Re: [Qemu-devel] [PATCH 4/4] scsi-disk: fix the block descriptor returned by the MODE SENSE command

[Bernhard Kohl]

Am 16.08.2010 19:34, schrieb ext Kevin Wolf:
> The patch itself looks okay. However, it made me wonder what this line
> wants to tell us:
>
>     if ((~dbd) & nb_sectors) {
>
> Is it just me or doesn't this make any sense at all? dbd is a single
> bit, 0x8 if set or 0x0 otherwise. nb_sectors is the number of sectors.
> Can this operation have any meaningful result? I suppose it was meant to
> be something like:
>
>     if (!dbd && nb_sectors) {
>
> Can you please check this and add a patch 5/4 if necessary?
For my opinion it is nonsense too. And it does not work. I tested it:

address@hidden ~ # sg_modes /dev/sdb -6 -v -H -H -p 0x3f
    inquiry cdb: 12 00 00 00 24 00
    QEMU      QEMU HARDDISK     0.13   peripheral_type: disk [0x0]
    mode sense (6) cdb: 1a 00 3f 00 fc 00
    mode sense (6): requested 252 bytes but got 220 bytes
Mode parameter header from MODE SENSE(6):
 00     1f 00 00 08
  Mode data length=32, medium type=0x00, WP=0, DpoFua=0, longlba=0
  Block descriptor length=8
> Direct access device block descriptors:
   Density code=0x0
 00     00 a0 00 00 00 00 02 00

>> page_code=0x8, page_control=0
 00     08 12 00 00 00 00 00 00  00 00 00 00 00 00 00 00
 10     00 00 00 00
address@hidden ~ # sg_modes /dev/sdb -6 -v -H -H -p 0x3f -d
    inquiry cdb: 12 00 00 00 24 00
    QEMU      QEMU HARDDISK     0.13   peripheral_type: disk [0x0]
    mode sense (6) cdb: 1a 08 3f 00 fc 00
    mode sense (6): requested 252 bytes but got 220 bytes
Mode parameter header from MODE SENSE(6):
 00     1f 00 00 08
  Mode data length=32, medium type=0x00, WP=0, DpoFua=0, longlba=0
  Block descriptor length=8
> Direct access device block descriptors:
   Density code=0x0
 00     00 a0 00 00 00 00 02 00

>> page_code=0x8, page_control=0
 00     08 12 00 00 00 00 00 00  00 00 00 00 00 00 00 00
 10     00 00 00 00
address@hidden ~ #

I will add a patch 5/5 in v2 as you proposed.

Bernhard