linux-kernel - Re: [PATCH 2/5] mm/mseal: move madvise() logic to mm/madvise.c

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <b9bb8416-24a0-4bd2-97c1-055460cece23@lucifer.local>
Date: Mon, 14 Jul 2025 15:56:24 +0100
From: Lorenzo Stoakes <lorenzo.stoakes@...cle.com>
To: David Hildenbrand <david@...hat.com>
Cc: Andrew Morton <akpm@...ux-foundation.org>,
        "Liam R . Howlett" <Liam.Howlett@...cle.com>,
        Vlastimil Babka <vbabka@...e.cz>, Jann Horn <jannh@...gle.com>,
        Pedro Falcato <pfalcato@...e.de>, linux-mm@...ck.org,
        linux-kernel@...r.kernel.org, Jeff Xu <jeffxu@...omium.org>
Subject: Re: [PATCH 2/5] mm/mseal: move madvise() logic to mm/madvise.c

On Mon, Jul 14, 2025 at 04:37:30PM +0200, David Hildenbrand wrote:
> On 14.07.25 15:00, Lorenzo Stoakes wrote:
> > The madvise() logic is inexplicably performed in mm/mseal.c - this ought to
> > be located in mm/madvise.c.
> >
> > Additionally can_modify_vma_madv() is inconsistently named and, in
> > combination with is_ro_anon(), is very confusing logic.
> >
> > Put a static function in mm/madvise.c instead - can_madvise_modify() - that
> > spells out exactly what's happening. Also explicitly check for an anon VMA.
> >
> > Additionally add commentary to explain what's going on.
> >
> > No functional change intended.
> >
> > Signed-off-by: Lorenzo Stoakes <lorenzo.stoakes@...cle.com>
> > ---
> >   mm/madvise.c | 62 +++++++++++++++++++++++++++++++++++++++++++++++++++-
> >   mm/mseal.c   | 49 -----------------------------------------
> >   mm/vma.h     |  7 ------
> >   3 files changed, 61 insertions(+), 57 deletions(-)
> >
> > diff --git a/mm/madvise.c b/mm/madvise.c
> > index 9de9b7c797c6..75757ba418a8 100644
> > --- a/mm/madvise.c
> > +++ b/mm/madvise.c
> > @@ -19,6 +19,7 @@
> >   #include <linux/sched.h>
> >   #include <linux/sched/mm.h>
> >   #include <linux/mm_inline.h>
> > +#include <linux/mmu_context.h>
> >   #include <linux/string.h>
> >   #include <linux/uio.h>
> >   #include <linux/ksm.h>
> > @@ -1256,6 +1257,65 @@ static long madvise_guard_remove(struct madvise_behavior *madv_behavior)
> >   			       &guard_remove_walk_ops, NULL);
> >   }
> >
> > +#ifdef CONFIG_64BIT
>
> It's consistent with mm/Makefile etc. but having a simple
>
> config MSEAL
> 	def_bool y if 64BIT
>
> or sth like that would surely clean that up further.

Well, I plan to make this not a thing soon so I'd rather not.

The intent is to make _all_ VMA flags work on 32-bit kernels. I have done some
preparatory work and next cycle intend to do more on this.

So I'd rather avoid any config changes on this until I've given this a shot.

>
> > +/* Does the madvise operation result in discarding of mapped data? */
> > +static bool is_discard(int behavior)
> > +{
> > +	switch (behavior) {
> > +	case MADV_FREE:
> > +	case MADV_DONTNEED:
> > +	case MADV_DONTNEED_LOCKED:
> > +	case MADV_REMOVE:
> > +	case MADV_DONTFORK:
> > +	case MADV_WIPEONFORK:
> > +	case MADV_GUARD_INSTALL:
> > +		return true;
> > +	}
> > +
> > +	return false;
> > +}
> > +
> > +/*
> > + * We are restricted from madvise()'ing mseal()'d VMAs only in very particular
> > + * circumstances - discarding of data from read-only anonymous SEALED mappings.
> > + *
> > + * This is because users cannot trivally discard data from these VMAs, and may
>
> s/trivally/trivially/

Ack thanks - Andrew can you fixup? Can send a fix-patch otherwise.

>
> > + * only do so via an appropriate madvise() call.
> > + */
> > +static bool can_madvise_modify(struct madvise_behavior *madv_behavior)
> > +{
> > +	struct vm_area_struct *vma = madv_behavior->vma;
> > +
> > +	/* If the operation won't discard, we're good. */
> > +	if (!is_discard(madv_behavior->behavior))
> > +		return true;
>
>
> Conceptually, I would do this first and then handle all the discard cases /
> exceptions.

Hm I'm confused :P we do do this first? I think the idea with this is we can
very cheaply ignore any MADV_ that isn't applicable.

Did you mean to put this comment under line below?

I mean it's not exactly a perf hotspot so don't mind moving them around.

>
> > +
> > +	/* If the VMA isn't sealed we're also good. */
> > +	if (can_modify_vma(vma))
> > +		return true;
> > +> +	/* File-backed mappings don't lose data on discard. */
> > +	if (!vma_is_anonymous(vma))
> > +		return true;
> > +
> > +	/*
> > +	 * If the VMA is writable and the architecture permits write access,
> > +	 * then discarding is fine.
> > +	 */
> > +	if ((vma->vm_flags & VM_WRITE) &&
> > +	    arch_vma_access_permitted(vma, /* write= */ true,
> > +			/* execute= */ false, /* foreign= */ false))
> > +		return true;
> > +
> > +	return false;
> > +}
> > +#else
> > +static bool can_madvise_modify(struct madvise_behavior *madv_behavior)
> > +{
> > +	return true;
> > +}
> > +#endif
> > +
> >   /*
>
> LGTM

Cheers!

>
> --
> Cheers,
>
> David / dhildenb
>