Messages in this thread |  | | From | "Edgecombe, Rick P" <> | | Subject | Re: [PATCH v4 12/16] x86/virt/tdx: Add helpers to allow for pre-allocating pages | | Date | Wed, 26 Nov 2025 22:33:08 +0000 |
| |
On Wed, 2025-11-26 at 11:40 +0800, Binbin Wu wrote: > > On 11/21/2025 8:51 AM, Rick Edgecombe wrote: > > In the KVM fault path page, tables and private pages need to be > "In the KVM fault path page, tables ..." should be > "In the KVM fault path, page tables ..."
Right, thanks.
> > > installed under a spin lock. This means that the operations around > > installing PAMT pages for them will not be able to allocate pages. > > > [...] > > @@ -141,7 +142,46 @@ int tdx_guest_keyid_alloc(void); > > u32 tdx_get_nr_guest_keyids(void); > > void tdx_guest_keyid_free(unsigned int keyid); > > > > -int tdx_pamt_get(struct page *page); > > +int tdx_dpamt_entry_pages(void); > > + > > +/* > > + * Simple structure for pre-allocating Dynamic > > + * PAMT pages outside of locks. > > It's not just for Dynamic PAMT pages, but also external page table pages.
I'll adjust.
> > > + */ > > +struct tdx_prealloc { > > + struct list_head page_list; > > + int cnt; > > +}; > > + > > +static inline struct page *get_tdx_prealloc_page(struct tdx_prealloc *prealloc) > > +{ > > + struct page *page; > > + > > + page = list_first_entry_or_null(&prealloc->page_list, struct page, lru); > > + if (page) { > > + list_del(&page->lru); > > + prealloc->cnt--; > > + } > > + > > + return page; > > +} > > + > > +static inline int topup_tdx_prealloc_page(struct tdx_prealloc *prealloc, unsigned int min_size) > > +{ > > + while (prealloc->cnt < min_size) { > > + struct page *page = alloc_page(GFP_KERNEL_ACCOUNT); > > + > > + if (!page) > > + return -ENOMEM; > > + > > + list_add(&page->lru, &prealloc->page_list); > > + prealloc->cnt++; > > + } > > + > > + return 0; > > +} > > + > > +int tdx_pamt_get(struct page *page, struct tdx_prealloc *prealloc); > > void tdx_pamt_put(struct page *page); > > > > struct page *tdx_alloc_page(void); > > @@ -219,6 +259,7 @@ static inline int tdx_enable(void) { return -ENODEV; } > > static inline u32 tdx_get_nr_guest_keyids(void) { return 0; } > > static inline const char *tdx_dump_mce_info(struct mce *m) { return NULL; } > > static inline const struct tdx_sys_info *tdx_get_sysinfo(void) { return NULL; } > > +static inline int tdx_dpamt_entry_pages(void) { return 0; } > > #endif /* CONFIG_INTEL_TDX_HOST */ > > > > #ifdef CONFIG_KEXEC_CORE > > diff --git a/arch/x86/kvm/vmx/tdx.c b/arch/x86/kvm/vmx/tdx.c > > index 260bb0e6eb44..61a058a8f159 100644 > > --- a/arch/x86/kvm/vmx/tdx.c > > +++ b/arch/x86/kvm/vmx/tdx.c > > @@ -1644,23 +1644,34 @@ static int tdx_mem_page_add(struct kvm *kvm, gfn_t gfn, enum pg_level level, > > > > static void *tdx_alloc_external_fault_cache(struct kvm_vcpu *vcpu) > > { > > - struct vcpu_tdx *tdx = to_tdx(vcpu); > > + struct page *page = get_tdx_prealloc_page(&to_tdx(vcpu)->prealloc); > > > > - return kvm_mmu_memory_cache_alloc(&tdx->mmu_external_spt_cache); > > + if (WARN_ON_ONCE(!page)) > > + return (void *)__get_free_page(GFP_ATOMIC | __GFP_ACCOUNT); > > kvm_mmu_memory_cache_alloc() calls BUG_ON() if the atomic allocation failed. > Do we want to follow?
Ignoring, per your other comment. But we don't want to add BUG_ON()s in general. > > > + > > + return page_address(page); > > } > > > > static int tdx_topup_external_fault_cache(struct kvm_vcpu *vcpu, unsigned int cnt) > > { > > - struct vcpu_tdx *tdx = to_tdx(vcpu); > > + struct tdx_prealloc *prealloc = &to_tdx(vcpu)->prealloc; > > + int min_fault_cache_size; > > > > - return kvm_mmu_topup_memory_cache(&tdx->mmu_external_spt_cache, cnt); > > + /* External page tables */ > > + min_fault_cache_size = cnt; > > + /* Dynamic PAMT pages (if enabled) */ > > + min_fault_cache_size += tdx_dpamt_entry_pages() * PT64_ROOT_MAX_LEVEL; > > Is the value PT64_ROOT_MAX_LEVEL intended, since dynamic PAMT pages are only > needed for 4KB level?
I'm not sure I follow. We need DPAMT backing for each S-EPT page table.
|  |