This message is from the T13 list server.
> --- d1532v1r4.pdf ATA/PI 7 vol 1 of 3 - 2003-12-23
> ...
> 6.16 IDENTIFY DEVICE
> ...
> 6.16.71 Word 255: Integrity word The use of this word is optional.
> If bits (7:0) of this word contain the signature A5h, bits (15:8)
> contain the data structure checksum. The data structure checksum is
> the
> twoâs complement of the sum of all bytes in words (254:0) and the byte
> consisting of bits (7:0) in word 255. Each byte shall be added with
> unsigned arithmetic, and overflow shall be ignored. The sum of all 512
> bytes is zero when the checksum is correct.
> ...
Kindly offline I was asked to draw my attention the case of:
"a stuck pattern for the top eight bits of every word returned"
I agree the ATA/PI 5 "word" 255 "integrity word" checksum by design does
not detect that case.
Anybody agree?
I argue, the unsigned byte sum without overflow of x100 words with XX in
the upper byte is xXX00. Truncated to eight bits that is x00. All
traces gone. Whoops.
I mis/remember I've seen this before in byte sums of 16-wide rom chips.
Any unsigned checksum of a power of two of bus words should be at least
as wide as the bus, aye, else by definition it checks just for transient
faults and the lo bits.
Indisputable, yes? Academic, yes? Anybody disagree?
Versus wires stuck hi, stuck lo, or cut to float hi, this checksum
doesn't actually tell us much that read-write-compare of x1F4 couldn't
tell us already.
Indisputable, yes? Academic, yes? Anybody disagree?
So do we want a less weak "word" 254 "integrity word" ... or can we hope
PATA will go away anyhow before "word" 254 gets adopted ... or can we
tell people, please don't get your upper byte stuck e.g. by losing one
PCMCIA byte enable ...
... or is my logic somehow faulty?
Pat LaVarre
