The driver exposes EFI runtime services to user-space through an IOCTL
interface, calling the EFI services function pointers directly without
using the efivar API.
Disallow access to the /dev/efi_test character device when the kernel is
locked down to prevent arbitrary user-space to call EFI runtime services.
Also require CAP_SYS_ADMIN to open the chardev to prevent unprivileged
users to call the EFI runtime services, instead of just relying on the
chardev file mode bits for this.
The main user of this driver is the fwts [0] tool that already checks if
the effective user ID is 0 and fails otherwise. So this change shouldn't
cause any regression to this tool.
[0]: https://wiki.ubuntu.com/FirmwareTestSuite/Reference/uefivarinfo
Signed-off-by: Javier Martinez Canillas <[email protected]>
Acked-by: Laszlo Ersek <[email protected]>
---
Changes in v2:
- Also disable /dev/efi_test access when the kernel is locked down as
suggested by Matthew Garrett.
- Add Acked-by tag from Laszlo Ersek.
drivers/firmware/efi/test/efi_test.c | 8 ++++++++
include/linux/security.h | 1 +
security/lockdown/lockdown.c | 1 +
3 files changed, 10 insertions(+)
diff --git a/drivers/firmware/efi/test/efi_test.c b/drivers/firmware/efi/test/efi_test.c
index 877745c3aaf..7baf48c01e7 100644
--- a/drivers/firmware/efi/test/efi_test.c
+++ b/drivers/firmware/efi/test/efi_test.c
@@ -14,6 +14,7 @@
#include <linux/init.h>
#include <linux/proc_fs.h>
#include <linux/efi.h>
+#include <linux/security.h>
#include <linux/slab.h>
#include <linux/uaccess.h>
@@ -717,6 +718,13 @@ static long efi_test_ioctl(struct file *file, unsigned int cmd,
static int efi_test_open(struct inode *inode, struct file *file)
{
+ int ret = security_locked_down(LOCKDOWN_EFI_TEST);
+
+ if (ret)
+ return ret;
+
+ if (!capable(CAP_SYS_ADMIN))
+ return -EACCES;
/*
* nothing special to do here
* We do accept multiple open files at the same time as we
diff --git a/include/linux/security.h b/include/linux/security.h
index a8d59d612d2..9df7547afc0 100644
--- a/include/linux/security.h
+++ b/include/linux/security.h
@@ -105,6 +105,7 @@ enum lockdown_reason {
LOCKDOWN_NONE,
LOCKDOWN_MODULE_SIGNATURE,
LOCKDOWN_DEV_MEM,
+ LOCKDOWN_EFI_TEST,
LOCKDOWN_KEXEC,
LOCKDOWN_HIBERNATION,
LOCKDOWN_PCI_ACCESS,
diff --git a/security/lockdown/lockdown.c b/security/lockdown/lockdown.c
index 8a10b43daf7..40b790536de 100644
--- a/security/lockdown/lockdown.c
+++ b/security/lockdown/lockdown.c
@@ -20,6 +20,7 @@ static const char *const lockdown_reasons[LOCKDOWN_CONFIDENTIALITY_MAX+1] = {
[LOCKDOWN_NONE] = "none",
[LOCKDOWN_MODULE_SIGNATURE] = "unsigned module loading",
[LOCKDOWN_DEV_MEM] = "/dev/mem,kmem,port",
+ [LOCKDOWN_EFI_TEST] = "/dev/efi_test access",
[LOCKDOWN_KEXEC] = "kexec of unsigned images",
[LOCKDOWN_HIBERNATION] = "hibernation",
[LOCKDOWN_PCI_ACCESS] = "direct PCI access",
--
2.21.0
On 10/08/19 12:55, Javier Martinez Canillas wrote:
> The driver exposes EFI runtime services to user-space through an IOCTL
> interface, calling the EFI services function pointers directly without
> using the efivar API.
>
> Disallow access to the /dev/efi_test character device when the kernel is
> locked down to prevent arbitrary user-space to call EFI runtime services.
>
> Also require CAP_SYS_ADMIN to open the chardev to prevent unprivileged
> users to call the EFI runtime services, instead of just relying on the
> chardev file mode bits for this.
>
> The main user of this driver is the fwts [0] tool that already checks if
> the effective user ID is 0 and fails otherwise. So this change shouldn't
> cause any regression to this tool.
