KVM: guest_memfd: Rework preparation/population flows in prep for in-place conversion
Subsystems (7)
v2 cover KVM: guest_memfd: Rework preparation/population flows in prep for in-place conversion
This patchset is also available at:
https://github.com/AMDESE/linux/tree/gmem-populate-rework-v2
and is based on top of kvm/next (e0c26d47def7)
Overview
--------
Yan previously posted a series[1] that reworked kvm_gmem_populate() to deal with potential locking issues that might arise once in-place conversion support[2] is added for guest_memfd. To quote Yan's original summary of the issues:
(1)
In Michael's series "KVM: gmem: 2MB THP support and preparedness tracking changes" [4], kvm_gmem_get_pfn() was modified to rely on the filemap invalidation lock for protecting its preparedness tracking. Similarly, the in-place conversion version of guest_memfd series by Ackerly also requires kvm_gmem_get_pfn() to acquire filemap invalidation lock [5].
kvm_gmem_get_pfn
filemap_invalidate_lock_shared(file_inode(file)->i_mapping); However, since kvm_gmem_get_pfn() is called by kvm_tdp_map_page(), which is in turn invoked within kvm_gmem_populate() in TDX, a deadlock occurs on the filemap invalidation lock.
(2)
Moreover, in step 2, get_user_pages_fast() may acquire mm->mmap_lock, resulting in the following lock sequence in tdx_vcpu_init_mem_region():
- filemap invalidation lock --> mm->mmap_lock
However, in future code, the shared filemap invalidation lock will be held in kvm_gmem_fault_shared() (see [6]), leading to the lock sequence:
- mm->mmap_lock --> filemap invalidation lock
This creates an AB-BA deadlock issue.
Sean has since then addressed (1) with his series[3] that avoids relying on calling kvm_gmem_get_pfn() within the TDX post-populate callback to re-fetch the PFN that was passed to it.
This series aims to address (2), which is still outstanding, and does so based heavily on Sean's suggested approach[4] of hoisting the get_user_pages_fast() out of the TDX post-populate callback so that it can be called prior to taking the filemap invalidate lock so that the ABBA deadlock is no longer possible. As preperation for this change, all the partial enablement for hugepages in the kvm_gmem_populate() path is stripped out so that it can be better considered once hugepage support is actually in place and code/design can be kept simpler in the meantime.
It additionally removes 'preparation' tracking from guest_memfd, which would similarly complicate locking considerations in the context of in-place conversion (and even moreso in the context of hugepage support). This has been discussed during both the guest_memfd calls and PUCK calls, and so far no strong objections have been given, so hopefully that particular change isn't too controversial.
Some items worth noting/discussing
----------------------------------
(A) While one of the aims of this rework is to implement things such that a separate source address can still be passed to kvm_gmem_populate() even though the gmem pages can be populated in-place from userspace beforehand, issues still arise if the source address itself has the KVM_MEMORY_ATTRIBUTE_PRIVATE attribute set, e.g. if source/target addresses are the same page. One line of reasoning would be to conclude that KVM_MEMORY_ATTRIBUTE_PRIVATE implies that it cannot be used as the source of a GUP/copy_from_user(), and thus cases like source==target are naturally disallowed. Thus userspace has no choice but to populate pages in-place *prior* to setting the KVM_MEMORY_ATTRIBUTE_PRIVATE attribute (as kvm_gmem_populate() requires), and passing in NULL for the source such that the GUP can be skipped (otherwise, it will trigger the shared memory fault path, which will then SIGBUS because it will see that it is faulting in pages for which KVM_MEMORY_ATTRIBUTE_PRIVATE is set).
While workable, this would at the very least involve documentation updates to KVM_TDX_INIT_MEM_REGION/KVM_SEV_SNP_LAUNCH_UPDATE to cover these soon-to-be-possible scenarios. Ira posted a patch separately that demonstrates how a NULL source could be safely handled within the TDX post-populate callback[5].
Known issues / TODO
-------------------
- Compile-tested only for the TDX bits (testing/feedback welcome!)
Changes since RFC v1
--------------------
- and a prep patch to remove partial hugepage enablement in kvm_gmem_populate() to simplify things until a hugepage implementation can actually make use of it (Yan, Ira, Vishal, Sean)
- begin retroactively enforcing that source pages must be page-aligned so that kvm_gmem_populate() callbacks can be simplified. add a patch to update SNP user-facing documentation to mention this.
- drop handling for GUP'ing multiple pages before issuing callbacks. This will only be needed for potentially for hugepages, and it must simpler to handle per-page in the meantime. (Yan, Vishal)
- make sure TDX actually builds (Ira, Yan)
Thanks,
Mike
[1] https://lore.kernel.org/kvm/[email protected]/
[2] https://lore.kernel.org/kvm/[email protected]/
[3] https://lore.kernel.org/kvm/[email protected]/
[4] https://lore.kernel.org/kvm/[email protected]/
[5] https://lore.kernel.org/kvm/[email protected]/
----------------------------------------------------------------
Michael Roth (5):
KVM: guest_memfd: Remove partial hugepage handling from kvm_gmem_populate()
KVM: guest_memfd: Remove preparation tracking
KVM: SEV: Document/enforce page-alignment for KVM_SEV_SNP_LAUNCH_UPDATE
KVM: TDX: Document alignment requirements for KVM_TDX_INIT_MEM_REGION
KVM: guest_memfd: GUP source pages prior to populating guest memory
.../virt/kvm/x86/amd-memory-encryption.rst | 2 +- Documentation/virt/kvm/x86/intel-tdx.rst | 2 +- arch/x86/kvm/svm/sev.c | 108 +++++++--------- arch/x86/kvm/vmx/tdx.c | 15 +-- include/linux/kvm_host.h | 4 +- virt/kvm/guest_memfd.c | 140 +++++++++++---------- 6 files changed, 129 insertions(+), 142 deletions(-)
PATCH v2 1/5 KVM: guest_memfd: Remove partial hugepage handling from kvm_gmem_populate()
kvm_gmem_populate(), and the associated post-populate callbacks, have some limited support for dealing with guests backed by hugepages by passing the order information along to each post-populate callback and iterating through the pages passed to kvm_gmem_populate() in hugepage-chunks.
However, guest_memfd doesn't yet support hugepages, and in most cases additional changes in the kvm_gmem_populate() path would also be needed to actually allow for this functionality.
This makes the existing code unecessarily complex, and makes changes difficult to work through upstream due to theoretical impacts on hugepage support that can't be considered properly without an actual hugepage implementation to reference. So for now, remove what's there so changes for things like in-place conversion can be implemented/reviewed more efficiently.
