public inbox for u-boot@lists.denx.de
 help / color / mirror / Atom feed
* [PATCH] efi_loader: fix use after free in efi_exit() with tcg2
@ 2026-01-23 10:58 Vincent Stehlé
  2026-01-27  7:58 ` Ilias Apalodimas
                   ` (2 more replies)
  0 siblings, 3 replies; 7+ messages in thread
From: Vincent Stehlé @ 2026-01-23 10:58 UTC (permalink / raw)
  To: u-boot
  Cc: Vincent Stehlé, Heinrich Schuchardt, Ilias Apalodimas,
	Tom Rini, Masahisa Kojima

The efi_exit() function frees the loaded image memory by calling
efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is enabled, the
image_obj->image_type structure member is accessed after the memory has
been freed.

Fix this by keeping a copy of image_type, as is already done for exit_jmp.

Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Cc: Tom Rini <trini@konsulko.com>
Cc: Masahisa Kojima <masahisa.kojima@linaro.org>
---

Hi,

This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
following command:

  valgrind --suppressions=scripts/u-boot.supp \
    ./u-boot -T -c "setenv efi_selftest start image return; \
                    bootefi selftest"

Best regards,
Vincent.

 lib/efi_loader/efi_boottime.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
index ddc935d2240..0b3f2fd276c 100644
--- a/lib/efi_loader/efi_boottime.c
+++ b/lib/efi_loader/efi_boottime.c
@@ -3453,6 +3453,7 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
 	struct efi_loaded_image_obj *image_obj =
 		(struct efi_loaded_image_obj *)image_handle;
 	jmp_buf *exit_jmp;
+	u16 image_type;
 
 	EFI_ENTRY("%p, %ld, %zu, %p", image_handle, exit_status,
 		  exit_data_size, exit_data);
@@ -3496,13 +3497,14 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
 	}
 	/* efi_delete_image() frees image_obj. Copy before the call. */
 	exit_jmp = image_obj->exit_jmp;
+	image_type = image_obj->image_type;
 	*image_obj->exit_status = exit_status;
-	if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
+	if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
 	    exit_status != EFI_SUCCESS)
 		efi_delete_image(image_obj, loaded_image_protocol);
 
 	if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
-		if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
+		if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
 			ret = efi_tcg2_measure_efi_app_exit();
 			if (ret != EFI_SUCCESS)
 				log_debug("tcg2 measurement fails (0x%lx)\n",
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* Re: [PATCH] efi_loader: fix use after free in efi_exit() with tcg2
  2026-01-23 10:58 [PATCH] efi_loader: fix use after free in efi_exit() with tcg2 Vincent Stehlé
@ 2026-01-27  7:58 ` Ilias Apalodimas
  2026-01-27 10:37   ` Vincent Stehlé
  2026-01-27 16:18 ` [PATCH v2] " Vincent Stehlé
  2026-01-30  7:25 ` [PATCH] " Ilias Apalodimas
  2 siblings, 1 reply; 7+ messages in thread
From: Ilias Apalodimas @ 2026-01-27  7:58 UTC (permalink / raw)
  To: Vincent Stehlé
  Cc: u-boot, Heinrich Schuchardt, Tom Rini, Masahisa Kojima