>
> [0]: https://wiki.ubuntu.com/FirmwareTestSuite/Reference/uefivarinfo
>
> Signed-off-by: Javier Martinez Canillas <[email protected]>
> Acked-by: Laszlo Ersek <[email protected]>
>
> ---
>
> Changes in v2:
> - Also disable /dev/efi_test access when the kernel is locked down as
> suggested by Matthew Garrett.
Right; if you remember the pre-patch discussion off-list, we kind of
expected that lockdown might affect this. :)
... And, I can see Matt's comment now, at
<https://bugzilla.redhat.com/show_bug.cgi?id=1759325#c1>. Thanks for that!
While this change decreases the usability of the module, I fully agree
it is justified for production use. While it's more convenient for me to
keep SB enabled in the test VM(s) in general, and just run the test
whenever I need it, security trumps convenience. I can disable SB when
necessary, or even dedicate separate VMs (with SB generally disabled) to
this kind of testing.
> - Add Acked-by tag from Laszlo Ersek.
My ACK stands -- I don't know enough to validate the
security_locked_down() call and its friends, but I'm OK with the intent.
Thanks all!
Laszlo
>
> drivers/firmware/efi/test/efi_test.c | 8 ++++++++
> include/linux/security.h | 1 +
> security/lockdown/lockdown.c | 1 +
> 3 files changed, 10 insertions(+)
>
> diff --git a/drivers/firmware/efi/test/efi_test.c b/drivers/firmware/efi/test/efi_test.c
> index 877745c3aaf..7baf48c01e7 100644
> --- a/drivers/firmware/efi/test/efi_test.c
> +++ b/drivers/firmware/efi/test/efi_test.c
> @@ -14,6 +14,7 @@
> #include <linux/init.h>
> #include <linux/proc_fs.h>
> #include <linux/efi.h>
> +#include <linux/security.h>
> #include <linux/slab.h>
> #include <linux/uaccess.h>
>
> @@ -717,6 +718,13 @@ static long efi_test_ioctl(struct file *file, unsigned int cmd,
>
> static int efi_test_open(struct inode *inode, struct file *file)
> {
> + int ret = security_locked_down(LOCKDOWN_EFI_TEST);
> +
> + if (ret)
> + return ret;
> +
> + if (!capable(CAP_SYS_ADMIN))
> + return -EACCES;
> /*
> * nothing special to do here
> * We do accept multiple open files at the same time as we
> diff --git a/include/linux/security.h b/include/linux/security.h
> index a8d59d612d2..9df7547afc0 100644
> --- a/include/linux/security.h
> +++ b/include/linux/security.h
> @@ -105,6 +105,7 @@ enum lockdown_reason {
> LOCKDOWN_NONE,
> LOCKDOWN_MODULE_SIGNATURE,
> LOCKDOWN_DEV_MEM,
> + LOCKDOWN_EFI_TEST,
> LOCKDOWN_KEXEC,
> LOCKDOWN_HIBERNATION,
> LOCKDOWN_PCI_ACCESS,
> diff --git a/security/lockdown/lockdown.c b/security/lockdown/lockdown.c
> index 8a10b43daf7..40b790536de 100644
> --- a/security/lockdown/lockdown.c
> +++ b/security/lockdown/lockdown.c
> @@ -20,6 +20,7 @@ static const char *const lockdown_reasons[LOCKDOWN_CONFIDENTIALITY_MAX+1] = {
> [LOCKDOWN_NONE] = "none",
> [LOCKDOWN_MODULE_SIGNATURE] = "unsigned module loading",
> [LOCKDOWN_DEV_MEM] = "/dev/mem,kmem,port",
> + [LOCKDOWN_EFI_TEST] = "/dev/efi_test access",
> [LOCKDOWN_KEXEC] = "kexec of unsigned images",
> [LOCKDOWN_HIBERNATION] = "hibernation",
> [LOCKDOWN_PCI_ACCESS] = "direct PCI access",
>
On Tue, Oct 8, 2019 at 9:55 PM Javier Martinez Canillas
<[email protected]> wrote:
> Signed-off-by: Javier Martinez Canillas <[email protected]>
> Acked-by: Laszlo Ersek <[email protected]>
Acked-by: Matthew Garrett <[email protected]>
On Wed, 9 Oct 2019 at 04:18, Matthew Garrett <[email protected]> wrote:
>
> On Tue, Oct 8, 2019 at 9:55 PM Javier Martinez Canillas
> <[email protected]> wrote:
> > Signed-off-by: Javier Martinez Canillas <[email protected]>
> > Acked-by: Laszlo Ersek <[email protected]>
>
> Acked-by: Matthew Garrett <[email protected]>
Thanks all. Queued as a fix.