LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Claudio Imbrenda <imbrenda@linux.ibm.com>
To: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: kvm@vger.kernel.org, cohuck@redhat.com, frankja@linux.ibm.com,
	thuth@redhat.com, pasic@linux.ibm.com, david@redhat.com,
	linux-s390@vger.kernel.org, linux-kernel@vger.kernel.org,
	Ulrich.Weigand@de.ibm.com
Subject: Re: [PATCH v4 05/14] KVM: s390: pv: leak the ASCE page when destroy fails
Date: Mon, 6 Sep 2021 17:54:53 +0200	[thread overview]
Message-ID: <20210906175453.5b98ca26@p-imbrenda> (raw)
In-Reply-To: <36ce2f10-a65d-ff2a-3a11-8f2cd853f3e9@de.ibm.com>

On Mon, 6 Sep 2021 17:32:36 +0200
Christian Borntraeger <borntraeger@de.ibm.com> wrote:

> The subject should say
> 
> KVM: s390: pv: leak the topmost page table when destroy fails
> 
> 
> On 18.08.21 15:26, Claudio Imbrenda wrote:
> > When a protected VM is created, the topmost level of page tables of its
> > ASCE is marked by the Ultravisor; any attempt to use that memory for
> > protected virtualization will result in failure.  
> 
> 
> maybe rephrase that to
> Each secure guest must have a unique address space control element and we
> must avoid that new guests will use the same ASCE to avoid an error. As
> the ASCE mostly consists of the top most page table address (and flags)
> we must not return that memory to the pool unless the ASCE is no longer
> used.
> 
> Only a a successful Destroy Configuration UVC will make the ASCE no longer
> collide.
> When the Destroy Configuration UVC fails, the ASCE cannot be reused for a
> secure guest ASCE. To avoid a collision, it must not be used again.
> 

ok

>   
> > Only a successful Destroy Configuration UVC will remove the marking.
> > 
> > When the Destroy Configuration UVC fails, the topmost level of page
> > tables of the VM does not get its marking cleared; to avoid issues it
> > must not be used again.
> > 
> > This is a permanent error and the page becomes in practice unusable, so
> > we set it aside and leak it.  
> 
> Maybe add: on failure we already leak other memory that has ultravisor marking (the
> variable and base storage for a guest) and not setting the ASCE aside (by
> leaking the topmost page table) was an oversight.
> 
> Or something like that
> 
> maybe also add that we usually do not expect to see such error under normal
> circumstances.
> 

makes sense

> > 
> > Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
> > ---
> >   arch/s390/include/asm/gmap.h |  2 ++
> >   arch/s390/kvm/pv.c           |  4 ++-
> >   arch/s390/mm/gmap.c          | 55 ++++++++++++++++++++++++++++++++++++
> >   3 files changed, 60 insertions(+), 1 deletion(-)
> > 
> > diff --git a/arch/s390/include/asm/gmap.h b/arch/s390/include/asm/gmap.h
> > index 40264f60b0da..746e18bf8984 100644
> > --- a/arch/s390/include/asm/gmap.h
> > +++ b/arch/s390/include/asm/gmap.h
> > @@ -148,4 +148,6 @@ void gmap_sync_dirty_log_pmd(struct gmap *gmap, unsigned long dirty_bitmap[4],
> >   			     unsigned long gaddr, unsigned long vmaddr);
> >   int gmap_mark_unmergeable(void);
> >   void s390_reset_acc(struct mm_struct *mm);
> > +void s390_remove_old_asce(struct gmap *gmap);
> > +int s390_replace_asce(struct gmap *gmap);
> >   #endif /* _ASM_S390_GMAP_H */
> > diff --git a/arch/s390/kvm/pv.c b/arch/s390/kvm/pv.c
> > index 00d272d134c2..76b0d64ce8fa 100644
> > --- a/arch/s390/kvm/pv.c
> > +++ b/arch/s390/kvm/pv.c
> > @@ -168,9 +168,11 @@ int kvm_s390_pv_deinit_vm(struct kvm *kvm, u16 *rc, u16 *rrc)
> >   	atomic_set(&kvm->mm->context.is_protected, 0);
> >   	KVM_UV_EVENT(kvm, 3, "PROTVIRT DESTROY VM: rc %x rrc %x", *rc, *rrc);
> >   	WARN_ONCE(cc, "protvirt destroy vm failed rc %x rrc %x", *rc, *rrc);
> > -	/* Inteded memory leak on "impossible" error */
> > +	/* Intended memory leak on "impossible" error */
> >   	if (!cc)
> >   		kvm_s390_pv_dealloc_vm(kvm);
> > +	else
> > +		s390_replace_asce(kvm->arch.gmap);
> >   	return cc ? -EIO : 0;
> >   }
> >   
> > diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c
> > index 9bb2c7512cd5..5a138f6220c4 100644
> > --- a/arch/s390/mm/gmap.c
> > +++ b/arch/s390/mm/gmap.c
> > @@ -2706,3 +2706,58 @@ void s390_reset_acc(struct mm_struct *mm)
> >   	mmput(mm);
> >   }
> >   EXPORT_SYMBOL_GPL(s390_reset_acc);
> > +
> > +/*
> > + * Remove the topmost level of page tables from the list of page tables of
> > + * the gmap.
> > + * This means that it will not be freed when the VM is torn down, and needs
> > + * to be handled separately by the caller, unless an intentional leak is
> > + * intended.
> > + */
> > +void s390_remove_old_asce(struct gmap *gmap)
> > +{
> > +	struct page *old;
> > +
> > +	old = virt_to_page(gmap->table);
> > +	spin_lock(&gmap->guest_table_lock);
> > +	list_del(&old->lru);
> > +	spin_unlock(&gmap->guest_table_lock);
> > +	/* in case the ASCE needs to be "removed" multiple times */
> > +	INIT_LIST_HEAD(&old->lru);  
> shouldn't that also be under the spin_lock?
> 
> > +}
> > +EXPORT_SYMBOL_GPL(s390_remove_old_asce);
> > +
> > +/*
> > + * Try to replace the current ASCE with another equivalent one.
> > + * If the allocation of the new top level page table fails, the ASCE is not
> > + * replaced.
> > + * In any case, the old ASCE is removed from the list, therefore the caller
> > + * has to make sure to save a pointer to it beforehands, unless an
> > + * intentional leak is intended.
> > + */
> > +int s390_replace_asce(struct gmap *gmap)
> > +{
> > +	unsigned long asce;
> > +	struct page *page;
> > +	void *table;
> > +
> > +	s390_remove_old_asce(gmap);
> > +
> > +	page = alloc_pages(GFP_KERNEL_ACCOUNT, CRST_ALLOC_ORDER);
> > +	if (!page)
> > +		return -ENOMEM;  
> 
> It seems that we do not handle errors in our caller?

for now, but it doesn't hurt to report an error

> 
> > +	table = page_to_virt(page);
> > +	memcpy(table, gmap->table, 1UL << (CRST_ALLOC_ORDER + PAGE_SHIFT));
> > +
> > +	spin_lock(&gmap->guest_table_lock);
> > +	list_add(&page->lru, &gmap->crst_list);
> > +	spin_unlock(&gmap->guest_table_lock);
> > +
> > +	asce = (gmap->asce & ~PAGE_MASK) | __pa(table);  
> 
> Instead of PAGE_MASK better use _ASCE_ORIGIN ?

ok

> > +	WRITE_ONCE(gmap->asce, asce);
> > +	WRITE_ONCE(gmap->mm->context.gmap_asce, asce);
> > +	WRITE_ONCE(gmap->table, table);
> > +
> > +	return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(s390_replace_asce);
> >   


  reply	other threads:[~2021-09-06 15:56 UTC|newest]

Thread overview: 31+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-18 13:26 [PATCH v4 00/14] KVM: s390: pv: implement lazy destroy for reboot Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 01/14] KVM: s390: pv: add macros for UVC CC values Claudio Imbrenda
2021-08-31 13:43   ` Christian Borntraeger
2021-08-18 13:26 ` [PATCH v4 02/14] KVM: s390: pv: avoid double free of sida page Claudio Imbrenda
2021-08-31 13:55   ` Christian Borntraeger
2021-09-08 18:50     ` Claudio Imbrenda
2021-08-31 13:59   ` Janosch Frank
2021-08-18 13:26 ` [PATCH v4 03/14] KVM: s390: pv: avoid stalls for kvm_s390_pv_init_vm Claudio Imbrenda
2021-08-31 14:10   ` Christian Borntraeger
2021-08-18 13:26 ` [PATCH v4 04/14] KVM: s390: pv: avoid stalls when making pages secure Claudio Imbrenda
2021-08-31 14:32   ` Christian Borntraeger
2021-08-31 15:00     ` Claudio Imbrenda
2021-08-31 15:11       ` Christian Borntraeger
2021-08-18 13:26 ` [PATCH v4 05/14] KVM: s390: pv: leak the ASCE page when destroy fails Claudio Imbrenda
2021-09-06 15:32   ` Christian Borntraeger
2021-09-06 15:54     ` Claudio Imbrenda [this message]
2021-08-18 13:26 ` [PATCH v4 06/14] KVM: s390: pv: properly handle page flags for protected guests Claudio Imbrenda
2021-09-06 15:46   ` Christian Borntraeger
2021-09-06 15:56     ` Claudio Imbrenda
2021-09-06 16:16       ` Christian Borntraeger
2021-09-17 14:57         ` Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 07/14] KVM: s390: pv: handle secure storage violations " Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 08/14] KVM: s390: pv: handle secure storage exceptions for normal guests Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 09/14] KVM: s390: pv: refactor s390_reset_acc Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 10/14] KVM: s390: pv: usage counter instead of flag Claudio Imbrenda
2021-08-26  7:58   ` Janis Schoetterl-Glausch
2021-08-18 13:26 ` [PATCH v4 11/14] KVM: s390: pv: add export before import Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 12/14] KVM: s390: pv: module parameter to fence lazy destroy Claudio Imbrenda
2021-08-18 13:26 ` [PATCH v4 13/14] KVM: s390: pv: lazy destroy for reboot Claudio Imbrenda
2021-08-26  8:33   ` Janis Schoetterl-Glausch
2021-08-18 13:26 ` [PATCH v4 14/14] KVM: s390: pv: avoid export before import if possible Claudio Imbrenda

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20210906175453.5b98ca26@p-imbrenda \
    --to=imbrenda@linux.ibm.com \
    --cc=Ulrich.Weigand@de.ibm.com \
    --cc=borntraeger@de.ibm.com \
    --cc=cohuck@redhat.com \
    --cc=david@redhat.com \
    --cc=frankja@linux.ibm.com \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-s390@vger.kernel.org \
    --cc=pasic@linux.ibm.com \
    --cc=thuth@redhat.com \
    --subject='Re: [PATCH v4 05/14] KVM: s390: pv: leak the ASCE page when destroy fails' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).