console only, not virtio serial
port.

It seems that the author of said testsuite did not report the problem so
I write it down so it does not get lost.

test (in bash):

n=0 ; while [ $n -lt 100 ] && grep -m 1 -F "WireGuard Test Suite on
Linux 4.8.6" <( /opt/qemu/bin/qemu-system-x86_64         -nodefaults
-nographic         -machine q35,accel=kvm         -cpu host         -smp
2         -m 64M         -object rng-
random,id=rng0,filename=/dev/urandom         -device virtio-rng-
pci,rng=rng0         -device virtio-serial,max_ports=2         -chardev
stdio,id=stdio         -device virtconsole,chardev=stdio
-chardev file,id=status,path=result.txt         -device
virtserialport,chardev=status         -monitor none         -kernel
wireguard-testing-harness-bzImage-e87cb2a7-145c-4985-907f-17e81fae329b
-append "console=hvc0 initcall_debug=1 loglevel=7" ) ; do echo $n ;
n=$(expr $n + 1) ; pkill -f wireguard ; done

This typically does 10-20 iterations but sometimes tens of iterations
run without issue.

** Affects: qemu
     Importance: Undecided
         Status: New

** Attachment added: "test kernel"
   https://bugs.launchpad.net/bugs/1639791/+attachment/4774019/+files/wireguard-testing-harness-bzImage-e87cb2a7-145c-4985-907f-17e81fae329b

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1639791

Title:
  early virtio console output is lost

Status in QEMU:
  New

Bug description:
  This is broken in git and reportedly in 2.5 through 2.7.

  Running a Linux kernel which includes a testsuite in initrd sometimes
  produces no output.

  Reportedly the console is sometimes not open when the early userspace
  tries to log output resulting in either the testsuite terminating
  early or not writing the output.

  Workaround patch is here:

  https://git.zx2c4.com/WireGuard/commit/?id=d2de8b0862a7fbb51a7f2f958d58f0efe4648259

  reportedly you would get -EBADF there when no output is generated.

  Also this reportedly happens with virtio console only, not virtio
  serial port.

  It seems that the author of said testsuite did not report the problem
  so I write it down so it does not get lost.

  test (in bash):

  n=0 ; while [ $n -lt 100 ] && grep -m 1 -F "WireGuard Test Suite on
  Linux 4.8.6" <( /opt/qemu/bin/qemu-system-x86_64         -nodefaults
  -nographic         -machine q35,accel=kvm         -cpu host
  -smp 2         -m 64M         -object rng-
  random,id=rng0,filename=/dev/urandom         -device virtio-rng-
  pci,rng=rng0         -device virtio-serial,max_ports=2
  -chardev stdio,id=stdio         -device virtconsole,chardev=stdio
  -chardev file,id=status,path=result.txt         -device
  virtserialport,chardev=status         -monitor none         -kernel
  wireguard-testing-harness-bzImage-e87cb2a7-145c-4985-907f-17e81fae329b
  -append "console=hvc0 initcall_debug=1 loglevel=7" ) ; do echo $n ;
  n=$(expr $n + 1) ; pkill -f wireguard ; done

  This typically does 10-20 iterations but sometimes tens of iterations
  run without issue.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1639791/+subscriptions

[Bug 1639791] Re: early virtio console output is lost

[Thomas Huth]

The QEMU project is currently considering to move its bug tracking to another system. For this we need to know which bugs are still valid and which could be closed already. Thus we are setting all older bugs to
"Incomplete" now.
If you still think this bug report here is valid, then please switch the state back to "New" within the next 60 days, otherwise this report will be marked as "Expired". Thank you and sorry for the inconvenience.


** Changed in: qemu
       Status: New => Incomplete

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1639791

Title:
  early virtio console output is lost

Status in QEMU:
  Incomplete

Bug description:
  This is broken in git and reportedly in 2.5 through 2.7.

