On Mon, Apr 16, 2012 at 03:52:10PM +0900, Kuniyasu Suzaki wrote: > > Marcelo, > > From: Marcelo Tosatti <mtosatti@xxxxxxxxxx> > Subject: Re: EuroSec'12 Presentation (ASLR reduces effect of KSM) > Date: Fri, 13 Apr 2012 21:47:47 -0300 > > > On Thu, Apr 12, 2012 at 08:24:57PM +0900, Kuniyasu Suzaki wrote: > > > > > > Dear, > > > > > > I made a presentation which measures OS security functions(ASLR, > > > Memory Santization, and Cache Page Flushing) on memory deduplication > > > "KSM with VKM" at EuroSec 2012. > > > > > > The titile is "Effects of Memory Randomization, Sanitization and Page > > > Cache on Memory Deduplication". > > > # This is one of papers related to my memory deduplication research. > > > > > > The slide is downloadable. > > > http://www.slideshare.net/suzaki/eurosec2012-effects-of-memory-randomization-sanitization-and-page-cache-on-memory-deduplication-by-ksuzaki > > > The paper will be downloadable form ACM Digital Library. > > > > > > The results show ALSR reduces the effect of memory deduplciation. > > > Please tell me, if you have comments. Thank you. > > > > > > ------ > > > Kuniyasu Suzaki, National Institute of Advanced Industrial Science and Technology, > > > http://staff.aist.go.jp/k.suzaki > > > > Very nice. ALSR is supposed to increase the number of unshared pages > > because translation tables that contain addresses of symbols will > > differ for every instance of an executable. > > Thank you for good suggestion. > Anyway, how much the size of translation tables? One entry per symbol that is accessed outside of the object or main executable, one table per shared object (GOT and PLT tables). See the ELF documentation. > In our experience, ALSR on 4 GuestOS (Linux) increased the memory consumption more than 50MB. > Does the translation table in a linux take more than 10MB? Increased memory consumption is due to larger number of pagetables (which is necessary to cover larger virtual address space). Increased number of unshared pages can be explained by translation tables. > > Can you share additional information about "HICAMP (hardware memory > > deduplication)" ? > > The detail of HICAMP was presented at ASPLOS 2011. > ASPLOS 2011 paper http://dl.acm.org/citation.cfm?id=2151007&preflayout=tabs > > ------ > suzaki -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