Hi Vincent,

On Fri, 23 Jan 2026 at 14:58, Vincent Stehlé <vincent.stehle@arm.com> wrote:
>
> The efi_exit() function frees the loaded image memory by calling
> efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is enabled, the
> image_obj->image_type structure member is accessed after the memory has
> been freed.
>
> Fix this by keeping a copy of image_type, as is already done for exit_jmp.
>
> Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
> Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
> Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
> Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Cc: Tom Rini <trini@konsulko.com>
> Cc: Masahisa Kojima <masahisa.kojima@linaro.org>
> ---
>
> Hi,
>
> This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
> following command:
>
>   valgrind --suppressions=scripts/u-boot.supp \
>     ./u-boot -T -c "setenv efi_selftest start image return; \
>                     bootefi selftest"
>
> Best regards,
> Vincent.
>
>  lib/efi_loader/efi_boottime.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
> index ddc935d2240..0b3f2fd276c 100644
> --- a/lib/efi_loader/efi_boottime.c
> +++ b/lib/efi_loader/efi_boottime.c
> @@ -3453,6 +3453,7 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
>         struct efi_loaded_image_obj *image_obj =
>                 (struct efi_loaded_image_obj *)image_handle;
>         jmp_buf *exit_jmp;
> +       u16 image_type;
>
>         EFI_ENTRY("%p, %ld, %zu, %p", image_handle, exit_status,
>                   exit_data_size, exit_data);
> @@ -3496,13 +3497,14 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
>         }
>         /* efi_delete_image() frees image_obj. Copy before the call. */
>         exit_jmp = image_obj->exit_jmp;
> +       image_type = image_obj->image_type;
>         *image_obj->exit_status = exit_status;
> -       if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> +       if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
>             exit_status != EFI_SUCCESS)
>                 efi_delete_image(image_obj, loaded_image_protocol);
>
>         if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
> -               if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
> +               if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {

This looks ok, but I don't remember the TCG spec on top of my head.
Can't we just move the measurement before deleting the handle?

Thanks
/Ilias
>                         ret = efi_tcg2_measure_efi_app_exit();
>                         if (ret != EFI_SUCCESS)
>                                 log_debug("tcg2 measurement fails (0x%lx)\n",
> --
> 2.51.0
>

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] efi_loader: fix use after free in efi_exit() with tcg2
  2026-01-27  7:58 ` Ilias Apalodimas
@ 2026-01-27 10:37   ` Vincent Stehlé
  0 siblings, 0 replies; 7+ messages in thread
From: Vincent Stehlé @ 2026-01-27 10:37 UTC (permalink / raw)
  To: Ilias Apalodimas; +Cc: u-boot, Heinrich Schuchardt, Tom Rini, Masahisa Kojima

(Updated Kojima-san's e-mail address.)

On Tue, Jan 27, 2026 at 09:58:33AM +0200, Ilias Apalodimas wrote:
> Hi Vincent,

Hi Ilias,

Thank you for having a look!
My comments below.

> On Fri, 23 Jan 2026 at 14:58, Vincent Stehlé <vincent.stehle@arm.com> wrote:
> >
> > The efi_exit() function frees the loaded image memory by calling
> > efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is enabled, the
> > image_obj->image_type structure member is accessed after the memory has
> > been freed.
> >
> > Fix this by keeping a copy of image_type, as is already done for exit_jmp.
> >
> > Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
> > Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
> > Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
> > Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> > Cc: Tom Rini <trini@konsulko.com>
> > Cc: Masahisa Kojima <masahisa.kojima@linaro.org>
> > ---
> >
> > Hi,
> >
> > This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
> > following command:
> >
> >   valgrind --suppressions=scripts/u-boot.supp \
> >     ./u-boot -T -c "setenv efi_selftest start image return; \
> >                     bootefi selftest"
> >
> > Best regards,
> > Vincent.
> >
> >  lib/efi_loader/efi_boottime.c | 6 ++++--
> >  1 file changed, 4 insertions(+), 2 deletions(-)
> >
> > diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
> > index ddc935d2240..0b3f2fd276c 100644
> > --- a/lib/efi_loader/efi_boottime.c
> > +++ b/lib/efi_loader/efi_boottime.c
> > @@ -3453,6 +3453,7 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
> >         struct efi_loaded_image_obj *image_obj =
> >                 (struct efi_loaded_image_obj *)image_handle;
> >         jmp_buf *exit_jmp;
> > +       u16 image_type;
> >
> >         EFI_ENTRY("%p, %ld, %zu, %p", image_handle, exit_status,
> >                   exit_data_size, exit_data);
> > @@ -3496,13 +3497,14 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
> >         }
> >         /* efi_delete_image() frees image_obj. Copy before the call. */
> >         exit_jmp = image_obj->exit_jmp;
> > +       image_type = image_obj->image_type;
> >         *image_obj->exit_status = exit_status;
> > -       if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> > +       if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> >             exit_status != EFI_SUCCESS)
> >                 efi_delete_image(image_obj, loaded_image_protocol);
> >
> >         if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
> > -               if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
> > +               if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
> 
> This looks ok, but I don't remember the TCG spec on top of my head.
> Can't we just move the measurement before deleting the handle?