  Running a Linux kernel which includes a testsuite in initrd sometimes
  produces no output.

  Reportedly the console is sometimes not open when the early userspace
  tries to log output resulting in either the testsuite terminating
  early or not writing the output.

  Workaround patch is here:

  https://git.zx2c4.com/WireGuard/commit/?id=d2de8b0862a7fbb51a7f2f958d58f0efe4648259

  reportedly you would get -EBADF there when no output is generated.

  Also this reportedly happens with virtio console only, not virtio
  serial port.

  It seems that the author of said testsuite did not report the problem
  so I write it down so it does not get lost.

  test (in bash):

  n=0 ; while [ $n -lt 100 ] && grep -m 1 -F "WireGuard Test Suite on
  Linux 4.8.6" <( /opt/qemu/bin/qemu-system-x86_64         -nodefaults
  -nographic         -machine q35,accel=kvm         -cpu host
  -smp 2         -m 64M         -object rng-
  random,id=rng0,filename=/dev/urandom         -device virtio-rng-
  pci,rng=rng0         -device virtio-serial,max_ports=2
  -chardev stdio,id=stdio         -device virtconsole,chardev=stdio
  -chardev file,id=status,path=result.txt         -device
  virtserialport,chardev=status         -monitor none         -kernel
  wireguard-testing-harness-bzImage-e87cb2a7-145c-4985-907f-17e81fae329b
  -append "console=hvc0 initcall_debug=1 loglevel=7" ) ; do echo $n ;
  n=$(expr $n + 1) ; pkill -f wireguard ; done

  This typically does 10-20 iterations but sometimes tens of iterations
  run without issue.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1639791/+subscriptions

[Bug 1639791] Re: early virtio console output is lost

[Launchpad Bug Tracker]

[Expired for QEMU because there has been no activity for 60 days.]

** Changed in: qemu
       Status: Incomplete => Expired

-- 
You received this bug notification because you are a member of qemu-
devel-ml, which is subscribed to QEMU.
https://bugs.launchpad.net/bugs/1639791

Title:
  early virtio console output is lost

Status in QEMU:
  Expired

Bug description:
  This is broken in git and reportedly in 2.5 through 2.7.

  Running a Linux kernel which includes a testsuite in initrd sometimes
  produces no output.

  Reportedly the console is sometimes not open when the early userspace
  tries to log output resulting in either the testsuite terminating
  early or not writing the output.

  Workaround patch is here:

  https://git.zx2c4.com/WireGuard/commit/?id=d2de8b0862a7fbb51a7f2f958d58f0efe4648259

  reportedly you would get -EBADF there when no output is generated.

  Also this reportedly happens with virtio console only, not virtio
  serial port.

  It seems that the author of said testsuite did not report the problem
  so I write it down so it does not get lost.

  test (in bash):

  n=0 ; while [ $n -lt 100 ] && grep -m 1 -F "WireGuard Test Suite on
  Linux 4.8.6" <( /opt/qemu/bin/qemu-system-x86_64         -nodefaults
  -nographic         -machine q35,accel=kvm         -cpu host
  -smp 2         -m 64M         -object rng-
  random,id=rng0,filename=/dev/urandom         -device virtio-rng-
  pci,rng=rng0         -device virtio-serial,max_ports=2
  -chardev stdio,id=stdio         -device virtconsole,chardev=stdio
  -chardev file,id=status,path=result.txt         -device
  virtserialport,chardev=status         -monitor none         -kernel
  wireguard-testing-harness-bzImage-e87cb2a7-145c-4985-907f-17e81fae329b
  -append "console=hvc0 initcall_debug=1 loglevel=7" ) ; do echo $n ;
  n=$(expr $n + 1) ; pkill -f wireguard ; done

  This typically does 10-20 iterations but sometimes tens of iterations
  run without issue.

To manage notifications about this bug go to:
https://bugs.launchpad.net/qemu/+bug/1639791/+subscriptions