16.09.2010 15:32, Joerg Roedel пишет: > On Thu, Sep 16, 2010 at 07:11:17PM +0800, Alec Joseph Rivera wrote: >> On Thu, 2010-09-16 at 13:02 +0200, Joerg Roedel wrote: >> >> Will wait for a 2.6.36 kernel then or compile a .36-rc myself on free >> time :-) >> >> Thanks, will update the list when either comes. > > Just checked, 2.6.36-rc4 contains the fix for the cmpxchg8b problem on > 32 bit hosts. What commit it is? I swear I saw it in 2.6.32-stable somewhere. Is it this one (2.6.32.12): From: Gleb Natapov <gleb@xxxxxxxxxx> Date: Fri, 19 Mar 2010 15:47:31 +0100 Subject: KVM: x86 emulator: fix memory access during x86 emulation To: stable@xxxxxxxxxx Cc: Marcelo Tosatti <mtosatti@xxxxxxxxxx>, Avi Kivity <avi@xxxxxxxxxx>, Gleb Natapov <gleb@xxxxxxxxxx> Message-ID: <1269010059-25309-4-git-send-email-stefan.bader@xxxxxxxxxxxxx> From: Gleb Natapov <gleb@xxxxxxxxxx> commit 1871c6020d7308afb99127bba51f04548e7ca84e upstream Currently when x86 emulator needs to access memory, page walk is done with broadest permission possible, so if emulated instruction was executed by userspace process it can still access kernel memory. Fix that by providing correct memory access to page walker during emulation. /mjt -- 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