I think this can work indeed, and it is more elegant; I will send a v2, thanks!

If I read the specification[1] correctly, the measurements sequence in 8.2.4 is
the following:

  f. Execute UEFI application
     i. If additional loads
          measure
  g. Measure EV_EFI_ACTION "Returning ..."
  h. If select next
       goto a
  i. If ExitBootServices
       measure EV_EFI_ACTION "Exit Boot Services ..."

Everything efi_exit() is doing is happening in step "g", therefore moving code
around efi_delete_image() should have no effect on the measurements.

Best regards,
Vincent.

[1] https://trustedcomputinggroup.org/wp-content/uploads/PC-Client-Platform-Firmware-Profile-Version-1.06-Revision-52_pub.pdf

> 
> Thanks
> /Ilias
> >                         ret = efi_tcg2_measure_efi_app_exit();
> >                         if (ret != EFI_SUCCESS)
> >                                 log_debug("tcg2 measurement fails (0x%lx)\n",
> > --
> > 2.51.0
> >

^ permalink raw reply	[flat|nested] 7+ messages in thread

* [PATCH v2] efi_loader: fix use after free in efi_exit() with tcg2
  2026-01-23 10:58 [PATCH] efi_loader: fix use after free in efi_exit() with tcg2 Vincent Stehlé
  2026-01-27  7:58 ` Ilias Apalodimas
@ 2026-01-27 16:18 ` Vincent Stehlé
  2026-01-27 23:43   ` kojima.masahisa
  2026-02-06  8:34   ` Heinrich Schuchardt
  2026-01-30  7:25 ` [PATCH] " Ilias Apalodimas
  2 siblings, 2 replies; 7+ messages in thread
From: Vincent Stehlé @ 2026-01-27 16:18 UTC (permalink / raw)
  To: u-boot
  Cc: Vincent Stehlé, Ilias Apalodimas, Heinrich Schuchardt,
	Tom Rini, Masahisa Kojima

The efi_exit() function frees the loaded image memory by calling
efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is enabled, the
image_obj->image_type structure member is accessed after the memory has
been freed.

Fix this by performing the tcg2 measurement before the image deletion.

Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
Suggested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Tom Rini <trini@konsulko.com>
Cc: Masahisa Kojima <kojima.masahisa@socionext.com>
---

Hi,

Here is a respin after feedbacks. [1]

Changes for v2:
- Move the event measurement before image deletion instead of keeping a
  copy of image_type (thanks Ilias!)

This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
following command:

  valgrind --suppressions=scripts/u-boot.supp \
    ./u-boot -T -c "setenv efi_selftest start image return; \
                    bootefi selftest"

This was lightly tested for regression with sandbox_defconfig and the
following commands:

  ./u-boot -T -c "ut measurement"

  ./test/py/test.py --build-dir="$PWD" -s -k "test_efi_bootmgr \
	or test_efi_loader or test_efi_selftest or test_efi_secboot"

Adding some instrumentation in efi_exit() and tcg2_log_append() shows no
change in the event measurements sequence.

[1] https://lore.kernel.org/u-boot/20260123105814.1083834-1-vincent.stehle@arm.com/

 lib/efi_loader/efi_boottime.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
index ddc935d2240..b424d924896 100644
--- a/lib/efi_loader/efi_boottime.c
+++ b/lib/efi_loader/efi_boottime.c
@@ -3494,12 +3494,6 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
 		if (ret != EFI_SUCCESS)
 			EFI_PRINT("%s: out of memory\n", __func__);
 	}
-	/* efi_delete_image() frees image_obj. Copy before the call. */
-	exit_jmp = image_obj->exit_jmp;
-	*image_obj->exit_status = exit_status;
-	if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
-	    exit_status != EFI_SUCCESS)
-		efi_delete_image(image_obj, loaded_image_protocol);
 
 	if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
 		if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
@@ -3510,6 +3504,13 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
 		}
 	}
 
