>
>
>Wow, that's definitely disk-dependent unless our driver (or FreeBSD's) is
>lying:
>
Maybe its just NetBSD? After viewing the KB/s versus transfer/s rates below
and performing some simple arithmetic its easy to see that Solaris reports
128 as well on both x86 and UltraSPARC using two different types of ATA
drives.
north_

blueberry# (dd if=/dev/rdsk/c0t0d0s0 of=/dev/null count=1024 
bs=128k)>&/dev/null & iostat -dl 1 dad0 1 6
[1] 8507
    dad0     
kps tps serv 
  4   0   10 
40886 319    3 
42115 329    3 
41984 328    3 
5761  45    3 
  0   0    0 
[1]  + Done                 ( dd if=/dev/rdsk/c0t0d0s0 of=/dev/null 
count=1024 bs=128k ) >& /dev/null
blueberry# uname -a
SunOS blueberry 5.9 Generic_112233-01 sun4u sparc SUNW,Sun-Blade-100
blueberry# dmesg | awk '{if($5 == "dada:") print $0}' -
Jan 31 03:23:29 blueberry dada: [ID 365881 kern.info]   <ST320011A cyl 
38790 alt 2 hd 16 sec 63>
blueberry#

---

givingtree# (dd if=/dev/rdsk/c0d0p0 of=/dev/null count=200 
bs=128k)>&/dev/null & iostat -dl 1 cmdk0 1 6
[1] 13213
   cmdk0     
kps tps serv 
183  11   66 
4173  34   29 
4292  34   30 
4287  33   30 
4270  33   30 
  0   0    0 
[1]  + Done                 ( dd if=/dev/rdsk/c0d0p0 of=/dev/null 
count=200 bs=128k ) >& /dev/null
givingtree# uname -a
SunOS givingtree 5.8 Generic i86pc i386 i86pc
givingtree#
Feb  3 19:27:47 givingtree ata: [ID 640982 kern.info]   IDE device at 
targ 0, lun 0 lastlun 0x0
Feb  3 19:27:47 givingtree ata: [ID 521533 kern.info]   model Maxtor 
90840D6, stat 50, err 0
Feb  3 19:27:47 givingtree ata: [ID 241969 kern.info]           cfg 
0x40, cyl 16276, hd 16, sec/trk 63
Feb  3 19:27:47 givingtree ata: [ID 674665 kern.info]           mult1 
0x8010, mult2 0x110, dwcap 0x0, cap 0x2f00
Feb  3 19:27:47 givingtree ata: [ID 350272 kern.info]           piomode 
0x200, dmamode 0x200, advpiomode 0x3
Feb  3 19:27:47 givingtree ata: [ID 245087 kern.info]           minpio 
120, minpioflow 120
Feb  3 19:27:47 givingtree ata: [ID 435839 kern.info]           valid 
0x7, dwdma 0x7, majver 0x1e