On Tue, 2023-04-04 at 10:45 +0800, Binbin Wu wrote: > On 4/4/2023 9:53 AM, Huang, Kai wrote: > > On Tue, 2023-04-04 at 09:21 +0800, Binbin Wu wrote: > > > On 4/3/2023 7:24 PM, Huang, Kai wrote: > > > > > > > > > > Anyway, I will seperate this patch from the LAM KVM enabling patch. And > > > > > send a patch seperately if > > > > > needed later. > > > > > > > > > I think your change for SGX is still needed based on the pseudo code of ENCLS. > > > Yes, I meant I would seperate VMX part since it is not a bug after all, > > > SGX will still be in the patchset. > > > > > > > > Shouldn't SGX part be also split out as a bug fix patch? > > > > Does it have anything to do with this LAM support series? > > It is related to LAM support because LAM only effective in 64-bit mode, > so the untag action should only be done in 64-bit mode. > > If the SGX fix patch is not included, that means LAM untag could be > called in compatiblity mode in SGX ENCLS handler. > > Yes I got this point, and your patch 6/7 depends on it. But my point is this fix is needed anyway regardless the LAM support, and it should be merged, for instance, asap as a bug fix (and CC stable perhaps) -- while the LAM support is a feature, and can be merged at a different time frame. Of course just my 2cents and this is up to maintainers.