--- arch/x86/kvm/svm/sev.c | 94 ++++++++++++++++------------------------ arch/x86/kvm/vmx/tdx.c | 2 +- include/linux/kvm_host.h | 2 +- virt/kvm/guest_memfd.c | 30 +++++++------ 4 files changed, 56 insertions(+), 72 deletions(-)
arch/x86/kvm/svm/sev.c +37 -57
@@ -2267,66 +2267,52 @@ struct sev_gmem_populate_args {int fw_error; }; -static int sev_gmem_post_populate (struct kvm *kvm,gfn_t gfn_start,kvm_pfn_t pfn, -void __user *src,int order,void *opaque) +static int sev_gmem_post_populate (struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, +void __user *src,void *opaque) {struct sev_gmem_populate_args *sev_populate_args = opaque; +struct sev_data_snp_launch_update fw_args = {0 };struct kvm_sev_info *sev =to_kvm_sev_info (kvm); -int n_private =0 , ret, i; -int npages = (1 << order); -gfn_t gfn; +bool assigned = false; +int level; +int ret;if (WARN_ON_ONCE (sev_populate_args->type !=KVM_SEV_SNP_PAGE_TYPE_ZERO && !src))return -EINVAL ; -for (gfn = gfn_start, i =0 ; gfn < gfn_start + npages; gfn++, i++) { -struct sev_data_snp_launch_update fw_args = {0 }; -bool assigned = false; -int level; - - ret =snp_lookup_rmpentry ((u64 )pfn + i, &assigned, &level); -if (ret || assigned) { -pr_debug ("%s: Failed to ensure GFN 0x%llx RMP entry is initial shared state, ret: %d assigned: %d\n" , - __func__, gfn, ret, assigned); - ret = ret ? -EINVAL : -EEXIST ; - goto err; - } + ret =snp_lookup_rmpentry ((u64 )pfn, &assigned, &level); +if (ret || assigned) { +pr_debug ("%s: Failed to ensure GFN 0x%llx RMP entry is initial shared state, ret: %d assigned: %d\n" , + __func__, gfn, ret, assigned); + ret = ret ? -EINVAL : -EEXIST ; + goto out; + } -if (src) { -void *vaddr =kmap_local_pfn (pfn + i); +if (src) { +void *vaddr =kmap_local_pfn (pfn); -if (copy_from_user (vaddr, src + i *PAGE_SIZE ,PAGE_SIZE )) { - ret = -EFAULT ; - goto err; - } -kunmap_local (vaddr); +if (copy_from_user (vaddr, src,PAGE_SIZE )) { + ret = -EFAULT ; + goto out; } - - ret =rmp_make_private (pfn + i, gfn <<PAGE_SHIFT ,PG_LEVEL_4K , -sev_get_asid (kvm), true); -if (ret) - goto err; - - n_private++; - - fw_args.gctx_paddr =__psp_pa (sev->snp_context); - fw_args.address =__sme_set (pfn_to_hpa (pfn + i)); - fw_args.page_size =PG_LEVEL_TO_RMP (PG_LEVEL_4K ); - fw_args.page_type = sev_populate_args->type; - - ret =__sev_issue_cmd (sev_populate_args->sev_fd,SEV_CMD_SNP_LAUNCH_UPDATE , - &fw_args, &sev_populate_args->fw_error); -if (ret) - goto fw_err; +kunmap_local (vaddr); } -return 0 ; + ret =rmp_make_private (pfn, gfn <<PAGE_SHIFT ,PG_LEVEL_4K , +sev_get_asid (kvm), true); +if (ret) + goto out; + + fw_args.gctx_paddr =__psp_pa (sev->snp_context); + fw_args.address =__sme_set (pfn_to_hpa (pfn)); + fw_args.page_size =PG_LEVEL_TO_RMP (PG_LEVEL_4K ); + fw_args.page_type = sev_populate_args->type; -fw_err: + ret =__sev_issue_cmd (sev_populate_args->sev_fd,SEV_CMD_SNP_LAUNCH_UPDATE , + &fw_args, &sev_populate_args->fw_error);/* * If the firmware command failed handle the reclaim and cleanup of that -* PFN specially vs. prior pages which can be cleaned up below without -* needing to reclaim in advance. +* PFN before reporting an error. * * Additionally, when invalid CPUID function entries are detected, * firmware writes the expected values into the page and leaves it
@@ -2336,26 +2322,20 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn_start, kvm_pfn_t pf* information to provide information on which CPUID leaves/fields * failed CPUID validation. */ -if (!snp_page_reclaim (kvm, pfn + i) && +if (ret && !snp_page_reclaim (kvm, pfn) && sev_populate_args->type ==KVM_SEV_SNP_PAGE_TYPE_CPUID && sev_populate_args->fw_error ==SEV_RET_INVALID_PARAM ) { -void *vaddr =kmap_local_pfn (pfn + i); +void *vaddr =kmap_local_pfn (pfn); -if (copy_to_user (src + i *PAGE_SIZE , vaddr,PAGE_SIZE )) +if (copy_to_user (src, vaddr,PAGE_SIZE ))pr_debug ("Failed to write CPUID page back to userspace\n" );kunmap_local (vaddr); } -/* pfn + i is hypervisor-owned now, so skip below cleanup for it. */ - n_private--; - -err: -pr_debug ("%s: exiting with error ret %d (fw_error %d), restoring %d gmem PFNs to shared.\n" , - __func__, ret, sev_populate_args->fw_error, n_private); -for (i =0 ; i < n_private; i++) -kvm_rmp_make_shared (kvm, pfn + i,PG_LEVEL_4K ); - +out: +pr_debug ("%s: exiting with return code %d (fw_error %d)\n" , + __func__, ret, sev_populate_args->fw_error);return ret; }
arch/x86/kvm/vmx/tdx.c +1 -1
@@ -3118,7 +3118,7 @@ struct tdx_gmem_post_populate_arg { };static int tdx_gmem_post_populate (struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, -void __user *src,int order,void *_arg) +void __user *src,void *_arg) {struct tdx_gmem_post_populate_arg *arg = _arg;struct kvm_tdx *kvm_tdx =to_kvm_tdx (kvm);
include/linux/kvm_host.h +1 -1
@@ -2581,7 +2581,7 @@ int kvm_arch_gmem_prepare(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, int max_ord* Returns the number of pages that were populated. */ typedef int (*kvm_gmem_populate_cb )(struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, -void __user *src,int order,void *opaque); +void __user *src,void *opaque);long kvm_gmem_populate (struct kvm *kvm,gfn_t gfn,void __user *src,long npages,kvm_gmem_populate_cb post_populate,void *opaque);
virt/kvm/guest_memfd.c +17 -13
@@ -151,6 +151,15 @@ static struct folio *kvm_gmem_get_folio(struct inode *inode, pgoff_t index) mapping_gfp_mask(inode->i_mapping), policy);mpol_cond_put (policy); +/* +* External interfaces like kvm_gmem_get_pfn() support dealing +* with hugepages to a degree, but internally, guest_memfd currently +* assumes that all folios are order-0 and handling would need +* to be updated for anything otherwise (e.g. page-clearing +* operations). +*/ +WARN_ON_ONCE (folio_order (folio)); +return folio; }
@@ -829,7 +838,7 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longstruct kvm_memory_slot *slot;void __user *p; -int ret =0 , max_order; +int ret =0 ;long i;lockdep_assert_held (&kvm->slots_lock);
@@ -848,7 +857,7 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longfilemap_invalidate_lock (file->f_mapping); npages =min_t (ulong, slot->npages - (start_gfn - slot->base_gfn), npages); -for (i =0 ; i < npages; i += (1 << max_order)) { +for (i =0 ; i < npages; i++) {struct folio *folio;gfn_t gfn = start_gfn + i;pgoff_t index =kvm_gmem_get_index (slot, gfn);
@@ -860,7 +869,7 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longbreak ; } - folio =__kvm_gmem_get_pfn (file, slot, index, &pfn, &is_prepared, &max_order); + folio =__kvm_gmem_get_pfn (file, slot, index, &pfn, &is_prepared,NULL );if (IS_ERR (folio)) { ret =PTR_ERR (folio);break ;
@@ -874,20 +883,15 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long }folio_unlock (folio); -WARN_ON (!IS_ALIGNED (gfn,1 << max_order) || - (npages - i) < (1 << max_order)); ret = -EINVAL ; -while (!kvm_range_has_memory_attributes (kvm, gfn, gfn + (1 << max_order), -KVM_MEMORY_ATTRIBUTE_PRIVATE , -KVM_MEMORY_ATTRIBUTE_PRIVATE )) { -if (!max_order) - goto put_folio_and_exit; - max_order--; - } +if (!kvm_range_has_memory_attributes (kvm, gfn, gfn +1 , +KVM_MEMORY_ATTRIBUTE_PRIVATE , +KVM_MEMORY_ATTRIBUTE_PRIVATE )) + goto put_folio_and_exit; p = src ? src + i *PAGE_SIZE :NULL ; - ret =post_populate (kvm, gfn, pfn, p, max_order, opaque); + ret =post_populate (kvm, gfn, pfn, p, opaque);if (!ret)kvm_gmem_mark_prepared (folio);
Signature · 1 line
2.25.1
Re: [PATCH v2 1/5] KVM: guest_memfd: Remove partial hugepage handling from kvm_gmem_populate()
Quoted 18 lines · Michael Roth
kvm_gmem_populate(), and the associated post-populate callbacks, have some limited support for dealing with guests backed by hugepages by passing the order information along to each post-populate callback and iterating through the pages passed tokvm_gmem_populate()in hugepage-chunks.However,
guest_memfddoesn't yet support hugepages, and in most cases additional changes in thekvm_gmem_populate()path would also be needed to actually allow for this functionality.This makes the existing code unecessarily complex, and makes changes difficult to work through upstream due to theoretical impacts on hugepage support that can't be considered properly without an actual hugepage implementation to reference. So for now, remove what's there so changes for things like in-place conversion can be implemented/reviewed more efficiently.
Quoted 4 lines
Suggested-by Vishal Annapurve <[email protected]>Co-developed-by Vishal Annapurve <[email protected]>Signed-off-by Vishal Annapurve <[email protected]>Signed-off-by Michael Roth <[email protected]>
Quoted 6 lines
virt/kvm/guest_memfd.c +0
-0@@ -151,6 +151,15 @@ static struct folio *kvm_gmem_get_folio(struct inode *inode, pgoff_t index) mapping_gfp_mask(inode->i_mapping), policy);
virt/kvm/guest_memfd.c static struct folio *kvm_gmem_get_folio(struct inode *inode, pgoff_t index) · line 161
Quoted 10 lines
virt/kvm/guest_memfd.c +8
-0mpol_cond_put (policy); +/* +* External interfaces like kvm_gmem_get_pfn() support dealing +* with hugepages to a degree, but internally, guest_memfd currently +* assumes that all folios are order-0 and handling would need +* to be updated for anything otherwise (e.g. page-clearing +* operations). +*/ +WARN_ON_ONCE (folio_order (folio));
I am not sure if this WARN_ON adds any value. i.e. The current code can't hit it. This note concerns future efforts to add hugepage support and could be omitted altogether from the current implementation.