+	/* efi_delete_image() frees image_obj. Copy before the call. */
+	exit_jmp = image_obj->exit_jmp;
+	*image_obj->exit_status = exit_status;
+	if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
+	    exit_status != EFI_SUCCESS)
+		efi_delete_image(image_obj, loaded_image_protocol);
+
 	/* Make sure entry/exit counts for EFI world cross-overs match */
 	EFI_EXIT(exit_status);
 
-- 
2.51.0


^ permalink raw reply related	[flat|nested] 7+ messages in thread

* RE: [PATCH v2] efi_loader: fix use after free in efi_exit() with tcg2
  2026-01-27 16:18 ` [PATCH v2] " Vincent Stehlé
@ 2026-01-27 23:43   ` kojima.masahisa
  2026-02-06  8:34   ` Heinrich Schuchardt
  1 sibling, 0 replies; 7+ messages in thread
From: kojima.masahisa @ 2026-01-27 23:43 UTC (permalink / raw)
  To: vincent.stehle, u-boot; +Cc: ilias.apalodimas, xypron.glpk, trini

Hi Vincent,

> -----Original Message-----
> From: Vincent Stehlé <vincent.stehle@arm.com>
> Sent: Wednesday, January 28, 2026 1:19 AM
> Subject: [PATCH v2] efi_loader: fix use after free in efi_exit() with tcg2
> 
> The efi_exit() function frees the loaded image memory by calling
> efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is
> enabled, the
> image_obj->image_type structure member is accessed after the memory has
> been freed.
> 
> Fix this by performing the tcg2 measurement before the image deletion.
> 
> Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
> Suggested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
> Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
> Cc: Tom Rini <trini@konsulko.com>
> Cc: Masahisa Kojima <kojima.masahisa@socionext.com>
> ---
> 
> Hi,
> 
> Here is a respin after feedbacks. [1]
> 
> Changes for v2:
> - Move the event measurement before image deletion instead of keeping a
>   copy of image_type (thanks Ilias!)
> 
> This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
> following command:
> 
>   valgrind --suppressions=scripts/u-boot.supp \
>     ./u-boot -T -c "setenv efi_selftest start image return; \
>                     bootefi selftest"
> 
> This was lightly tested for regression with sandbox_defconfig and the
> following commands:
> 
>   ./u-boot -T -c "ut measurement"
> 
>   ./test/py/test.py --build-dir="$PWD" -s -k "test_efi_bootmgr \
> 	or test_efi_loader or test_efi_selftest or test_efi_secboot"
> 
> Adding some instrumentation in efi_exit() and tcg2_log_append() shows no
> change in the event measurements sequence.
> 
> [1]
> https://lore.kernel.org/u-boot/20260123105814.1083834-1-vincent.stehle@ar
> m.com/
> 
>  lib/efi_loader/efi_boottime.c | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
> index ddc935d2240..b424d924896 100644
> --- a/lib/efi_loader/efi_boottime.c
> +++ b/lib/efi_loader/efi_boottime.c
> @@ -3494,12 +3494,6 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t
> image_handle,
>  		if (ret != EFI_SUCCESS)
>  			EFI_PRINT("%s: out of memory\n", __func__);
>  	}
> -	/* efi_delete_image() frees image_obj. Copy before the call. */
> -	exit_jmp = image_obj->exit_jmp;
> -	*image_obj->exit_status = exit_status;
> -	if (image_obj->image_type ==
> IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> -	    exit_status != EFI_SUCCESS)
> -		efi_delete_image(image_obj, loaded_image_protocol);
> 
>  	if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
>  		if (image_obj->image_type ==
> IMAGE_SUBSYSTEM_EFI_APPLICATION) {
> @@ -3510,6 +3504,13 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t
> image_handle,
>  		}
>  	}
> 
> +	/* efi_delete_image() frees image_obj. Copy before the call. */
> +	exit_jmp = image_obj->exit_jmp;
> +	*image_obj->exit_status = exit_status;
> +	if (image_obj->image_type ==
> IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> +	    exit_status != EFI_SUCCESS)
> +		efi_delete_image(image_obj, loaded_image_protocol);
> +
>  	/* Make sure entry/exit counts for EFI world cross-overs match */
>  	EFI_EXIT(exit_status);

