Re: [PATCH 10 of 20] ipath - core driver, part 3 of 4

[Linus Torvalds]

All your user page lookup/pinning code is terminally broken.

You can't do it that way. You have serveral major conceptual bugs, like 
keeping track of pages without incrementing their page count, and just 
expecting that they are magically "pinned" even you do nothing at all to 
pin them. The process exits or does an munmap, and the page will be used 
for something else, and you'll just corrupt totally random memory.

Similarly, you do page_address() on the page, which just can't work on 
highmem pages.

Crap like this must not be merged. Drivers aren't supposed to play VM 
tricks in the first place - even if they were to get it right (which they 
never do). Don't do it.

		Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/