Quoted 4 lines
virt/kvm/guest_memfd.c +1
-0+return folio; }
Sean Christopherson
maintainerRe: [PATCH v2 1/5] KVM: guest_memfd: Remove partial hugepage handling from kvm_gmem_populate()
Quoted 9 lines · Vishal Annapurve
Quoted 8 lines · Michael Roth
virt/kvm/guest_memfd.c +0
-0@@ -151,6 +151,15 @@ static struct folio *kvm_gmem_get_folio(struct inode *inode, pgoff_t index) mapping_gfp_mask(inode->i_mapping), policy);mpol_cond_put (policy);
Quoted 11 lines
virt/kvm/guest_memfd.c static struct folio *kvm_gmem_get_folio(struct inode *inode, pgoff_t index) · line 161
Quoted 8 lines
virt/kvm/guest_memfd.c +8
-0+/* +* External interfaces like kvm_gmem_get_pfn() support dealing +* with hugepages to a degree, but internally, guest_memfd currently +* assumes that all folios are order-0 and handling would need +* to be updated for anything otherwise (e.g. page-clearing +* operations). +*/ +WARN_ON_ONCE (folio_order (folio));I am not sure if this
WARN_ONadds any value. i.e. The current code can't hit it.
The current code _shouldn't_ hit it.
Quoted 2 lines
This note concerns future efforts to add hugepage support and could be omitted altogether from the current implementation.
IMO, this is a good use of WARN_ON_ONCE(). It documents guest_memfd's assumptions and/or limitations, which is extremely helpful to readers/contributors that aren't familiar with guest_memfd and/or its history of hugepage support.
Re: [PATCH v2 1/5] KVM: guest_memfd: Remove partial hugepage handling from kvm_gmem_populate()
In reply to Michael RothQuoted 36 lines
(quoted diff) +12
-17-static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn_start, kvm_pfn_t pfn, - void __user *src, int order, void *opaque) +static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, + void __user *src, void *opaque) { struct sev_gmem_populate_args *sev_populate_args = opaque; + struct sev_data_snp_launch_update fw_args = {0}; struct kvm_sev_info *sev = to_kvm_sev_info(kvm); - int n_private = 0, ret, i; - int npages = (1 << order); - gfn_t gfn; + bool assigned = false; + int level; + int ret; if (WARN_ON_ONCE(sev_populate_args->type != KVM_SEV_SNP_PAGE_TYPE_ZERO && !src)) return -EINVAL; - for (gfn = gfn_start, i = 0; gfn < gfn_start + npages; gfn++, i++) { - struct sev_data_snp_launch_update fw_args = {0}; - bool assigned = false; - int level; - - ret = snp_lookup_rmpentry((u64)pfn + i, &assigned, &level); - if (ret || assigned) { - pr_debug("%s: Failed to ensure GFN 0x%llx RMP entry is initial shared state, ret: %d assigned: %d\n", - __func__, gfn, ret, assigned); - ret = ret ? -EINVAL : -EEXIST; - goto err; - } + ret = snp_lookup_rmpentry((u64)pfn, &assigned, &level); + if (ret || assigned) { + pr_debug("%s: Failed to ensure GFN 0x%llx RMP entry is initial shared state, ret: %d assigned: %d\n", + __func__, gfn, ret, assigned); + ret = ret ? -EINVAL : -EEXIST; + goto out;
Review
Quoted 10 lines
(quoted diff) +3
-5+ } - if (src) { - void *vaddr = kmap_local_pfn(pfn + i); + if (src) { + void *vaddr = kmap_local_pfn(pfn); - if (copy_from_user(vaddr, src + i * PAGE_SIZE, PAGE_SIZE)) { - ret = -EFAULT; - goto err;
Please consider apply the following fix before this patch. Thanks!
commit 2714522d42263e0e250f21a0b171c10c4bb17ed3
Author: Yan Zhao <[email protected]> Date: Mon Nov 10 11:22:28 2025 +0800
KVM: SVM: Fix a missing kunmap_local() in sev_gmem_post_populate()
sev_gmem_post_populate() needs to unmap the target vaddr after copy_from_user() to the vaddr fails.
arch/x86/kvm/svm/sev.c +1 -0
@@ -2296,6 +2296,7 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn_start, kvm_pfn_t pfvoid *vaddr =kmap_local_pfn (pfn + i);if (copy_from_user (vaddr, src + i *PAGE_SIZE ,PAGE_SIZE )) { +kunmap_local (vaddr); ret = -EFAULT ; goto err; }
Quoted 80 lines
(quoted diff) +22
-35- } - kunmap_local(vaddr); + if (copy_from_user(vaddr, src, PAGE_SIZE)) { + ret = -EFAULT; + goto out; } - - ret = rmp_make_private(pfn + i, gfn << PAGE_SHIFT, PG_LEVEL_4K, - sev_get_asid(kvm), true); - if (ret) - goto err; - - n_private++; - - fw_args.gctx_paddr = __psp_pa(sev->snp_context); - fw_args.address = __sme_set(pfn_to_hpa(pfn + i)); - fw_args.page_size = PG_LEVEL_TO_RMP(PG_LEVEL_4K); - fw_args.page_type = sev_populate_args->type; - - ret = __sev_issue_cmd(sev_populate_args->sev_fd, SEV_CMD_SNP_LAUNCH_UPDATE, - &fw_args, &sev_populate_args->fw_error); - if (ret) - goto fw_err; + kunmap_local(vaddr); } - return 0; + ret = rmp_make_private(pfn, gfn << PAGE_SHIFT, PG_LEVEL_4K, + sev_get_asid(kvm), true); + if (ret) + goto out; + + fw_args.gctx_paddr = __psp_pa(sev->snp_context); + fw_args.address = __sme_set(pfn_to_hpa(pfn)); + fw_args.page_size = PG_LEVEL_TO_RMP(PG_LEVEL_4K); + fw_args.page_type = sev_populate_args->type; -fw_err: + ret = __sev_issue_cmd(sev_populate_args->sev_fd, SEV_CMD_SNP_LAUNCH_UPDATE, + &fw_args, &sev_populate_args->fw_error); /* * If the firmware command failed handle the reclaim and cleanup of that - * PFN specially vs. prior pages which can be cleaned up below without - * needing to reclaim in advance. + * PFN before reporting an error. * * Additionally, when invalid CPUID function entries are detected, * firmware writes the expected values into the page and leaves it@@ -2336,26 +2322,20 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn_start, kvm_pfn_t pf * information to provide information on which CPUID leaves/fields * failed CPUID validation. */ - if (!snp_page_reclaim(kvm, pfn + i) && + if (ret && !snp_page_reclaim(kvm, pfn) && sev_populate_args->type == KVM_SEV_SNP_PAGE_TYPE_CPUID && sev_populate_args->fw_error == SEV_RET_INVALID_PARAM) { - void *vaddr = kmap_local_pfn(pfn + i); + void *vaddr = kmap_local_pfn(pfn); - if (copy_to_user(src + i * PAGE_SIZE, vaddr, PAGE_SIZE)) + if (copy_to_user(src, vaddr, PAGE_SIZE)) pr_debug("Failed to write CPUID page back to userspace\n"); kunmap_local(vaddr); } - /* pfn + i is hypervisor-owned now, so skip below cleanup for it. */ - n_private--; - -err: - pr_debug("%s: exiting with error ret %d (fw_error %d), restoring %d gmem PFNs to shared.\n", - __func__, ret, sev_populate_args->fw_error, n_private); - for (i = 0; i < n_private; i++) - kvm_rmp_make_shared(kvm, pfn + i, PG_LEVEL_4K); - +out: + pr_debug("%s: exiting with return code %d (fw_error %d)\n", + __func__, ret, sev_populate_args->fw_error); return ret; }
PATCH v2 2/5 KVM: guest_memfd: Remove preparation tracking
In reply to Michael Rothguest_memfd currently uses the folio uptodate flag to track:
1) whether or not a page has been cleared before initial usage
2) whether or not the architecture hooks have been issued to put the page in a private state as defined by the architecture
In practice, 2) is only actually being tracked for SEV-SNP VMs, and there do not seem to be any plans/reasons that would suggest this will change in the future, so this additional tracking/complexity is not really providing any general benefit to guest_memfd users. Future plans around in-place conversion and hugepage support, where the per-folio uptodate flag is planned to be used purely to track the initial clearing of folios, whereas conversion operations could trigger multiple transitions between 'prepared' and 'unprepared' and thus need separate tracking, will make the burden of tracking this information within guest_memfd even more complex, since preparation generally happens during fault time, on the "read-side" of any global locks that might protect state tracked by guest_memfd, and so may require more complex locking schemes to allow for concurrent handling of page faults for multiple vCPUs where the "preparedness" state tracked by guest_memfd might need to be updated as part of handling the fault.