Thanks for the fix.
Acked-by: Masahisa Kojima <kojima.masahisa@socionext.com>

Best Regards,
Masahisa Kojima

> 
> --
> 2.51.0


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH] efi_loader: fix use after free in efi_exit() with tcg2
  2026-01-23 10:58 [PATCH] efi_loader: fix use after free in efi_exit() with tcg2 Vincent Stehlé
  2026-01-27  7:58 ` Ilias Apalodimas
  2026-01-27 16:18 ` [PATCH v2] " Vincent Stehlé
@ 2026-01-30  7:25 ` Ilias Apalodimas
  2 siblings, 0 replies; 7+ messages in thread
From: Ilias Apalodimas @ 2026-01-30  7:25 UTC (permalink / raw)
  To: Vincent Stehlé
  Cc: u-boot, Heinrich Schuchardt, Tom Rini, Masahisa Kojima

Thanks Vincent

I also tested that PCR4 remains unchanged after the patch.

On Fri, 23 Jan 2026 at 14:58, Vincent Stehlé <vincent.stehle@arm.com> wrote:
>
> The efi_exit() function frees the loaded image memory by calling
> efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is enabled, the
> image_obj->image_type structure member is accessed after the memory has
> been freed.
>
> Fix this by keeping a copy of image_type, as is already done for exit_jmp.
>
> Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
> Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
> Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
> Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Cc: Tom Rini <trini@konsulko.com>
> Cc: Masahisa Kojima <masahisa.kojima@linaro.org>
> ---

Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Tested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>

>
> Hi,
>
> This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
> following command:
>
>   valgrind --suppressions=scripts/u-boot.supp \
>     ./u-boot -T -c "setenv efi_selftest start image return; \
>                     bootefi selftest"
>
> Best regards,
> Vincent.
>
>  lib/efi_loader/efi_boottime.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
> index ddc935d2240..0b3f2fd276c 100644
> --- a/lib/efi_loader/efi_boottime.c
> +++ b/lib/efi_loader/efi_boottime.c
> @@ -3453,6 +3453,7 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
>         struct efi_loaded_image_obj *image_obj =
>                 (struct efi_loaded_image_obj *)image_handle;
>         jmp_buf *exit_jmp;
> +       u16 image_type;
>
>         EFI_ENTRY("%p, %ld, %zu, %p", image_handle, exit_status,
>                   exit_data_size, exit_data);
> @@ -3496,13 +3497,14 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
>         }
>         /* efi_delete_image() frees image_obj. Copy before the call. */
>         exit_jmp = image_obj->exit_jmp;
> +       image_type = image_obj->image_type;
>         *image_obj->exit_status = exit_status;
> -       if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> +       if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
>             exit_status != EFI_SUCCESS)
>                 efi_delete_image(image_obj, loaded_image_protocol);
>
>         if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
> -               if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
> +               if (image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
>                         ret = efi_tcg2_measure_efi_app_exit();
>                         if (ret != EFI_SUCCESS)
>                                 log_debug("tcg2 measurement fails (0x%lx)\n",
> --
> 2.51.0
>

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [PATCH v2] efi_loader: fix use after free in efi_exit() with tcg2
  2026-01-27 16:18 ` [PATCH v2] " Vincent Stehlé
  2026-01-27 23:43   ` kojima.masahisa
@ 2026-02-06  8:34   ` Heinrich Schuchardt
  1 sibling, 0 replies; 7+ messages in thread
From: Heinrich Schuchardt @ 2026-02-06  8:34 UTC (permalink / raw)
  To: Vincent Stehlé, u-boot; +Cc: Ilias Apalodimas, Tom Rini, Masahisa Kojima

