|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH v3 05/18] arm: poison initmem when it is freed.
Hi Konrad, On 11/09/2016 22:35, Konrad Rzeszutek Wilk wrote: Whilst I agree it is much better to get a trap rather than executing a random instruction (though co-processor 12 is reserved and a trap should (?) occur), I see two problems with this choice. Firstly, it might be possible that in the future we decide to use svc in the hypervisor (it is dumb but valid use case). Secondly, AArch64 is using a different set (and therefore encoding). So far this encoding is not allocated, but it is not rule out that this encoding will not be used in the future. So I would suggest to point initmem with:- AArch32: udf instruction i.e 0xe7f000f0 (see A8.8.247 in ARM DDI 0406C.c) - AArch64: a break point (possibly the encoding AARCH64_BREAK_FAULT (see asm-arm/arm64/brk.h). What do you think? Regardless the final decision, I think we should document the meaning of the value in the code. Regards, -- Julien Grall _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxx https://lists.xen.org/xen-devel
|
![]() |
Lists.xenproject.org is hosted with RackSpace, monitoring our |