Instead of keeping this current/future complexity within guest_memfd for what is essentially just SEV-SNP, just drop the tracking for 2) and have the arch-specific preparation hooks get triggered unconditionally on every fault so the arch-specific hooks can check the preparation state directly and decide whether or not a folio still needs additional preparation. In the case of SEV-SNP, the preparation state is already checked again via the preparation hooks to avoid double-preparation, so nothing extra needs to be done to update the handling of things there.
--- virt/kvm/guest_memfd.c | 44 ++++++++++++------------------------------ 1 file changed, 12 insertions(+), 32 deletions(-)
virt/kvm/guest_memfd.c +12 -32
@@ -76,11 +76,6 @@ static int __kvm_gmem_prepare_folio(struct kvm *kvm, struct kvm_memory_slot *sloreturn 0 ; } -static inline void kvm_gmem_mark_prepared (struct folio *folio) -{ -folio_mark_uptodate (folio); -} -/* * Process @folio, which contains @gfn, so that the guest can use it. * The folio must be locked and the gfn must be contained in @slot.
@@ -90,13 +85,7 @@ static inline void kvm_gmem_mark_prepared(struct folio *folio)static int kvm_gmem_prepare_folio (struct kvm *kvm,struct kvm_memory_slot *slot,gfn_t gfn,struct folio *folio) { -unsigned long nr_pages, i;pgoff_t index; -int r; - - nr_pages =folio_nr_pages (folio); -for (i =0 ; i < nr_pages; i++) -clear_highpage (folio_page (folio, i));/* * Preparing huge folios should always be safe, since it should
@@ -114,11 +103,8 @@ static int kvm_gmem_prepare_folio(struct kvm *kvm, struct kvm_memory_slot *slot,WARN_ON (!IS_ALIGNED (slot->gmem.pgoff,folio_nr_pages (folio))); index =kvm_gmem_get_index (slot, gfn); index =ALIGN_DOWN (index,folio_nr_pages (folio)); - r =__kvm_gmem_prepare_folio (kvm, slot, index, folio); -if (!r) -kvm_gmem_mark_prepared (folio); -return r; +return __kvm_gmem_prepare_folio (kvm, slot, index, folio); }/*
@@ -429,7 +415,7 @@ static vm_fault_t kvm_gmem_fault_user_mapping(struct vm_fault *vmf)if (!folio_test_uptodate (folio)) {clear_highpage (folio_page (folio,0 )); -kvm_gmem_mark_prepared (folio); +folio_mark_uptodate (folio); } vmf->page =folio_file_page (folio, vmf->pgoff);
@@ -766,7 +752,7 @@ void kvm_gmem_unbind(struct kvm_memory_slot *slot)static struct folio *__kvm_gmem_get_pfn (struct file *file,struct kvm_memory_slot *slot,pgoff_t index,kvm_pfn_t *pfn, -bool *is_prepared,int *max_order) +int *max_order) {struct file *slot_file =READ_ONCE (slot->gmem.file);struct gmem_file *f = file->private_data;
@@ -796,7 +782,6 @@ static struct folio *__kvm_gmem_get_pfn(struct file *file,if (max_order) *max_order =0 ; - *is_prepared =folio_test_uptodate (folio);return folio; }
@@ -806,19 +791,22 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot, {pgoff_t index =kvm_gmem_get_index (slot, gfn);struct folio *folio; -bool is_prepared = false;int r =0 ;CLASS (gmem_get_file, file)(slot);if (!file)return -EFAULT ; - folio =__kvm_gmem_get_pfn (file, slot, index, pfn, &is_prepared, max_order); + folio =__kvm_gmem_get_pfn (file, slot, index, pfn, max_order);if (IS_ERR (folio))return PTR_ERR (folio); -if (!is_prepared) - r =kvm_gmem_prepare_folio (kvm, slot, gfn, folio); +if (!folio_test_uptodate (folio)) { +clear_highpage (folio_page (folio,0 )); +folio_mark_uptodate (folio); + } + + r =kvm_gmem_prepare_folio (kvm, slot, gfn, folio);folio_unlock (folio);
@@ -861,7 +849,6 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longstruct folio *folio;gfn_t gfn = start_gfn + i;pgoff_t index =kvm_gmem_get_index (slot, gfn); -bool is_prepared = false;kvm_pfn_t pfn;if (signal_pending (current)) {
@@ -869,19 +856,12 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longbreak ; } - folio =__kvm_gmem_get_pfn (file, slot, index, &pfn, &is_prepared,NULL ); + folio =__kvm_gmem_get_pfn (file, slot, index, &pfn,NULL );if (IS_ERR (folio)) { ret =PTR_ERR (folio);break ; } -if (is_prepared) { -folio_unlock (folio); -folio_put (folio); - ret = -EEXIST ; -break ; - } -folio_unlock (folio); ret = -EINVAL ;
@@ -893,7 +873,7 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long p = src ? src + i *PAGE_SIZE :NULL ; ret =post_populate (kvm, gfn, pfn, p, opaque);if (!ret) -kvm_gmem_mark_prepared (folio); +folio_mark_uptodate (folio); put_folio_and_exit:folio_put (folio);
Signature · 1 line
2.25.1
Re: [PATCH v2 2/5] KVM: guest_memfd: Remove preparation tracking
Quoted 32 lines · Michael Roth
guest_memfdcurrently uses the folio uptodate flag to track:1) whether or not a page has been cleared before initial usage
2) whether or not the architecture hooks have been issued to put the page in a private state as defined by the architectureIn practice, 2) is only actually being tracked for SEV-SNP VMs, and there do not seem to be any plans/reasons that would suggest this will change in the future, so this additional tracking/complexity is not really providing any general benefit to
guest_memfdusers. Future plans around in-place conversion and hugepage support, where the per-folio uptodate flag is planned to be used purely to track the initial clearing of folios, whereas conversion operations could trigger multiple transitions between 'prepared' and 'unprepared' and thus need separate tracking, will make the burden of tracking this information withinguest_memfdeven more complex, since preparation generally happens during fault time, on the "read-side" of any global locks that might protect state tracked byguest_memfd, and so may require more complex locking schemes to allow for concurrent handling of page faults for multiple vCPUs where the "preparedness" state tracked byguest_memfdmight need to be updated as part of handling the fault.Instead of keeping this current/future complexity within
guest_memfdfor what is essentially just SEV-SNP, just drop the tracking for 2) and have the arch-specific preparation hooks get triggered unconditionally on every fault so the arch-specific hooks can check the preparation state directly and decide whether or not a folio still needs additional preparation. In the case of SEV-SNP, the preparation state is already checked again via the preparation hooks to avoid double-preparation, so nothing extra needs to be done to update the handling of things there.
Quoted 1 line
Signed-off-by Michael Roth <[email protected]>
Re: [PATCH v2 2/5] KVM: guest_memfd: Remove preparation tracking
In reply to Michael RothQuoted 32 lines
guest_memfdcurrently uses the folio uptodate flag to track:1) whether or not a page has been cleared before initial usage
2) whether or not the architecture hooks have been issued to put the page in a private state as defined by the architectureIn practice, 2) is only actually being tracked for SEV-SNP VMs, and there do not seem to be any plans/reasons that would suggest this will change in the future, so this additional tracking/complexity is not really providing any general benefit to
guest_memfdusers. Future plans around in-place conversion and hugepage support, where the per-folio uptodate flag is planned to be used purely to track the initial clearing of folios, whereas conversion operations could trigger multiple transitions between 'prepared' and 'unprepared' and thus need separate tracking, will make the burden of tracking this information withinguest_memfdeven more complex, since preparation generally happens during fault time, on the "read-side" of any global locks that might protect state tracked byguest_memfd, and so may require more complex locking schemes to allow for concurrent handling of page faults for multiple vCPUs where the "preparedness" state tracked byguest_memfdmight need to be updated as part of handling the fault.Instead of keeping this current/future complexity within
guest_memfdfor what is essentially just SEV-SNP, just drop the tracking for 2) and have the arch-specific preparation hooks get triggered unconditionally on every fault so the arch-specific hooks can check the preparation state directly and decide whether or not a folio still needs additional preparation. In the case of SEV-SNP, the preparation state is already checked again via the preparation hooks to avoid double-preparation, so nothing extra needs to be done to update the handling of things there.