On 1/27/26 17:18, Vincent Stehlé wrote:
> The efi_exit() function frees the loaded image memory by calling
> efi_delete_image(). However, when CONFIG_EFI_TCG2_PROTOCOL is enabled, the
> image_obj->image_type structure member is accessed after the memory has
> been freed.
> 
> Fix this by performing the tcg2 measurement before the image deletion.
> 
> Fixes: 8fc4e0b4273a ("efi_loader: add boot variable measurement")
> Suggested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
> Signed-off-by: Vincent Stehlé <vincent.stehle@arm.com>
> Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
> Cc: Tom Rini <trini@konsulko.com>
> Cc: Masahisa Kojima <kojima.masahisa@socionext.com>
> ---

Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>

> 
> Hi,
> 
> Here is a respin after feedbacks. [1]
> 
> Changes for v2:
> - Move the event measurement before image deletion instead of keeping a
>    copy of image_type (thanks Ilias!)
> 
> This can be verified with sandbox_defconfig + CONFIG_VALGRIND=y and the
> following command:
> 
>    valgrind --suppressions=scripts/u-boot.supp \
>      ./u-boot -T -c "setenv efi_selftest start image return; \
>                      bootefi selftest"
> 
> This was lightly tested for regression with sandbox_defconfig and the
> following commands:
> 
>    ./u-boot -T -c "ut measurement"
> 
>    ./test/py/test.py --build-dir="$PWD" -s -k "test_efi_bootmgr \
> 	or test_efi_loader or test_efi_selftest or test_efi_secboot"
> 
> Adding some instrumentation in efi_exit() and tcg2_log_append() shows no
> change in the event measurements sequence.
> 
> [1] https://lore.kernel.org/u-boot/20260123105814.1083834-1-vincent.stehle@arm.com/
> 
>   lib/efi_loader/efi_boottime.c | 13 +++++++------
>   1 file changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/lib/efi_loader/efi_boottime.c b/lib/efi_loader/efi_boottime.c
> index ddc935d2240..b424d924896 100644
> --- a/lib/efi_loader/efi_boottime.c
> +++ b/lib/efi_loader/efi_boottime.c
> @@ -3494,12 +3494,6 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
>   		if (ret != EFI_SUCCESS)
>   			EFI_PRINT("%s: out of memory\n", __func__);
>   	}
> -	/* efi_delete_image() frees image_obj. Copy before the call. */
> -	exit_jmp = image_obj->exit_jmp;
> -	*image_obj->exit_status = exit_status;
> -	if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> -	    exit_status != EFI_SUCCESS)
> -		efi_delete_image(image_obj, loaded_image_protocol);
>   
>   	if (IS_ENABLED(CONFIG_EFI_TCG2_PROTOCOL)) {
>   		if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION) {
> @@ -3510,6 +3504,13 @@ static efi_status_t EFIAPI efi_exit(efi_handle_t image_handle,
>   		}
>   	}
>   
> +	/* efi_delete_image() frees image_obj. Copy before the call. */
> +	exit_jmp = image_obj->exit_jmp;
> +	*image_obj->exit_status = exit_status;
> +	if (image_obj->image_type == IMAGE_SUBSYSTEM_EFI_APPLICATION ||
> +	    exit_status != EFI_SUCCESS)
> +		efi_delete_image(image_obj, loaded_image_protocol);
> +
>   	/* Make sure entry/exit counts for EFI world cross-overs match */
>   	EFI_EXIT(exit_status);
>   


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2026-02-06  8:34 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-01-23 10:58 [PATCH] efi_loader: fix use after free in efi_exit() with tcg2 Vincent Stehlé
2026-01-27  7:58 ` Ilias Apalodimas
2026-01-27 10:37   ` Vincent Stehlé
2026-01-27 16:18 ` [PATCH v2] " Vincent Stehlé
2026-01-27 23:43   ` kojima.masahisa
2026-02-06  8:34   ` Heinrich Schuchardt
2026-01-30  7:25 ` [PATCH] " Ilias Apalodimas

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox