All of lore.kernel.org
 help / color / mirror / Atom feed
From: Kees Cook <kees@kernel.org>
To: Bill Wendling <morbo@google.com>
Cc: "Christophe Leroy (CS GROUP)" <chleroy@kernel.org>,
	"Gustavo A. R. Silva" <gustavoars@kernel.org>,
	Nathan Chancellor <nathan@kernel.org>,
	Nick Desaulniers <ndesaulniers@google.com>,
	Justin Stitt <justinstitt@google.com>,
	linuxppc-dev@lists.ozlabs.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-hardening@vger.kernel.org,
	llvm@lists.linux.dev, codemender-patching+linux@google.com
Subject: Re: [PATCH] bus: fsl-mc: Annotate fsl_mc_io.portal_virt_addr with __counted_by_ptr
Date: Wed, 23 Sep 2026 01:12:14 -0700	[thread overview]
Message-ID: <202609230106.7960C12@keescook> (raw)
In-Reply-To: <20260922223538.2287676-1-morbo@google.com>

On Tue, Sep 22, 2026 at 10:35:38PM +0000, Bill Wendling wrote:
> The GCC and Clang compilers provide the __counted_by_ptr attribute,
> which is used by KASAN and compiler bounds-checking to detect
> out-of-bounds accesses to pointer fields.
> 
> In "struct fsl_mc_io", the "portal_virt_addr" pointer points to the MC
> command portal virtual address. The size of this allocated portal in
> bytes is tracked by the "portal_size" field within the same structure.
> 
> Annotate the "portal_virt_addr" pointer field with
> "__counted_by_ptr(portal_size)" to enable compiler bounds-checking and
> harden against potential out-of-bounds accesses.

Another one where I hope things agree. :)

        mc_portal_size = resource_size(dpmcp_dev->regions);

        error = fsl_create_mc_io(&mc_bus_dev->dev,
                                 mc_portal_phys_addr,
                                 mc_portal_size, dpmcp_dev,
                                 mc_io_flags, &mc_io);
...

        mc_io->portal_size = mc_portal_size;
	...
        mc_portal_virt_addr = devm_ioremap(dev,
                                                   mc_portal_phys_addr,
                                                   mc_portal_size);
	...
        mc_io->portal_virt_addr = mc_portal_virt_addr;

But it actually reminds me that I still want a warning for having
compile-time warning about pointers being stripped from their counter:

                status = mc_read_response(mc_io->portal_virt_addr, cmd);
...
static inline enum mc_cmd_status mc_read_response(struct fsl_mc_command __iomem
                                                  *portal,
                                                  struct fsl_mc_command *resp)
{
        int i;
        enum mc_cmd_status status;

        /* Copy command response header from MC portal: */
        resp->header = cpu_to_le64(readq_relaxed(&portal->header));

Not only is mc_io->portal_virt_addr separated from mc_io->portal_size
via getting passed to mc_read_response(), but it then immediately gets
cast to struct fsl_mc_command.

We should get the __singleton attribute so we can mark function arg
pointers as "not an array", and then these kinds of casts could generate
a run-time check at function call time to check
sizeof(struct fsl_mc_command) against mc_io->portal_size when
__singleton was there, or kick up a warning that the counter got
stripped.

Because as-is, this patch is a no-op: nothing dereferences
mc_io->portal_virt_addr with the counter in context.


-Kees

-- 
Kees Cook

  parent reply	other threads:[~2026-09-23  8:12 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-22 22:35 [PATCH] bus: fsl-mc: Annotate fsl_mc_io.portal_virt_addr with __counted_by_ptr Bill Wendling
2026-09-23  5:46 ` Gustavo A. R. Silva
2026-09-23  8:12 ` Kees Cook [this message]
2026-10-02 12:55 ` Ioana Ciornei
2026-10-02 13:35 ` Christophe Leroy (CS GROUP)

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=202609230106.7960C12@keescook \
    --to=kees@kernel.org \
    --cc=chleroy@kernel.org \
    --cc=codemender-patching+linux@google.com \
    --cc=gustavoars@kernel.org \
    --cc=justinstitt@google.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-hardening@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linuxppc-dev@lists.ozlabs.org \
    --cc=llvm@lists.linux.dev \
    --cc=morbo@google.com \
    --cc=nathan@kernel.org \
    --cc=ndesaulniers@google.com \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.