Quoted 1 line
Signed-off-by Michael Roth <[email protected]>
Quoted 138 lines
--- virt/kvm/guest_memfd.c | 44 ++++++++++++------------------------------ 1 file changed, 12 insertions(+), 32 deletions(-)virt/kvm/guest_memfd.c +12
-32@@ -76,11 +76,6 @@ static int __kvm_gmem_prepare_folio(struct kvm *kvm, struct kvm_memory_slot *sloreturn 0 ; } -static inline void kvm_gmem_mark_prepared (struct folio *folio) -{ -folio_mark_uptodate (folio); -} -/* * Process @folio, which contains @gfn, so that the guest can use it. * The folio must be locked and the gfn must be contained in @slot. @@ -90,13 +85,7 @@ static inline void kvm_gmem_mark_prepared(struct folio *folio)static int kvm_gmem_prepare_folio (struct kvm *kvm,struct kvm_memory_slot *slot,gfn_t gfn,struct folio *folio) { -unsigned long nr_pages, i;pgoff_t index; -int r; - - nr_pages =folio_nr_pages (folio); -for (i =0 ; i < nr_pages; i++) -clear_highpage (folio_page (folio, i));/* * Preparing huge folios should always be safe, since it should @@ -114,11 +103,8 @@ static int kvm_gmem_prepare_folio(struct kvm *kvm, struct kvm_memory_slot *slot,WARN_ON (!IS_ALIGNED (slot->gmem.pgoff,folio_nr_pages (folio))); index =kvm_gmem_get_index (slot, gfn); index =ALIGN_DOWN (index,folio_nr_pages (folio)); - r =__kvm_gmem_prepare_folio (kvm, slot, index, folio); -if (!r) -kvm_gmem_mark_prepared (folio); -return r; +return __kvm_gmem_prepare_folio (kvm, slot, index, folio); }/* @@ -429,7 +415,7 @@ static vm_fault_t kvm_gmem_fault_user_mapping(struct vm_fault *vmf)if (!folio_test_uptodate (folio)) {clear_highpage (folio_page (folio,0 )); -kvm_gmem_mark_prepared (folio); +folio_mark_uptodate (folio); } vmf->page =folio_file_page (folio, vmf->pgoff);@@ -766,7 +752,7 @@ void kvm_gmem_unbind(struct kvm_memory_slot *slot)static struct folio *__kvm_gmem_get_pfn (struct file *file,struct kvm_memory_slot *slot,pgoff_t index,kvm_pfn_t *pfn, -bool *is_prepared,int *max_order) +int *max_order) {struct file *slot_file =READ_ONCE (slot->gmem.file);struct gmem_file *f = file->private_data;@@ -796,7 +782,6 @@ static struct folio *__kvm_gmem_get_pfn(struct file *file,if (max_order) *max_order =0 ; - *is_prepared =folio_test_uptodate (folio);return folio; }@@ -806,19 +791,22 @@ int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot, {pgoff_t index =kvm_gmem_get_index (slot, gfn);struct folio *folio; -bool is_prepared = false;int r =0 ;CLASS (gmem_get_file, file)(slot);if (!file)return -EFAULT ; - folio =__kvm_gmem_get_pfn (file, slot, index, pfn, &is_prepared, max_order); + folio =__kvm_gmem_get_pfn (file, slot, index, pfn, max_order);if (IS_ERR (folio))return PTR_ERR (folio); -if (!is_prepared) - r =kvm_gmem_prepare_folio (kvm, slot, gfn, folio); +if (!folio_test_uptodate (folio)) { +clear_highpage (folio_page (folio,0 )); +folio_mark_uptodate (folio); + } + + r =kvm_gmem_prepare_folio (kvm, slot, gfn, folio);folio_unlock (folio);@@ -861,7 +849,6 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longstruct folio *folio;gfn_t gfn = start_gfn + i;pgoff_t index =kvm_gmem_get_index (slot, gfn); -bool is_prepared = false;kvm_pfn_t pfn;if (signal_pending (current)) {@@ -869,19 +856,12 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longbreak ; } - folio =__kvm_gmem_get_pfn (file, slot, index, &pfn, &is_prepared,NULL ); + folio =__kvm_gmem_get_pfn (file, slot, index, &pfn,NULL );if (IS_ERR (folio)) { ret =PTR_ERR (folio);break ; } -if (is_prepared) { -folio_unlock (folio); -folio_put (folio); - ret = -EEXIST ; -break ; - } -folio_unlock (folio); ret = -EINVAL ;@@ -893,7 +873,7 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long p = src ? src + i *PAGE_SIZE :NULL ; ret =post_populate (kvm, gfn, pfn, p, opaque);if (!ret) -kvm_gmem_mark_prepared (folio); +folio_mark_uptodate (folio); put_folio_and_exit:folio_put (folio);
Re: [PATCH v2 2/5] KVM: guest_memfd: Remove preparation tracking
In reply to Michael RothQuoted 6 lines
/*
* Process @folio, which contains @gfn, so that the guest can use it.
* The folio must be locked and the gfn must be contained in @slot.virt/kvm/guest_memfd.c +0
-0@@ -90,13 +85,7 @@ static inline void kvm_gmem_mark_prepared(struct folio *folio) static int kvm_gmem_prepare_folio(struct kvm *kvm, struct kvm_memory_slot *slot,
virt/kvm/guest_memfd.c static inline void kvm_gmem_mark_prepared(struct folio *folio) · line 89
Quoted 10 lines
virt/kvm/guest_memfd.c +0
-6gfn_t gfn, struct folio *folio) { - unsigned long nr_pages, i; pgoff_t index; -int r; - - nr_pages =folio_nr_pages (folio); -for (i =0 ; i < nr_pages; i++) -clear_highpage (folio_page (folio, i));
Here the entire folio is cleared, but ...
[...]
virt/kvm/guest_memfd.c int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot, · line 807
Quoted 11 lines
virt/kvm/guest_memfd.c +1
-1- folio =__kvm_gmem_get_pfn (file, slot, index, pfn, &is_prepared, max_order); + folio =__kvm_gmem_get_pfn (file, slot, index, pfn, max_order);if (IS_ERR (folio))virt/kvm/guest_memfd.c +4
-2return PTR_ERR (folio); -if (!is_prepared) - r =kvm_gmem_prepare_folio (kvm, slot, gfn, folio); +if (!folio_test_uptodate (folio)) { +clear_highpage (folio_page (folio,0 )); +folio_mark_uptodate (folio); + }
... here only the first page is cleared.
I understand currently there's no huge folio coming out of gmem now, but since both __kvm_gmem_get_pfn() and kvm_gmem_get_pfn() still have @max_level as output, it's kinda inconsistent here.
I also see kvm_gmem_fault_user_mapping() only clears the first page too, but I think that already has assumption that folio can never be huge currently?
Given this, and the fact that the first patch of this series has introduced
WARN_ON_ONCE(folio_order(folio));
in kvm_gmem_get_folio(), I think it's fine to only clear the first page, but for the sake of consistency, perhaps we should just remove @max_order from __kvm_gmem_get_pfn() and kvm_gmem_get_pfn()?
Then we can handle huge folio logic when that comes to play.
Btw:
I actually looked into the RFC v1 discussion but the code there actually does a loop to clear all pages in the folio. There were some other discussions about AFAICT they were more related to issues regarding to "mark entire folio as uptodate while only one page is processed in post_populate()".
Btw2:
There was also discussion that clearing page isn't required for TDX. To that end, maybe we can remove clearing page from gmem common code but to SEV code, e.g., as part of "folio preparation"?
Re: [PATCH v2 2/5] KVM: guest_memfd: Remove preparation tracking
Quoted 46 lines · Kai
Quoted 6 lines
/*
* Process @folio, which contains @gfn, so that the guest can use it.
* The folio must be locked and the gfn must be contained in @slot.virt/kvm/guest_memfd.c +0
-0@@ -90,13 +85,7 @@ static inline void kvm_gmem_mark_prepared(struct folio *folio) static int kvm_gmem_prepare_folio(struct kvm *kvm, struct kvm_memory_slot *slot,virt/kvm/guest_memfd.c static inline void kvm_gmem_mark_prepared(struct folio *folio) · line 89
Quoted 10 lines
virt/kvm/guest_memfd.c +0
-6gfn_t gfn, struct folio *folio) { - unsigned long nr_pages, i; pgoff_t index; -int r; - - nr_pages =folio_nr_pages (folio); -for (i =0 ; i < nr_pages; i++) -clear_highpage (folio_page (folio, i));Here the entire folio is cleared, but ...
[...]
virt/kvm/guest_memfd.c int kvm_gmem_get_pfn(struct kvm *kvm, struct kvm_memory_slot *slot, · line 807
Quoted 11 lines
virt/kvm/guest_memfd.c +1
-1- folio =__kvm_gmem_get_pfn (file, slot, index, pfn, &is_prepared, max_order); + folio =__kvm_gmem_get_pfn (file, slot, index, pfn, max_order);if (IS_ERR (folio))virt/kvm/guest_memfd.c +4
-2return PTR_ERR (folio); -if (!is_prepared) - r =kvm_gmem_prepare_folio (kvm, slot, gfn, folio); +if (!folio_test_uptodate (folio)) { +clear_highpage (folio_page (folio,0 )); +folio_mark_uptodate (folio); + }... here only the first page is cleared.
