On 28.01.2014 14:00, Eduardo Otubo wrote:
> On 01/26/2014 10:21 AM, Felix Geyer wrote:
>> libusb calls timerfd_create() and timerfd_settime() when it's built with
>> timerfd support.
>>
>> Command to reproduce:
>>
>> qemu -sandbox on -monitor stdio -device piix3-usb-uhci,id=usb
>> -device usb-host,hostbus=1,hostaddr=3,id=hostdev0
>>
>> Log messages:
>>
>> audit(1390730418.924:135): auid=4294967295 uid=121 gid=103 ses=4294967295
>> pid=5232 comm="qemu-system-x86" sig=31
>> syscall=283
>> compat=0 ip=0x7f2b0f4e96a7 code=0x0
>> audit(1390733100.580:142): auid=4294967295 uid=121 gid=103 ses=4294967295
>> pid=16909 comm="qemu-system-x86" sig=31
>> syscall=286
>> compat=0 ip=0x7f03513a06da code=0x0
>>
>> Signed-off-by: Felix Geyer <de...@fobos.de>
>> ---
>> qemu-seccomp.c | 4 +++-
>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/qemu-seccomp.c b/qemu-seccomp.c
>> index caa926e..2705468 100644
>> --- a/qemu-seccomp.c
>> +++ b/qemu-seccomp.c
>> @@ -225,7 +225,9 @@ static const struct QemuSeccompSyscall
>> seccomp_whitelist[] = {
>> { SCMP_SYS(fchmod), 240 },
>> { SCMP_SYS(shmget), 240 },
>> { SCMP_SYS(shmat), 240 },
>> - { SCMP_SYS(shmdt), 240 }
>> + { SCMP_SYS(shmdt), 240 },
>> + { SCMP_SYS(timerfd_create), 240 },
>> + { SCMP_SYS(timerfd_settime), 240 }
>
> Did you deliberately set the priority to 240? Or did you run any sort of
> benchmark (strace) to
> find this value?
>
> Regards,
Not really, sorry.
I've now done a benchmark on x86_64, copying a few hundred MB from a USB drive:
calls syscall
--------- ----------------
5303600 write
2240554 read
2167030 ppoll
2134828 ioctl
704023 timerfd_settime
689105 poll
83122 futex
803 writev
476 rt_sigprocmask
287 recvmsg
178 brk
timerfd_create is basically only called once so it can have the lowest priority.
timerfd_settime should probably have priority around 242.
Regards,
Felix
