From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-pl1-f201.google.com (mail-pl1-f201.google.com [209.85.214.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 844073B27E2 for ; Wed, 3 Jun 2026 23:29:16 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.214.201 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1780529357; cv=none; b=FsRXBDdQ6e5XIzqPS89ArnC0EMBY7Navr/Lb/Ons1SJNOqLBedS2E4oKtrcAnSBdkL+Mi3Nn33ZjBOBPxlZ9ihdgoEpqDIq1Yc5CsLrrksCiXhgg8RtjiQUSzODyCAX7R/JTBrDCOgT9M6wnG3GS2cnRtWpWShZPlu6S8ihmO0I= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1780529357; c=relaxed/simple; bh=1EwM9K/dFJXLqTPP7s4Oxg+UqsfTRVwjax+NCb2B6mI=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Cc:Content-Type; b=d9sK9f+y5jnquqjTJTF+mIN4jVX2Te3id/zqQ8dms6pkZOSaq8dMJ5todi8Vhoq41WyJuc1rTfXiuzsp/1+eel3Fq37kkIQ6q6MfiohjbByf8kqaOhLPH1ZoVm0DcpTU9rpQP8ymCdtC9wmQuwQcgG9sg6SmOVHknGzoMWOipVc= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--seanjc.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=oxN6N5Hr; arc=none smtp.client-ip=209.85.214.201 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--seanjc.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="oxN6N5Hr" Received: by mail-pl1-f201.google.com with SMTP id d9443c01a7336-2c0c331eda3so1052475ad.2 for ; Wed, 03 Jun 2026 16:29:16 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20251104; t=1780529356; x=1781134156; darn=vger.kernel.org; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:from:to:cc:subject:date:message-id:reply-to; bh=m9TAkhtRZEf2sqIn4AaIXqhnULFz6dUCOQHgczry4fc=; b=oxN6N5HrsOU9rXTakI+OJ6sFO3BJ6cV5fAAp4Qtw78q7LY+hMJq91gEjjQgBmJBdC7 8Jjyppyl6A8GmG0UDrYkLzZDnni3nEHgba0C00aiQS4EEgkxIWj1Ki6qW7Oy7FQW5SLD gyOo/JeR46Oy3i6TtIBMarOVPMPg/ba2V8ta/MRoj7QngJaOwk2Ug7yP+0h1jqvQhsrG yRXV5y7mr5gK4EQMIyYOL/19kr4+KYrdItPtyk52i4DAsSSVFnywaS4xSNlGWVRVthD7 GIQan4C644fJbqxy2Bmmrz3mPWHX3eOqr6C/eNR2YkrWDn3I5keD5mqQmw8fABxcp7t7 zdzg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1780529356; x=1781134156; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=m9TAkhtRZEf2sqIn4AaIXqhnULFz6dUCOQHgczry4fc=; b=cP62FOmKn/tcjryfab9sBJ5akrwkSl/qFVVnepjCMVztdsvgeHVQbGkWy5X99OtX42 FkQc+vuXKfVKqbfh7y9CMviENfxgp0C5oAEf0neiCYca9DLJk2XDE/sgTDbq4v1rlSJM cdajcRmvuwdcuz2TJZWOhHsS0eRFNBptuyItv2u56pXus0pAGeGyhs90OM7HfjfwfMau Q5alnM+AKTRJK0A3pnXv/hLJLrQwXUQpD4xLeYdp8CjDwYY60gCEDhzYyOEJpRxHMQhK 6dzT+mwkySJahyijsFoK+thftX+mD1l3bA3HEuu4ylLUgAHwylq2jw4YqGMjl768+Vyr hg5w== X-Forwarded-Encrypted: i=1; AFNElJ8qXmiEqb1/JZONdyT4h2ix5Y5JCFN1FnYlxi52kcIalkjijmE4x5zx3VWsZ8xJhwCEd2lSnFE/hUTTiUE=@vger.kernel.org X-Gm-Message-State: AOJu0Yz2nmbF7RQZLLgRWieJHUYl/5tm+gH/8uka8+gveETxA6a2RycW BL29OXDoGik38ZWB4QcnkpusfhAZ1U0XbN8aRO/9Ay/Yja33NKPLfxVwmg4lfanw1lo+6ntW8cR dlsmBNQ== X-Received: from pllk1.prod.google.com ([2002:a17:902:7601:b0:2c1:9d1:eeb0]) (user=seanjc job=prod-delivery.src-stubby-dispatcher) by 2002:a17:902:d592:b0:2c0:f807:56b2 with SMTP id d9443c01a7336-2c1644ca2f2mr54951835ad.34.1780529355634; Wed, 03 Jun 2026 16:29:15 -0700 (PDT) Date: Wed, 3 Jun 2026 16:29:14 -0700 In-Reply-To: Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20251113232229.1698886-1-seanjc@google.com> Message-ID: Subject: Re: [PATCH] KVM: guest_memfd: Elaborate on how release() vs. get_pfn() is safe against UAF From: Sean Christopherson To: Ackerley Tng Cc: Paolo Bonzini , kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Yan Zhao , Vishal Annapurve Content-Type: text/plain; charset="us-ascii" On Wed, Jun 03, 2026, Ackerley Tng wrote: > Sean Christopherson writes: > > > Add more context and information to the comment in kvm_gmem_release() that > > explains why there's no synchronization on