I understand currently there's no huge folio coming out of gmem now, but since both
__kvm_gmem_get_pfn()andkvm_gmem_get_pfn()still have @max_level as output, it's kinda inconsistent here.I also see
kvm_gmem_fault_user_mapping()only clears the first page too, but I think that already has assumption that folio can never be huge currently?Given this, and the fact that the first patch of this series has introduced
Quoted 7 lines
WARN_ON_ONCE(folio_order(folio));in
kvm_gmem_get_folio(), I think it's fine to only clear the first page, but for the sake of consistency, perhaps we should just remove @max_order from__kvm_gmem_get_pfn()andkvm_gmem_get_pfn()?Then we can handle huge folio logic when that comes to play.
Sean had mentioned during the PUCK prior to this that he was okay with stripping out traces of hugepage support from kvm_gmem_populate() path, since it's bringing about unecessary complexity for a use-case we'll potentially never support. We will however eventually support hugepages outside the kvm_gmem_populate() path, and the bits and pieces of the API that plumb those details into KVM MMU code are more useful to keep around since there's existing hugepage support in KVM MMU that make it clearer where/when we'll need it. So I'm not sure we gain much from the churn of stripping it out. However, if as part of wiring up hugepage support those interfaces prove insufficient, then I wouldn't be opposed to similarly adding pre-patches to strip it out for a cleaner base implementation, but I don't really think that need to be part fo this series which is focused more on the population path rather than fault handling at run-time, so I've left things as-is for v3 for now.
Quoted 5 lines
Btw:
I actually looked into the RFC v1 discussion but the code there actually does a loop to clear all pages in the folio. There were some other
The thinking there was to try to not actively break the hugepage-related bits that were already in place, but since we decided to implement hugepage-related stuff for kvm_gmem_populate() as a clean follow-up implementation, there's no need to consider the hugepage case and loop through the page.
Quoted 9 lines
discussions about AFAICT they were more related to issues regarding to "mark entire folio as uptodate while only one page is processed in
post_populate()".Btw2:
There was also discussion that clearing page isn't required for TDX. To that end, maybe we can remove clearing page from gmem common code but to SEV code, e.g., as part of "folio preparation"?
I think we are considering this approach for TDX and there was some discussion of having gmem-internal flags to select for this type of handling, but I think that would make more sense as part of TDX-specific enablement of in-place conversion. I'm planning to post the SNP-specific enablement of in-place conversion patches on top of this series, so maybe we can consider this in response to that series or as part of the TDX-specific enablement.
-Mike
PATCH v2 3/5 KVM: SEV: Document/enforce page-alignment for KVM_SEV_SNP_LAUNCH_UPDATE
In reply to Michael RothIn the past, KVM_SEV_SNP_LAUNCH_UPDATE accepted a non-page-aligned 'uaddr' parameter to copy data from, but continuing to support this with new functionality like in-place conversion and hugepages in the pipeline has proven to be more trouble than it is worth, since there are no known users that have been identified who use a non-page-aligned 'uaddr' parameter.
Rather than locking guest_memfd into continuing to support this, go ahead and document page-alignment as a requirement and begin enforcing this in the handling function.
--- Documentation/virt/kvm/x86/amd-memory-encryption.rst | 2 +- arch/x86/kvm/svm/sev.c | 6 +++++- 2 files changed, 6 insertions(+), 2 deletions(-)
Documentation/virt/kvm/x86/amd-memory-encryption.rst +1 -1
@@ -523,7 +523,7 @@ Returns: 0 on success, < 0 on error, -EAGAIN if caller should retry struct kvm_sev_snp_launch_update { __u64 gfn_start; /* Guest page number to load/encrypt data into. */ - __u64 uaddr; /* Userspace address of data to be loaded/encrypted. */ + __u64 uaddr; /* 4k-aligned address of data to be loaded/encrypted. */ __u64 len; /* 4k-aligned length in bytes to copy into guest memory.*/ __u8 type; /* The type of the guest pages being initialized. */ __u8 pad0;
arch/x86/kvm/svm/sev.c +5 -1
@@ -2366,6 +2366,11 @@ static int snp_launch_update(struct kvm *kvm, struct kvm_sev_cmd *argp) params.type != KVM_SEV_SNP_PAGE_TYPE_CPUID))return -EINVAL ; + src = params.type ==KVM_SEV_SNP_PAGE_TYPE_ZERO ?NULL :u64_to_user_ptr (params.uaddr); + +if (!PAGE_ALIGNED (src)) +return -EINVAL ; + npages = params.len /PAGE_SIZE ;/*
@@ -2397,7 +2402,6 @@ static int snp_launch_update(struct kvm *kvm, struct kvm_sev_cmd *argp) sev_populate_args.sev_fd = argp->sev_fd; sev_populate_args.type = params.type; - src = params.type ==KVM_SEV_SNP_PAGE_TYPE_ZERO ?NULL :u64_to_user_ptr (params.uaddr); count =kvm_gmem_populate (kvm, params.gfn_start, src, npages, sev_gmem_post_populate, &sev_populate_args);
Signature · 1 line
2.25.1
Re: [PATCH v2 3/5] KVM: SEV: Document/enforce page-alignment for KVM_SEV_SNP_LAUNCH_UPDATE
Quoted 8 lines · Michael Roth
In the past,
KVM_SEV_SNP_LAUNCH_UPDATEaccepted a non-page-aligned 'uaddr' parameter to copy data from, but continuing to support this with new functionality like in-place conversion and hugepages in the pipeline has proven to be more trouble than it is worth, since there are no known users that have been identified who use a non-page-aligned 'uaddr' parameter.
Quoted 5 lines
Rather than locking
guest_memfdinto continuing to support this, go ahead and document page-alignment as a requirement and begin enforcing this in the handling function.Signed-off-by Michael Roth <[email protected]>
PATCH v2 4/5 KVM: TDX: Document alignment requirements for KVM_TDX_INIT_MEM_REGION
In reply to Michael RothSince it was never possible to use a non-PAGE_SIZE-aligned @source_addr, go ahead and document this as a requirement. This is in preparation for enforcing page-aligned @source_addr for all architectures in guest_memfd.
--- Documentation/virt/kvm/x86/intel-tdx.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Documentation/virt/kvm/x86/intel-tdx.rst +1 -1
@@ -156,7 +156,7 @@ KVM_TDX_INIT_MEM_REGION :Returns: 0 on success, <0 on error Initialize @nr_pages TDX guest private memory starting from @gpa with userspace -provided data from @source_addr. +provided data from @source_addr. @source_addr must be PAGE_SIZE-aligned. Note, before calling this sub command, memory attribute of the range [gpa, gpa + nr_pages] needs to be private. Userspace can use
Signature · 1 line
2.25.1
Re: [PATCH v2 4/5] KVM: TDX: Document alignment requirements for KVM_TDX_INIT_MEM_REGION
Quoted 16 lines · Michael Roth
Since it was never possible to use a non-PAGE_SIZE-aligned @source_addr, go ahead and document this as a requirement. This is in preparation for enforcing page-aligned @source_addr for all architectures in
guest_memfd.Signed-off-by Michael Roth <[email protected]>--- Documentation/virt/kvm/x86/intel-tdx.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)Documentation/virt/kvm/x86/intel-tdx.rst +0
-0
Documentation/virt/kvm/x86/intel-tdx.rst KVM_TDX_INIT_MEM_REGION · line 159
Quoted 5 lines
Documentation/virt/kvm/x86/intel-tdx.rst +1
-1@@ -156,7 +156,7 @@ KVM_TDX_INIT_MEM_REGION :Returns: 0 on success, <0 on error Initialize @nr_pages TDX guest private memory starting from @gpa with userspace -provided data from @source_addr. +provided data from @source_addr. @source_addr must be PAGE_SIZE-aligned.
PATCH v2 5/5 KVM: guest_memfd: GUP source pages prior to populating guest memory
In reply to Michael RothCurrently the post-populate callbacks handle copying source pages into private GPA ranges backed by guest_memfd, where kvm_gmem_populate() acquires the filemap invalidate lock, then calls a post-populate callback which may issue a get_user_pages() on the source pages prior to copying them into the private GPA (e.g. TDX).
This will not be compatible with in-place conversion, where the userspace page fault path will attempt to acquire filemap invalidate lock while holding the mm->mmap_lock, leading to a potential ABBA deadlock[1].
Address this by hoisting the GUP above the filemap invalidate lock so that these page faults path can be taken early, prior to acquiring the filemap invalidate lock.
It's not currently clear whether this issue is reachable with the current implementation of guest_memfd, which doesn't support in-place conversion, however it does provide a consistent mechanism to provide stable source/target PFNs to callbacks rather than punting to vendor-specific code, which allows for more commonality across architectures, which may be worthwhile even without in-place conversion.
As part of this change, also begin enforcing that the 'src' argument to kvm_gmem_populate() must be page-aligned, as this greatly reduces the complexity around how the post-populate callbacks are implemented, and since no current in-tree users support using a non-page-aligned 'src' argument.
