Re: [PATCH v15 08/23] x86/virt/tdx: Use all system memory when initializing TDX module as TDX memory

From: Mike Rapoport
Date: Thu Dec 05 2024 - 11:30:38 EST


On Thu, Dec 05, 2024 at 11:06:17AM +0200, Nikolay Borisov wrote:
>
>
> On 5.12.24 г. 9:57 ч., Mike Rapoport wrote:
> > Hi,
> >
> > I've been auditing for_each_mem_pfn_range() users and it's usage in TDX is
> > dubious for me.
> >
> > On Fri, Nov 10, 2023 at 12:55:45AM +1300, Kai Huang wrote:
> > >
> > > As TDX-usable memory is a fixed configuration, take a snapshot of the
> > > memory configuration from memblocks at the time of module initialization
> > > (memblocks are modified on memory hotplug). This snapshot is used to
> >
> > AFAUI this could happen long after free_initmem() which discards all
> > memblock data on x86.
> > >> enable TDX support for *this* memory configuration only. Use a memory
> > > hotplug notifier to ensure that no other RAM can be added outside of
> > > this configuration.
> > ...
> >
> > > +/*
> > > + * Ensure that all memblock memory regions are convertible to TDX
> > > + * memory. Once this has been established, stash the memblock
> > > + * ranges off in a secondary structure because memblock is modified
> > > + * in memory hotplug while TDX memory regions are fixed.
> > > + */
> > > +static int build_tdx_memlist(struct list_head *tmb_list)
> > > +{
> > > + unsigned long start_pfn, end_pfn;
> > > + int i, ret;
> > > +
> > > + for_each_mem_pfn_range(i, MAX_NUMNODES, &start_pfn, &end_pfn, NULL) {
> >
> > Unles ARCH_KEEP_MEMBLOCK is defined this won't work after free_initmem()
>
> TDX_HOST actually selects ARCH_KEEP_MEMBLOCK:

Oh, I've missed that, thanks!

> 6 config INTEL_TDX_HOST
> 5 bool "Intel Trust Domain Extensions (TDX) host support"
> 4 depends on CPU_SUP_INTEL
> 3 depends on X86_64
> 2 depends on KVM_INTEL
> 1 depends on X86_X2APIC
> 1980 select ARCH_KEEP_MEMBLOCK
> 1 depends on CONTIG_ALLOC
> 2 depends on !KEXEC_CORE
> 3 depends on X86_MCE
>
>
> <snip>
>

--
Sincerely yours,
Mike.