Hello Haixu,
There are such strange usage scenarios that frame size exceeds 32
bits, even 4096 bits. For these strange cases, It's impossible to use
only 32 bits bits_per_word_mask to enumerate all the conditions. Seems
that a similar situation exists in Linux spi driver also.
Assume a case where the spi controller supports the frame size to be
any value from 16 to 64, 32 bits mask can only cover the 16 ~ 32
conditions.
Even so, it's not a good idea to expand the bits_per_word_mask, 32
bits can cover most of the cases, except for some situations that are
not commonly used. Besides, it's hard to determine the bit width to
cover all the cases, and expanding will make the logic more
complicated. To summarize, I think the following statement is probably
proper:
"\field{bits_per_word_mask} is a mask indicating which values of
bits_per_word are supported. If bit n of \field{bits_per_word_mask} is
set, the bits_per_word with value (n+1) is supported. If
\field{bits_per_word_mask} is 0, there is no limitation for
bits_per_word."
Do you think it is acceptable and appropriate?
Perfect.
Currently in the middle (almost done) of updating my software to draft V10.
Regards
Harald Mommer
---------------------------------------------------------------------
To unsubscribe, e-mail: virtio-dev-unsubscr...@lists.oasis-open.org
For additional commands, e-mail: virtio-dev-h...@lists.oasis-open.org