--- arch/x86/kvm/svm/sev.c | 32 ++++++++------- arch/x86/kvm/vmx/tdx.c | 15 +------ include/linux/kvm_host.h | 4 +- virt/kvm/guest_memfd.c | 84 +++++++++++++++++++++++++++------------- 4 files changed, 77 insertions(+), 58 deletions(-)
arch/x86/kvm/svm/sev.c +17 -15
@@ -2268,7 +2268,7 @@ struct sev_gmem_populate_args { };static int sev_gmem_post_populate (struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, -void __user *src,void *opaque) +struct page *src_page,void *opaque) {struct sev_gmem_populate_args *sev_populate_args = opaque;struct sev_data_snp_launch_update fw_args = {0 };
@@ -2277,7 +2277,7 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn,int level;int ret; -if (WARN_ON_ONCE (sev_populate_args->type !=KVM_SEV_SNP_PAGE_TYPE_ZERO && !src)) +if (WARN_ON_ONCE (sev_populate_args->type !=KVM_SEV_SNP_PAGE_TYPE_ZERO && !src_page))return -EINVAL ; ret =snp_lookup_rmpentry ((u64 )pfn, &assigned, &level);
@@ -2288,14 +2288,14 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, goto out; } -if (src) { -void *vaddr =kmap_local_pfn (pfn); +if (src_page) { +void *src_vaddr =kmap_local_pfn (page_to_pfn (src_page)); +void *dst_vaddr =kmap_local_pfn (pfn); -if (copy_from_user (vaddr, src,PAGE_SIZE )) { - ret = -EFAULT ; - goto out; - } -kunmap_local (vaddr); +memcpy (dst_vaddr, src_vaddr,PAGE_SIZE ); + +kunmap_local (src_vaddr); +kunmap_local (dst_vaddr); } ret = rmp_make_private(pfn, gfn << PAGE_SHIFT, PG_LEVEL_4K,
@@ -2325,17 +2325,19 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn,if (ret && !snp_page_reclaim (kvm, pfn) && sev_populate_args->type ==KVM_SEV_SNP_PAGE_TYPE_CPUID && sev_populate_args->fw_error ==SEV_RET_INVALID_PARAM ) { -void *vaddr =kmap_local_pfn (pfn); +void *src_vaddr =kmap_local_pfn (page_to_pfn (src_page)); +void *dst_vaddr =kmap_local_pfn (pfn); -if (copy_to_user (src, vaddr,PAGE_SIZE )) -pr_debug ("Failed to write CPUID page back to userspace\n" ); +memcpy (src_vaddr, dst_vaddr,PAGE_SIZE ); -kunmap_local (vaddr); +kunmap_local (src_vaddr); +kunmap_local (dst_vaddr); } out: -pr_debug ("%s: exiting with return code %d (fw_error %d)\n" , - __func__, ret, sev_populate_args->fw_error); +if (ret) +pr_debug ("%s: error updating GFN %llx, return code %d (fw_error %d)\n" , + __func__, gfn, ret, sev_populate_args->fw_error);return ret; }
arch/x86/kvm/vmx/tdx.c +1 -14
@@ -3118,34 +3118,21 @@ struct tdx_gmem_post_populate_arg { };static int tdx_gmem_post_populate (struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, -void __user *src,void *_arg) +struct page *src_page,void *_arg) {struct tdx_gmem_post_populate_arg *arg = _arg;struct kvm_tdx *kvm_tdx =to_kvm_tdx (kvm);u64 err, entry, level_state;gpa_t gpa =gfn_to_gpa (gfn); -struct page *src_page;int ret, i;if (KVM_BUG_ON (kvm_tdx->page_add_src, kvm))return -EIO ; -/* -* Get the source page if it has been faulted in. Return failure if the -* source page has been swapped out or unmapped in primary memory. -*/ - ret =get_user_pages_fast ((unsigned long )src,1 ,0 , &src_page); -if (ret <0 ) -return ret; -if (ret !=1 ) -return -ENOMEM ; - kvm_tdx->page_add_src = src_page; ret =kvm_tdp_mmu_map_private_pfn (arg->vcpu, gfn, pfn); kvm_tdx->page_add_src =NULL ; -put_page (src_page); -if (ret || !(arg->flags &KVM_TDX_MEASURE_MEMORY_REGION ))return ret;
include/linux/kvm_host.h +2 -2
@@ -2566,7 +2566,7 @@ int kvm_arch_gmem_prepare(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, int max_ord* @gfn: starting GFN to be populated * @src: userspace-provided buffer containing data to copy into GFN range * (passed to @post_populate, and incremented on each iteration -* if not NULL) +* if not NULL). Must be page-aligned. * @npages: number of pages to copy from userspace-buffer * @post_populate: callback to issue for each gmem page that backs the GPA * range
@@ -2581,7 +2581,7 @@ int kvm_arch_gmem_prepare(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, int max_ord* Returns the number of pages that were populated. */ typedef int (*kvm_gmem_populate_cb )(struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, -void __user *src,void *opaque); +struct page *page,void *opaque);long kvm_gmem_populate (struct kvm *kvm,gfn_t gfn,void __user *src,long npages,kvm_gmem_populate_cb post_populate,void *opaque);
virt/kvm/guest_memfd.c +57 -27
122 lines not on this page · raw message
Signature · 1 line
2.25.1
Re: [PATCH v2 5/5] KVM: guest_memfd: GUP source pages prior to populating guest memory
Quoted 29 lines · Michael Roth
Currently the post-populate callbacks handle copying source pages into private GPA ranges backed by
guest_memfd, wherekvm_gmem_populate()acquires the filemap invalidate lock, then calls a post-populate callback which may issue aget_user_pages()on the source pages prior to copying them into the private GPA (e.g. TDX).This will not be compatible with in-place conversion, where the userspace page fault path will attempt to acquire filemap invalidate lock while holding the
mm->mmap_lock, leading to a potential ABBA deadlock[1].Address this by hoisting the GUP above the filemap invalidate lock so that these page faults path can be taken early, prior to acquiring the filemap invalidate lock.
It's not currently clear whether this issue is reachable with the current implementation of
guest_memfd, which doesn't support in-place conversion, however it does provide a consistent mechanism to provide stable source/target PFNs to callbacks rather than punting to vendor-specific code, which allows for more commonality across architectures, which may be worthwhile even without in-place conversion.As part of this change, also begin enforcing that the 'src' argument to
kvm_gmem_populate()must be page-aligned, as this greatly reduces the complexity around how the post-populate callbacks are implemented, and since no current in-tree users support using a non-page-aligned 'src' argument.
Quoted 6 lines
Suggested-by Sean Christopherson <[email protected]>Co-developed-by Sean Christopherson <[email protected]>Signed-off-by Sean Christopherson <[email protected]>Co-developed-by Vishal Annapurve <[email protected]>Signed-off-by Vishal Annapurve <[email protected]>Signed-off-by Michael Roth <[email protected]>
Re: [PATCH v2 5/5] KVM: guest_memfd: GUP source pages prior to populating guest memory
In reply to Michael RothQuoted 10 lines · Michael Roth
Currently the post-populate callbacks handle copying source pages into private GPA ranges backed by
guest_memfd, wherekvm_gmem_populate()acquires the filemap invalidate lock, then calls a post-populate callback which may issue aget_user_pages()on the source pages prior to copying them into the private GPA (e.g. TDX).This will not be compatible with in-place conversion, where the userspace page fault path will attempt to acquire filemap invalidate lock while holding the
mm->mmap_lock, leading to a potential ABBA deadlock[1].
Nit: there's no link to mention [1].
[...]
Quoted 7 lines
Suggested-by Sean Christopherson <[email protected]>Co-developed-by Sean Christopherson <[email protected]>Signed-off-by Sean Christopherson <[email protected]>Co-developed-by Vishal Annapurve <[email protected]>Signed-off-by Vishal Annapurve <[email protected]>Signed-off-by Michael Roth <[email protected]>
[...]
arch/x86/kvm/svm/sev.c static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, · line 2292
Quoted 2 lines
arch/x86/kvm/svm/sev.c +2
-0+if (src_page) { +void *src_vaddr =kmap_local_pfn (page_to_pfn (src_page));
Nit: maybe you can use kmap_local_page(src_page) directly.
Quoted 15 lines
arch/x86/kvm/svm/sev.c +5
-5+void *dst_vaddr =kmap_local_pfn (pfn); -if (copy_from_user (vaddr, src,PAGE_SIZE )) { - ret = -EFAULT ; - goto out; - } -kunmap_local (vaddr); +memcpy (dst_vaddr, src_vaddr,PAGE_SIZE ); + +kunmap_local (src_vaddr); +kunmap_local (dst_vaddr); } ret = rmp_make_private(pfn, gfn << PAGE_SHIFT, PG_LEVEL_4K,
arch/x86/kvm/svm/sev.c static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn, · line 2328
Quoted 5 lines
arch/x86/kvm/svm/sev.c +1
-1@@ -2325,17 +2325,19 @@ static int sev_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn,if (ret && !snp_page_reclaim (kvm, pfn) && sev_populate_args->type ==KVM_SEV_SNP_PAGE_TYPE_CPUID && sev_populate_args->fw_error ==SEV_RET_INVALID_PARAM ) { -void *vaddr =kmap_local_pfn (pfn); +void *src_vaddr =kmap_local_pfn (page_to_pfn (src_page));
Ditto.
