On Thu, 5 May 2011, Christian Robottom Reis wrote:
On Thu, May 05, 2011 at 07:28:33PM -0400, Nicolas Pitre wrote:
On Thu, 5 May 2011, Christian Robottom Reis wrote:
Last quasi-random question of the night: would reviving the kmemcheck ARM port (that IIRC was hacked up a while back) be something useful, or is it something that is too niche to be worth it? At the board's request, I was looking across platform gaps and this feature was one of the things that I found.
I can't find any ARM port for this. And this is not exactly trivial to make this work on ARM either (even though this would be an interesting challenge).
Hmmm. I thought I had seen one. Mmm. What about
http://www.spinics.net/lists/arm-kernel/msg56221.html
This won't work. The fundamental part for this feature is the kmemcheck_hide_pages() function which has about a dozen lines. But it assumes that kernel memory is mapped with 4KB page granularity which on ARM it is not. That's the prerequisite part I was talking about.
The prerequisite for kmemcheck on ARM would be to have the kernel direct mapping of memory use second level page tables, which might possibly be useful to the multimedia memory allocator work too.
Does the patch above implement that, though?
No. It merely copied the x86 code, stripped some of it about instruction decoding and pasted in parts of the ARM kprobes code instead. End result is a large duplication of existing code segments which incidentally doesn't have the favor these days.
Nicolas