RCU _or_ kvm->srcu. Point (b) > > from commit 67b43038ce14 ("KVM: guest_memfd: Remove RCU-protected attribute > > from slot->gmem.file") > > > > b) kvm->srcu ensures that kvm_gmem_unbind() and freeing of a memslot > > occur after the memslot is no longer visible to kvm_gmem_get_pfn(). > > > > is especially difficult to fully grok, particularly in light of commit > > ae431059e75d ("KVM: guest_memfd: Remove bindings on memslot deletion when > > gmem is dying"), which addressed a race between unbind() and release(). > > > > No functional change intended. > > > > Cc: Yan Zhao > > Cc: Vishal Annapurve > > Signed-off-by: Sean Christopherson > > --- > > virt/kvm/guest_memfd.c | 20 ++++++++++++++------ > > 1 file changed, 14 insertions(+), 6 deletions(-) > > > > diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c > > index fdaea3422c30..2e09d7ec0cfc 100644 > > --- a/virt/kvm/guest_memfd.c > > +++ b/virt/kvm/guest_memfd.c > > @@ -338,17 +338,25 @@ static int kvm_gmem_release(struct inode *inode, struct file *file) > > * dereferencing the slot for existing bindings needs to be protected > > * against memslot updates, specifically so that unbind doesn't race > > * and free the memslot (kvm_gmem_get_file() will return NULL). > > - * > > - * Since .release is called only when the reference count is zero, > > - * after which file_ref_get() and get_file_active() fail, > > - * kvm_gmem_get_pfn() cannot be using the file concurrently. > > - * file_ref_put() provides a full barrier, and get_file_active() the > > - * matching acquire barrier. > > */ > > mutex_lock(&kvm->slots_lock); > > > > filemap_invalidate_lock(inode->i_mapping); > > > > + /* > > + * Note! synchronize_srcu() is _not_ needed after nullifying memslot > > + * bindings as slot->gmem.file cannot be set back to a non-null value > > + * without the memslot first being deleted. I.e. this relies on the > > + * synchronize_srcu_expedited() in kvm_swap_active_memslots() to ensure > > Is this kvm_swap_active_memslots() referring not to swapping out of the > deleted memslot, but swapping in of a new memslot? No, it's the deletion that matters. When doing an (S)RCU-protect pointer swap, readers can see _either_ pointer (the old or the new). I.e. when creating a new memslot, synchronize_srcu_expedited() would only guarantee readers would see the new slot *after* synchronization completes. But that's not what we care about. What we care about is that an in-flight kvm_gmem_get_pfn() _can't_ reach the file for new, re-allocated memslot. That's guaranteed because deleting the old memslot waiting for readers to away, i.e. guarantees that any readers that saw a non-NULL slot->gmem.file saw the one associated with the old memslot. > > + * kvm_gmem_get_pfn() (which runs with kvm->srcu held for read) can't > > + * grab a reference to slot->gmem.file even if the struct file object > > + * is reallocated. > > + * > > I think the part that's missing for me is: why is synchronize_srcu() > required after NULLifying memslot bindings? Normally, synchronize_srcu() > will be needed after memslot updates to make sure the next user of > kvm->srcu gets the new state of the memslots? > > The "new state" here could be > > 1. slot->gmem.file == NULL > 2. slot->gmem.file == some new file, set after the old memslot was > replaced Only because there's a synchronize_srcu(). Without that, it could be: 1. slot->gmem.file == old == 2. slot->gmem.file == NULL |==> window for readers 3. slot->gmem.file == new == versus 1. slot->gmem.file == old == synchronize_srcu() |===> window for readers 2. slot->gmem.file == NULL == synchronize_srcu() |===> window #2 for readers 3. slot->gmem.file == new ==