Re: [PATCH v2 5/5] KVM: guest_memfd: GUP source pages prior to populating guest memory
In reply to Michael RothQuoted 10 lines · Michael Roth
arch/x86/kvm/vmx/tdx.c +1
-1@@ -3118,34 +3118,21 @@ struct tdx_gmem_post_populate_arg { };static int tdx_gmem_post_populate (struct kvm *kvm,gfn_t gfn,kvm_pfn_t pfn, -void __user *src,void *_arg) +struct page *src_page,void *_arg)
arch/x86/kvm/vmx/tdx.c struct tdx_gmem_post_populate_arg { · line 3130
Quoted 10 lines
arch/x86/kvm/vmx/tdx.c +0
-1{struct tdx_gmem_post_populate_arg *arg = _arg;struct kvm_tdx *kvm_tdx =to_kvm_tdx (kvm);u64 err, entry, level_state;gpa_t gpa =gfn_to_gpa (gfn); -struct page *src_page;int ret, i;if (KVM_BUG_ON (kvm_tdx->page_add_src, kvm))return -EIO ;
Check if src_page is NULL.
arch/x86/kvm/vmx/tdx.c +3 -0
@@ -3190,6 +3190,9 @@ static int tdx_gmem_post_populate(struct kvm *kvm, gfn_t gfn, kvm_pfn_t pfn,if (KVM_BUG_ON (kvm_tdx->page_add_src, kvm))return -EIO ; +if (!src_page) +return -EOPNOTSUPP ; + kvm_tdx->page_add_src = src_page; ret =kvm_tdp_mmu_map_private_pfn (arg->vcpu, gfn, pfn); kvm_tdx->page_add_src =NULL ;
Quoted 19 lines
arch/x86/kvm/vmx/tdx.c +0
-12-/* -* Get the source page if it has been faulted in. Return failure if the -* source page has been swapped out or unmapped in primary memory. -*/ - ret =get_user_pages_fast ((unsigned long )src,1 ,0 , &src_page); -if (ret <0 ) -return ret; -if (ret !=1 ) -return -ENOMEM ; - kvm_tdx->page_add_src = src_page; ret =kvm_tdp_mmu_map_private_pfn (arg->vcpu, gfn, pfn); kvm_tdx->page_add_src =NULL ; -put_page (src_page); -if (ret || !(arg->flags &KVM_TDX_MEASURE_MEMORY_REGION ))return ret;
...
Quoted 42 lines
long kvm_gmem_populate(struct kvm *kvm,
gfn_tstart_gfn, void__user*src, long npages,arch/x86/kvm/vmx/tdx.c +5
-11kvm_gmem_populate_cb post_populate, void *opaque) { struct kvm_memory_slot *slot; -void __user *p; -int ret =0 ;long i;@@ -834,6 +870,9 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longif (WARN_ON_ONCE (npages <=0 ))return -EINVAL ; +if (WARN_ON_ONCE (!PAGE_ALIGNED (src))) +return -EINVAL ; + slot =gfn_to_memslot (kvm, start_gfn);if (!kvm_slot_has_gmem (slot))return -EINVAL ;@@ -842,47 +881,38 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longif (!file)return -EFAULT ; -filemap_invalidate_lock (file->f_mapping); - npages =min_t (ulong, slot->npages - (start_gfn - slot->base_gfn), npages);for (i =0 ; i < npages; i++) { -struct folio *folio; -gfn_t gfn = start_gfn + i; -pgoff_t index =kvm_gmem_get_index (slot, gfn); -kvm_pfn_t pfn; +struct page *src_page =NULL ; +void __user *p;if (signal_pending (current)) { ret = -EINTR ;break ; } - folio =__kvm_gmem_get_pfn (file, slot, index, &pfn,NULL ); -if (IS_ERR (folio)) { - ret =PTR_ERR (folio);
arch/x86/kvm/vmx/tdx.c long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long · line 900
Quoted 10 lines
arch/x86/kvm/vmx/tdx.c +6
-3-break ; - } + p = src ? src + i *PAGE_SIZE :NULL ; -folio_unlock (folio); +if (p) { + ret =get_user_pages_fast ((unsigned long )p,1 ,0 , &src_page); +if (ret <0 ) +break ; +if (ret !=1 ) {
Put pages in this case? e.g.,
virt/kvm/guest_memfd.c +3 -0
@@ -1645,6 +1645,9 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longif (ret <0 )break ;if (ret !=1 ) { +while (ret--) +put_page (src_page++); + ret = -ENOMEM ;break ; }
Quoted 34 lines
arch/x86/kvm/vmx/tdx.c +8
-13+ ret = -ENOMEM ; +break ; + } + } - ret = -EINVAL ; -if (!kvm_range_has_memory_attributes (kvm, gfn, gfn +1 , -KVM_MEMORY_ATTRIBUTE_PRIVATE , -KVM_MEMORY_ATTRIBUTE_PRIVATE )) - goto put_folio_and_exit; + ret =__kvm_gmem_populate (kvm, slot, file, start_gfn + i, src_page, + post_populate, opaque); - p = src ? src + i *PAGE_SIZE :NULL ; - ret =post_populate (kvm, gfn, pfn, p, opaque); -if (!ret) -folio_mark_uptodate (folio); +if (src_page) +put_page (src_page); -put_folio_and_exit: -folio_put (folio);if (ret)break ; } -filemap_invalidate_unlock (file->f_mapping); -return ret && !i ? ret : i; }EXPORT_SYMBOL_FOR_KVM_INTERNAL (kvm_gmem_populate);Signature · 1 line
2.25.1
Re: [PATCH v2 5/5] KVM: guest_memfd: GUP source pages prior to populating guest memory
Quoted 38 lines · Yan Zhao
Quoted 23 lines
(quoted diff) +2
-9@@ -842,47 +881,38 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long if (!file) return -EFAULT; - filemap_invalidate_lock(file->f_mapping); - npages = min_t(ulong, slot->npages - (start_gfn - slot->base_gfn), npages); for (i = 0; i < npages; i++) { - struct folio *folio; - gfn_t gfn = start_gfn + i; - pgoff_t index = kvm_gmem_get_index(slot, gfn); - kvm_pfn_t pfn; + struct page *src_page = NULL; + void __user *p; if (signal_pending(current)) { ret = -EINTR; break; } - folio = __kvm_gmem_get_pfn(file, slot, index, &pfn, NULL); - if (IS_ERR(folio)) { - ret = PTR_ERR(folio);Review long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long · line 900
Quoted 10 lines
(quoted diff) +6
-3- break; - } + p = src ? src + i * PAGE_SIZE : NULL; - folio_unlock(folio); + if (p) { + ret = get_user_pages_fast((unsigned long)p, 1, 0, &src_page); + if (ret < 0) + break; + if (ret != 1) {Put pages in this case? e.g.,
virt/kvm/guest_memfd.c long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, long · line 1648
Quoted 4 lines
virt/kvm/guest_memfd.c +1
-0@@ -1645,6 +1645,9 @@ long kvm_gmem_populate(struct kvm *kvm, gfn_t start_gfn, void __user *src, longif (ret <0 )break ;if (ret !=1 ) { +while (ret--)
Oops. Need to check if ret == 0, and looks put_page() is not required in this case given nr_pages == 1.
So, please ignore this comment.
Quoted 43 lines
virt/kvm/guest_memfd.c +2
-0+put_page (src_page++); + ret = -ENOMEM ;break ; }Quoted 38 lines
(quoted diff) +8
-13+ ret = -ENOMEM; + break; + } + } - ret = -EINVAL; - if (!kvm_range_has_memory_attributes(kvm, gfn, gfn + 1, - KVM_MEMORY_ATTRIBUTE_PRIVATE, - KVM_MEMORY_ATTRIBUTE_PRIVATE)) - goto put_folio_and_exit; + ret = __kvm_gmem_populate(kvm, slot, file, start_gfn + i, src_page, + post_populate, opaque); - p = src ? src + i * PAGE_SIZE : NULL; - ret = post_populate(kvm, gfn, pfn, p, opaque); - if (!ret) - folio_mark_uptodate(folio); + if (src_page) + put_page(src_page); -put_folio_and_exit: - folio_put(folio); if (ret) break; } - filemap_invalidate_unlock(file->f_mapping); - return ret && !i ? ret : i; } EXPORT_SYMBOL_FOR_KVM_INTERNAL(kvm_gmem_populate);Signature · 1 line
2.25.1
Quoted 4 lines · Michael Roth
Known issues / TODO
-------------------- Compile-tested only for the TDX bits (testing/feedback welcome!)
Applied this series to 6.19-rc1, and tested booting/destroying TD worked fine, so: