linux-next: manual merge of the kvm tree with the arm64 tree

Posted by Stephen Rothwell 4 weeks ago

Hi all,

Today's linux-next merge of the kvm tree got a conflict in:

  arch/arm64/kvm/guest.c

between commit:

  25c17c4b55de ("hugetlb: arm64: add mte support")

from the arm64 tree and commit:

  570d666c11af ("KVM: arm64: Use __gfn_to_page() when copying MTE tags to/from userspace")

from the kvm tree.

I fixed it up (see below) and can carry the fix as necessary. This
is now fixed as far as linux-next is concerned, but any non trivial
conflicts should be mentioned to your upstream maintainer when your tree
is submitted for merging.  You may also want to consider cooperating
with the maintainer of the conflicting tree to minimise any particularly
complex conflicts.

-- 
Cheers,
Stephen Rothwell

diff --cc arch/arm64/kvm/guest.c
index e738a353b20e,4cd7ffa76794..000000000000
--- a/arch/arm64/kvm/guest.c
+++ b/arch/arm64/kvm/guest.c
@@@ -1051,13 -1051,11 +1051,12 @@@ int kvm_vm_ioctl_mte_copy_tags(struct k
  	}
  
  	while (length > 0) {
- 		kvm_pfn_t pfn = gfn_to_pfn_prot(kvm, gfn, write, NULL);
+ 		struct page *page = __gfn_to_page(kvm, gfn, write);
  		void *maddr;
  		unsigned long num_tags;
- 		struct page *page;
 +		struct folio *folio;
  
- 		if (is_error_noslot_pfn(pfn)) {
+ 		if (!page) {
  			ret = -EFAULT;
  			goto out;
  		}
@@@ -1099,12 -1090,8 +1097,12 @@@
  			/* uaccess failed, don't leave stale tags */
  			if (num_tags != MTE_GRANULES_PER_PAGE)
  				mte_clear_page_tags(maddr);
 -			set_page_mte_tagged(page);
 +			if (folio_test_hugetlb(folio))
 +				folio_set_hugetlb_mte_tagged(folio);
 +			else
 +				set_page_mte_tagged(page);
 +
- 			kvm_release_pfn_dirty(pfn);
+ 			kvm_release_page_dirty(page);
  		}
  
  		if (num_tags != MTE_GRANULES_PER_PAGE) {

Re: linux-next: manual merge of the kvm tree with the arm64 tree

Posted by Catalin Marinas 3 weeks, 6 days ago

On Mon, Oct 28, 2024 at 05:03:10PM +1100, Stephen Rothwell wrote:
> Today's linux-next merge of the kvm tree got a conflict in:
> 
>   arch/arm64/kvm/guest.c
> 
> between commit:
> 
>   25c17c4b55de ("hugetlb: arm64: add mte support")
> 
> from the arm64 tree and commit:
> 
>   570d666c11af ("KVM: arm64: Use __gfn_to_page() when copying MTE tags to/from userspace")
> 
> from the kvm tree.
[...]
> diff --cc arch/arm64/kvm/guest.c
> index e738a353b20e,4cd7ffa76794..000000000000
> --- a/arch/arm64/kvm/guest.c
> +++ b/arch/arm64/kvm/guest.c
> @@@ -1051,13 -1051,11 +1051,12 @@@ int kvm_vm_ioctl_mte_copy_tags(struct k
>   	}
>   
>   	while (length > 0) {
> - 		kvm_pfn_t pfn = gfn_to_pfn_prot(kvm, gfn, write, NULL);
> + 		struct page *page = __gfn_to_page(kvm, gfn, write);
>   		void *maddr;
>   		unsigned long num_tags;
> - 		struct page *page;
>  +		struct folio *folio;
>   
> - 		if (is_error_noslot_pfn(pfn)) {
> + 		if (!page) {
>   			ret = -EFAULT;
>   			goto out;
>   		}
> @@@ -1099,12 -1090,8 +1097,12 @@@
>   			/* uaccess failed, don't leave stale tags */
>   			if (num_tags != MTE_GRANULES_PER_PAGE)
>   				mte_clear_page_tags(maddr);
>  -			set_page_mte_tagged(page);
>  +			if (folio_test_hugetlb(folio))
>  +				folio_set_hugetlb_mte_tagged(folio);
>  +			else
>  +				set_page_mte_tagged(page);
>  +
> - 			kvm_release_pfn_dirty(pfn);
> + 			kvm_release_page_dirty(page);
>   		}
>   
>   		if (num_tags != MTE_GRANULES_PER_PAGE) {

Thanks Stephen. The resolution looks fine and I'm happy to leave to
Linus to fix it up during the merging window.

To the KVM maintainers, if you prefer a conflict-free linux-next, feel
free to pull the arm64 for-next/mte branch with the above commit (and a
kselftest). The other way around is not something I'd suggest we do,
there are over 80 patches in that kvm series.

-- 
Catalin

Re: linux-next: manual merge of the kvm tree with the arm64 tree

Posted by Sean Christopherson 3 weeks, 6 days ago

On Mon, Oct 28, 2024, Catalin Marinas wrote:
> On Mon, Oct 28, 2024 at 05:03:10PM +1100, Stephen Rothwell wrote:
> > Today's linux-next merge of the kvm tree got a conflict in:
> > 
> >   arch/arm64/kvm/guest.c
> > 
> > between commit:
> > 
> >   25c17c4b55de ("hugetlb: arm64: add mte support")
> > 
> > from the arm64 tree and commit:
> > 
> >   570d666c11af ("KVM: arm64: Use __gfn_to_page() when copying MTE tags to/from userspace")
> > 
> > from the kvm tree.
> [...]
> > diff --cc arch/arm64/kvm/guest.c
> > index e738a353b20e,4cd7ffa76794..000000000000
> > --- a/arch/arm64/kvm/guest.c
> > +++ b/arch/arm64/kvm/guest.c
> > @@@ -1051,13 -1051,11 +1051,12 @@@ int kvm_vm_ioctl_mte_copy_tags(struct k
> >   	}
> >   
> >   	while (length > 0) {
> > - 		kvm_pfn_t pfn = gfn_to_pfn_prot(kvm, gfn, write, NULL);
> > + 		struct page *page = __gfn_to_page(kvm, gfn, write);
> >   		void *maddr;
> >   		unsigned long num_tags;
> > - 		struct page *page;
> >  +		struct folio *folio;
> >   
> > - 		if (is_error_noslot_pfn(pfn)) {
> > + 		if (!page) {
> >   			ret = -EFAULT;
> >   			goto out;
> >   		}
> > @@@ -1099,12 -1090,8 +1097,12 @@@
> >   			/* uaccess failed, don't leave stale tags */
> >   			if (num_tags != MTE_GRANULES_PER_PAGE)
> >   				mte_clear_page_tags(maddr);
> >  -			set_page_mte_tagged(page);
> >  +			if (folio_test_hugetlb(folio))
> >  +				folio_set_hugetlb_mte_tagged(folio);
> >  +			else
> >  +				set_page_mte_tagged(page);
> >  +
> > - 			kvm_release_pfn_dirty(pfn);
> > + 			kvm_release_page_dirty(page);
> >   		}
> >   
> >   		if (num_tags != MTE_GRANULES_PER_PAGE) {
> 
> Thanks Stephen. The resolution looks fine

Looks correct to my eyes, too.  Thanks Stephen!

> and I'm happy to leave to Linus to fix it up during the merging window.
> 
> To the KVM maintainers, if you prefer a conflict-free linux-next, feel
> free to pull the arm64 for-next/mte branch with the above commit (and a
> kselftest). The other way around is not something I'd suggest we do,
> there are over 80 patches in that kvm series.

Not feeling lucky today? ;-)