On Sun, Jun 11, 2023 at 11:25:27PM -0500, Michael Roth wrote:
> From: Ashish Kalra <ashish.kalra@xxxxxxx>
>
> Pages are unsafe to be released back to the page-allocator, if they
> have been transitioned to firmware/guest state and can't be reclaimed
> or transitioned back to hypervisor/shared state. In this case add
> them to an internal leaked pages list to ensure that they are not freed
> or touched/accessed to cause fatal page faults.
>
> Signed-off-by: Ashish Kalra <ashish.kalra@xxxxxxx>
> [mdr: relocate to arch/x86/coco/sev/host.c]
> Signed-off-by: Michael Roth <michael.roth@xxxxxxx>
> ---
> arch/x86/coco/sev/host.c | 28 ++++++++++++++++++++++++++++
> arch/x86/include/asm/sev-host.h | 3 +++
> 2 files changed, 31 insertions(+)
>
> diff --git a/arch/x86/coco/sev/host.c b/arch/x86/coco/sev/host.c
> index cd3b4c6a25bc..373e91f5a337 100644
> --- a/arch/x86/coco/sev/host.c
> +++ b/arch/x86/coco/sev/host.c
> @@ -64,6 +64,12 @@ struct rmpentry {
> static unsigned long rmptable_start __ro_after_init;
> static unsigned long rmptable_end __ro_after_init;
>
> +/* list of pages which are leaked and cannot be reclaimed */
> +static LIST_HEAD(snp_leaked_pages_list);
> +static DEFINE_SPINLOCK(snp_leaked_pages_list_lock);
> +
> +static atomic_long_t snp_nr_leaked_pages = ATOMIC_LONG_INIT(0);
> +
> #undef pr_fmt
> #define pr_fmt(fmt) "SEV-SNP: " fmt
>
> @@ -494,3 +500,25 @@ int rmp_make_shared(u64 pfn, enum pg_level level)
> return rmpupdate(pfn, &val);
> }
> EXPORT_SYMBOL_GPL(rmp_make_shared);
> +
> +void snp_leak_pages(unsigned long pfn, unsigned int npages)
> +{
> + struct page *page = pfn_to_page(pfn);
> +
> + WARN(1, "psc failed, pfn 0x%lx pages %d (marked offline)\n", pfn, npages);
> +
> + spin_lock(&snp_leaked_pages_list_lock);
> + while (npages--) {
> + /*
> + * Reuse the page's buddy list for chaining into the leaked
> + * pages list. This page should not be on a free list currently
> + * and is also unsafe to be added to a free list.
> + */
> + list_add_tail(&page->buddy_list, &snp_leaked_pages_list);
> + sev_dump_rmpentry(pfn);
> + pfn++;
> + }
> + spin_unlock(&snp_leaked_pages_list_lock);
> + atomic_long_inc(&snp_nr_leaked_pages);
> +}
> +EXPORT_SYMBOL_GPL(snp_leak_pages);
> diff --git a/arch/x86/include/asm/sev-host.h b/arch/x86/include/asm/sev-host.h
> index 753e80d16433..bab3b226777a 100644
> --- a/arch/x86/include/asm/sev-host.h
> +++ b/arch/x86/include/asm/sev-host.h
> @@ -19,6 +19,8 @@ void sev_dump_rmpentry(u64 pfn);
> int psmash(u64 pfn);
> int rmp_make_private(u64 pfn, u64 gpa, enum pg_level level, int asid, bool immutable);
> int rmp_make_shared(u64 pfn, enum pg_level level);
> +void snp_leak_pages(unsigned long pfn, unsigned int npages);
> +
> #else
> static inline int snp_lookup_rmpentry(u64 pfn, bool *assigned, int *level) { return 0; }
> static inline void sev_dump_rmpentry(u64 pfn) {}
> @@ -29,6 +31,7 @@ static inline int rmp_make_private(u64 pfn, u64 gpa, enum pg_level level, int as
> return -ENODEV;
> }
> static inline int rmp_make_shared(u64 pfn, enum pg_level level) { return -ENODEV; }
> +void snp_leak_pages(unsigned long pfn, unsigned int npages) {}
This needs to be 'static inline' or the build fails with multiple definition errors.
I'm building a guest kernel with CONFIG_KVM_AMD_SEV disabled.
Jeremi
> #endif
>
> #endif
> --
> 2.25.1
>
