Am 17.10.2011 16:04, schrieb Paolo Bonzini: > On 10/17/2011 04:05 PM, Kevin Wolf wrote: >>>> diff --git a/hw/ide/atapi.c b/hw/ide/atapi.c >>>> index 347c38d..10f161f 100644 >>>> --- a/hw/ide/atapi.c >>>> +++ b/hw/ide/atapi.c >>>> @@ -731,6 +731,8 @@ static void cmd_mode_sense(IDEState *s, uint8_t *buf) >>>> buf[7] = 0; >>>> >>>> /* Fill with CDROM audio volume */ >>>> + buf[8] = 0x0e; >> >> 0x0e would be more readable if it was written as MODE_PAGE_AUDIO_CTL. >> >>>> + buf[9] = 0x0e; >> >> And this is really the length. But I'm not sure if writing it as 24 - 10 >> would help a lot. > > Yes, the coincidence is ugly.
Maybe we should define structs for the mode pages so that you have real names instead of just offsets into the output buffer? >> Of course the same is true for the other pages, so it could be done for >> all of them on top. Or maybe you actually have included it later in the >> series? > > Only for SCSI. I can queue it as a cleanup for ATAPI. That would be good. Kevin
