Re: Physical memory topology when locking virtual memory

Hy Maxim,
Since I'm working on drivers that performs DMAs (a shared DMA those times)
with HARDWARE DEVICES, I have to deel with PHYSICAL MEMORY considerations at
last.
Anyway, Tim answered to my question below and I did not know that aspect
that seems to be a hardware consideration too.
Thanks anyway for all,
Francois.


"Maxim S. Shatskih" wrote:

Then I would not rely on anything about them.

--
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
maxim@xxxxxxxxxxxxxxxx
http://www.storagecraft.com

"Francois Champs" <FrancoisChamps@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:04CEF209-D77E-49BB-BF63-292625F71D86@xxxxxxxxxxxxxxxx
Yes they are DMA scatter-gather list entries provided by the DMA adapter.

"Maxim S. Shatskih" wrote:

What are "physical blocks"? MDLs? or DMA scatter-gather list entries?

--
Maxim Shatskih, Windows DDK MVP
StorageCraft Corporation
maxim@xxxxxxxxxxxxxxxx
http://www.storagecraft.com

"Francois Champs" <FrancoisChamps@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote in
message
news:6A4977D2-D848-421F-A204-B147FCB734BB@xxxxxxxxxxxxxxxx
Dear All,
When parsing the physical blocks describing a virtual memory portion once
locked I can see the following:
- the first block may start in the middle of a 4k page and its end is
page
aligned
- the same way, the last block starts page aligned and ends in the middle
of
a page
- all other blocks are page aligned (address and size), most of the time
one
page size, sometimes more.

My question is:
Is there a chance in normal conditions that the blocks other than fist
and
end may not be paged aligned in address and / or size???
Or should I consider this should never occur????

Many thanks in advance and have a nice day all!!!
Francois.





.