On 06/15/2009 04:45 PM, Anthony Liguori wrote:
> Avi Kivity wrote:
>> On 06/15/2009 04:20 PM, Anthony Liguori wrote:
>>> It's not at all that simple. SCSI has a hierarchical address
>>> mechanism with 0-7 targets but then potentially multiple LUNs per
>>> target. Today, we always emulate a single LUN per target but if we
>>> ever wanted to support more than 7 disks on a SCSI controller, we
>>> would have to add multiple LUN support too. So the current linear
>>> unit= parameter is actually pretty broken for SCSI.
>>
>> Well, another level in the hierarchy, but I don't think it materially
>> changes things.
>
> Depends on whether you expect to say index=0,lun=3 or index=3. If you
> mean the later, then it's quite conceivable that each target supports
> less than the maximum number of LUNs. This makes things pretty
> confusing to the user because they have to know that in the current
> implementation, index=0 is valid, index=1 isn't, but index=8 is.
I'd object to any implicit addressing rules. If we have to say
target=2,lun=7,street=8,city=9,state=99,zip=12345 instead of
index=8345345235 so be it.
>>> No, I meant drive file=foo.img,bus=3. If that doesn't seem obvious
>>> what it should do to you that's because it isn't at all obvious :-)
>>> It ends up skipping a predefined number of locations in the drive
>>> table. This is pretty broken fundamentally because it assumes
>>> controllers always support a fixed number of devices. Nothing
>>> really respects bus_id though so in practice, I assume it's almost
>>> universally broken.
>>
>> Isn't the drive table something totally internal? And how does bus=
>> relate to it?
>
> The reality of unit=X,bus=Y,if=Z is that they expand to:
>
> drive_table_index=Y*max_devs[Z] + X
>
> Whereas max_devs = {"ide":4, "scsi": 7, *:0}
>
> How drive_table_index is interpreted is "if" specific. For if=scsi,
> each lsi device gets a base drive table index that starts at bus_index
> * 7. For virtio, the first empty spot in drive_table results in no
> more drives being created.
>
> It's broken by design.
Agreed. Pity that it's exposed to the poor users.
--
error compiling committee.c: too many arguments to function
_______________________________________________
Virtualization mailing list
Virtualization@xxxxxxxxxxxxxxxxxxxxxxxxxx
https://lists.linux-foundation.org/mailman/listinfo/virtualization
