From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-qt1-f181.google.com (mail-qt1-f181.google.com [209.85.160.181]) (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 858772BCFB for ; Mon, 28 Jul 2025 17:17:12 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.160.181 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1753723034; cv=none; b=asn7ONhVuBOlAfyK3s49cISTYfzzgzTeMPOvBtu9mjpLQDPeWcQ+U9I/qRt26c43O5FDUNFdgnQcz2VOUQc8yUZTJgOu6OlLpa692figYu50FW73LPomR4v2+Toc4lGy0hIZXa//IClCjtYGVOqpDvW+49BjYsY+LSta5k5+Ypk= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1753723034; c=relaxed/simple; bh=dCGL9BqXQ+3CzZDvRjNFmap92b/+aEqgRPHbjk3YeS0=; h=MIME-Version:References:In-Reply-To:From:Date:Message-ID:Subject: To:Cc:Content-Type; b=LEHe4oLDmxlaYhE3O8eUoCY5V7xyUGG+NyGoRxNAIuf7cChhMJ3TQkGl0kPUjH64COGhlhXCWttD8IAVZkrXgbkugGKp/PaDWKV2CDHX79jAwWO1yukMVKtW8/rWeFTsUT3KHDz54ugf4KjlajCM5ixZtnteRRSGnd6RWUj1TC4= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=FiURPs8k; arc=none smtp.client-ip=209.85.160.181 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=google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="FiURPs8k" Received: by mail-qt1-f181.google.com with SMTP id d75a77b69052e-4ab86a29c98so17781cf.0 for ; Mon, 28 Jul 2025 10:17:12 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1753723031; x=1754327831; darn=vger.kernel.org; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:from:to:cc:subject:date :message-id:reply-to; bh=2vKuqsMcT46hBPIeeydmoWyK22Wv2QkCzgzDxJHhBAg=; b=FiURPs8ktf2WhT4S8/DpNal/kWtF1Qbkm+d9DM7yjflLS14v/F44dJQXoJ2u02MLnD /iXLO98nQRNVxfrE/lbvZRaUscaQHvOu/riO8L+m/kzBF81d3qF7ZEQeDqrZJ449TRAz PsANhdyTabCqRIN9Yp3NNmlbTEiBrtdua7kogwofq7gjlPdPIYZnkDEojj5JbmtsMEfT fbJLOK1aKlblQbmZ/zuvRSnQIn7YHv8ghkbnAvpb0qUDJT8atM7hRHLWMP1WHMQxp1vd JBBZTCljY7K8enhbWR1dCUuKs9lxxZEIjOXYdMlMFb+aMMA+pCnaPtKzOqBXXvcqq1ZF s+mw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1753723031; x=1754327831; h=content-transfer-encoding:cc:to:subject:message-id:date:from :in-reply-to:references:mime-version:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=2vKuqsMcT46hBPIeeydmoWyK22Wv2QkCzgzDxJHhBAg=; b=l3oo/q/+ZT8IqmTGxO/usT8Xim64xSjgCUnwJRmt2/IstdAEipml5opkw5fHiGRQse /lowQvCBK8VFZNTpqfgvlERnYULa83Z1atbARXieZFW1cITu+3MUW666G6N1eaPFAeV0 Uce526r51Igb2haTZKzpQ83c3AD8oOJXWlTq0zuPNVGvdXSjPfhYJPaNBV5RXb72pF3o ezaXTh7raeMx44FKCM3pkQQ7EzbRNXyFUz/SSgRx95nFzc1kfvXSdtXnyRO7Wt1vQOnC M0TVQtsJiyInw0rQT3bzExW5Ly7euq+masd0S1mEhpqCH1Cc0TS96n0cTCosJvdzQi3K t7jw== X-Forwarded-Encrypted: i=1; AJvYcCX6frDp+aaRXo1qPY+OI+naz+qm52sjgnkOr1vMM4vdJrmwENr1BQ4wI9yX4SC0bae8GX7mjhWSToJhhjc=@vger.kernel.org X-Gm-Message-State: AOJu0Yy4ZYI4ppvQ05mgH5dnQU2T5t+2OlEkG46NAHbeXNLr9O23h+0E 3hh5+SZtq+OlrSOs/xCG1XSxTfelOSzUdMmmD37zFWCsY7MXLQdFzhzi8wfFzuGYXX64UlT92vs rmE+LKlUd+du7BSEpjwW/KBNNt1QWlVv7/zM9UYN2 X-Gm-Gg: ASbGncvqlnKFVgFTYPnk/JLuJyPulDIrgLdw40CPhJvhAit5dY1F/26Pv/kJMSK3XxI /cG2iWPkITo5vyhrOBGA+uI5AQM7yr8X6TwBU4yd0JbKMv/DDyM9+2QI7+cQnSH5BKtadnr+ihp zei+5ipQmPLivleZRTuuafBCeP06nw14UPi2pu7B/9SwdzZxgZAbk1Y+kqz2F6WMUQCNjt2zdrZ mMa5dpQkBcF4zl5E5DoKPfrCzQtWURWNMeuoPEXQQMnm7Av X-Google-Smtp-Source: AGHT+IELxZqzgYWAQP9Ngohy4ApnkBo9oFMQtzvoZOAJ7yizj5sLIqWpkPtI4MboAEZl0uq5UgdDsHMj4GTXCTR/jaM= X-Received: by 2002:a05:622a:148c:b0:4a7:e3b:50be with SMTP id d75a77b69052e-4ae9cc0b2f3mr8431311cf.16.1753723030772; Mon, 28 Jul 2025 10:17:10 -0700 (PDT) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 References: <20250728170950.2216966-1-surenb@google.com> In-Reply-To: <20250728170950.2216966-1-surenb@google.com> From: Suren Baghdasaryan Date: Mon, 28 Jul 2025 10:16:59 -0700 X-Gm-Features: Ac12FXx5Ajyiy-dotDY0A444StrFRdZUfUHPwsQH_l1PjmWS7zmm2-g_4bnZRTc Message-ID: Subject: Re: [PATCH 1/1] mm: fix a UAF when vma->mm is freed after vma->vm_refcnt got dropped To: akpm@linux-foundation.org Cc: jannh@google.com, Liam.Howlett@oracle.com, lorenzo.stoakes@oracle.com, vbabka@suse.cz, pfalcato@suse.de, linux-mm@kvack.org, linux-kernel@vger.kernel.org, stable@vger.kernel.org Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Mon, Jul 28, 2025 at 10:09=E2=80=AFAM Suren Baghdasaryan wrote: > > By inducing delays in the right places, Jann Horn created a reproducer > for a hard to hit UAF issue that became possible after VMAs were allowed > to be recycled by adding SLAB_TYPESAFE_BY_RCU to their cache. > > Race description is borrowed from Jann's discovery report: > lock_vma_under_rcu() looks up a VMA locklessly with mas_walk() under > rcu_read_lock(). At that point, the VMA may be concurrently freed, and > it can be recycled by another process. vma_start_read() then > increments the vma->vm_refcnt (if it is in an acceptable range), and > if this succeeds, vma_start_read() can return a recycled VMA. > > In this scenario where the VMA has been recycled, lock_vma_under_rcu() > will then detect the mismatching ->vm_mm pointer and drop the VMA > through vma_end_read(), which calls vma_refcount_put(). > vma_refcount_put() drops the refcount and then calls rcuwait_wake_up() > using a copy of vma->vm_mm. This is wrong: It implicitly assumes that > the caller is keeping the VMA's mm alive, but in this scenario the caller > has no relation to the VMA's mm, so the rcuwait_wake_up() can cause UAF. > > The diagram depicting the race: > T1 T2 T3 > =3D=3D =3D=3D =3D=3D > lock_vma_under_rcu > mas_walk > > mmap > > vma_start_read > __refcount_inc_not_zero_limited_acquire > munmap > __vma_enter_locked > refcount_add_not_zero > vma_end_read > vma_refcount_put > __refcount_dec_and_test > rcuwait_wait_event > > rcuwait_wake_up [UAF] > > Note that rcuwait_wait_event() in T3 does not block because refcount > was already dropped by T1. At this point T3 can exit and free the mm > causing UAF in T1. > To avoid this we move vma->vm_mm verification into vma_start_read() and > grab vma->vm_mm to stabilize it before vma_refcount_put() operation. > > Fixes: 3104138517fc ("mm: make vma cache SLAB_TYPESAFE_BY_RCU") > Reported-by: Jann Horn > Closes: https://lore.kernel.org/all/CAG48ez0-deFbVH=3DE3jbkWx=3DX3uVbd8nW= eo6kbJPQ0KoUD+m2tA@mail.gmail.com/ > Signed-off-by: Suren Baghdasaryan > Cc: > --- > - Applies cleanly over mm-unstable. > - Should be applied to 6.15 and 6.16 but these branches do not > have lock_next_vma() function, so the change in lock_next_vma() should be > skipped when applying to those branches. Andrew, if you would like me to post a separate patch for 6.15 and 6.16 please let me know. The merge conflict in those branches should be trivial: just skip the change in lock_next_vma() which does not exist in those branches. > > include/linux/mmap_lock.h | 21 +++++++++++++++++++++ > mm/mmap_lock.c | 10 +++------- > 2 files changed, 24 insertions(+), 7 deletions(-) > > diff --git a/include/linux/mmap_lock.h b/include/linux/mmap_lock.h > index 1f4f44951abe..4ee4ab835c41 100644 > --- a/include/linux/mmap_lock.h > +++ b/include/linux/mmap_lock.h > @@ -12,6 +12,7 @@ extern int rcuwait_wake_up(struct rcuwait *w); > #include > #include > #include > +#include > > #define MMAP_LOCK_INITIALIZER(name) \ > .mmap_lock =3D __RWSEM_INITIALIZER((name).mmap_lock), > @@ -183,6 +184,26 @@ static inline struct vm_area_struct *vma_start_read(= struct mm_struct *mm, > } > > rwsem_acquire_read(&vma->vmlock_dep_map, 0, 1, _RET_IP_); > + > + /* > + * If vma got attached to another mm from under us, that mm is no= t > + * stable and can be freed in the narrow window after vma->vm_ref= cnt > + * is dropped and before rcuwait_wake_up(mm) is called. Grab it b= efore > + * releasing vma->vm_refcnt. > + */ > + if (unlikely(vma->vm_mm !=3D mm)) { > + /* > + * __mmdrop() is a heavy operation and we don't need RCU > + * protection here. Release RCU lock during these operati= ons. > + */ > + rcu_read_unlock(); > + mmgrab(vma->vm_mm); > + vma_refcount_put(vma); > + mmdrop(vma->vm_mm); > + rcu_read_lock(); > + return NULL; > + } > + > /* > * Overflow of vm_lock_seq/mm_lock_seq might produce false locked= result. > * False unlocked result is impossible because we modify and chec= k > diff --git a/mm/mmap_lock.c b/mm/mmap_lock.c > index 729fb7d0dd59..aa3bc42ecde0 100644 > --- a/mm/mmap_lock.c > +++ b/mm/mmap_lock.c > @@ -164,8 +164,7 @@ struct vm_area_struct *lock_vma_under_rcu(struct mm_s= truct *mm, > */ > > /* Check if the vma we locked is the right one. */ > - if (unlikely(vma->vm_mm !=3D mm || > - address < vma->vm_start || address >=3D vma->vm_end)= ) > + if (unlikely(address < vma->vm_start || address >=3D vma->vm_end)= ) > goto inval_end_read; > > rcu_read_unlock(); > @@ -236,11 +235,8 @@ struct vm_area_struct *lock_next_vma(struct mm_struc= t *mm, > goto fallback; > } > > - /* > - * Verify the vma we locked belongs to the same address space and= it's > - * not behind of the last search position. > - */ > - if (unlikely(vma->vm_mm !=3D mm || from_addr >=3D vma->vm_end)) > + /* Verify the vma is not behind of the last search position. */ > + if (unlikely(from_addr >=3D vma->vm_end)) > goto fallback_unlock; > > /* > > base-commit: c617a4dd7102e691fa0fb2bc4f6b369e37d7f509 > -- > 2.50.1.487.gc89ff58d15-goog >