From: Michael Kelley <mhklinux@xxxxxxxxxxx> [ Upstream commit 9fef276f9f416a1e85eb48d3bd38e6018a220bf5 ] In preparation for temporarily marking pages not present during a transition between encrypted and decrypted, use slow_virt_to_phys() in the hypervisor callback. As long as the PFN is correct, slow_virt_to_phys() works even if the leaf PTE is not present. The existing functions that depend on vmalloc_to_page() all require that the leaf PTE be marked present, so they don't work. Update the comments for slow_virt_to_phys() to note this broader usage and the requirement to work even if the PTE is not marked present. Signed-off-by: Michael Kelley <mhklinux@xxxxxxxxxxx> Acked-by: Kirill A. Shutemov <kirill.shutemov@xxxxxxxxxxxxxxx> Reviewed-by: Rick Edgecombe <rick.p.edgecombe@xxxxxxxxx> Link: https://lore.kernel.org/r/20240116022008.1023398-2-mhklinux@xxxxxxxxxxx Signed-off-by: Wei Liu <wei.liu@xxxxxxxxxx> Message-ID: <20240116022008.1023398-2-mhklinux@xxxxxxxxxxx> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx> --- arch/x86/hyperv/ivm.c | 12 +++++++++++- arch/x86/mm/pat/set_memory.c | 12 ++++++++---- 2 files changed, 19 insertions(+), 5 deletions(-) diff --git a/arch/x86/hyperv/ivm.c b/arch/x86/hyperv/ivm.c index 8c6bf07f7d2b8..4bf3805aa8ab5 100644 --- a/arch/x86/hyperv/ivm.c +++ b/arch/x86/hyperv/ivm.c @@ -515,6 +515,8 @@ static bool hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bo enum hv_mem_host_visibility visibility = enc ? VMBUS_PAGE_NOT_VISIBLE : VMBUS_PAGE_VISIBLE_READ_WRITE; u64 *pfn_array; + phys_addr_t paddr; + void *vaddr; int ret = 0; bool result = true; int i, pfn; @@ -524,7 +526,15 @@ static bool hv_vtom_set_host_visibility(unsigned long kbuffer, int pagecount, bo return false; for (i = 0, pfn = 0; i < pagecount; i++) { - pfn_array[pfn] = virt_to_hvpfn((void *)kbuffer + i * HV_HYP_PAGE_SIZE); + /* + * Use slow_virt_to_phys() because the PRESENT bit has been + * temporarily cleared in the PTEs. slow_virt_to_phys() works + * without the PRESENT bit while virt_to_hvpfn() or similar + * does not. + */ + vaddr = (void *)kbuffer + (i * HV_HYP_PAGE_SIZE); + paddr = slow_virt_to_phys(vaddr); + pfn_array[pfn] = paddr >> HV_HYP_PAGE_SHIFT; pfn++; if (pfn == HV_MAX_MODIFY_GPA_REP_COUNT || i == pagecount - 1) { diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c index bda9f129835e9..355dc8f5cb7dd 100644 --- a/arch/x86/mm/pat/set_memory.c +++ b/arch/x86/mm/pat/set_memory.c @@ -755,10 +755,14 @@ pmd_t *lookup_pmd_address(unsigned long address) * areas on 32-bit NUMA systems. The percpu areas can * end up in this kind of memory, for instance. * - * This could be optimized, but it is only intended to be - * used at initialization time, and keeping it - * unoptimized should increase the testing coverage for - * the more obscure platforms. + * Note that as long as the PTEs are well-formed with correct PFNs, this + * works without checking the PRESENT bit in the leaf PTE. This is unlike + * the similar vmalloc_to_page() and derivatives. Callers may depend on + * this behavior. + * + * This could be optimized, but it is only used in paths that are not perf + * sensitive, and keeping it unoptimized should increase the testing coverage + * for the more obscure platforms. */ phys_addr_t slow_virt_to_phys(void *__virt_addr) { -- 2.43.0
