Message ID | 20181016130641.GA603@embeddedor.com |
---|---|
State | Accepted, archived |
Delegated to: | David Miller |
Headers | show |
Series | ptp: fix Spectre v1 vulnerability | expand |
On Tue, Oct 16, 2018 at 03:06:41PM +0200, Gustavo A. R. Silva wrote: > pin_index can be indirectly controlled by user-space, hence leading > to a potential exploitation of the Spectre variant 1 vulnerability. > > This issue was detected with the help of Smatch: > > drivers/ptp/ptp_chardev.c:253 ptp_ioctl() warn: potential spectre issue > 'ops->pin_config' [r] (local cap) > > Fix this by sanitizing pin_index before using it to index > ops->pin_config, and before passing it as an argument to > function ptp_set_pinfunc(), in which it is used to index > info->pin_config. Acked-by: Richard Cochran <richardcochran@gmail.com>
From: "Gustavo A. R. Silva" <gustavo@embeddedor.com> Date: Tue, 16 Oct 2018 15:06:41 +0200 > pin_index can be indirectly controlled by user-space, hence leading > to a potential exploitation of the Spectre variant 1 vulnerability. > > This issue was detected with the help of Smatch: > > drivers/ptp/ptp_chardev.c:253 ptp_ioctl() warn: potential spectre issue > 'ops->pin_config' [r] (local cap) > > Fix this by sanitizing pin_index before using it to index > ops->pin_config, and before passing it as an argument to > function ptp_set_pinfunc(), in which it is used to index > info->pin_config. > > Notice that given that speculation windows are large, the policy is > to kill the speculation on the first load and not worry if it can be > completed with a dependent load/store [1]. > > [1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2 > > Cc: stable@vger.kernel.org > Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com> Applied.
diff --git a/drivers/ptp/ptp_chardev.c b/drivers/ptp/ptp_chardev.c index 01b0e2b..2012551 100644 --- a/drivers/ptp/ptp_chardev.c +++ b/drivers/ptp/ptp_chardev.c @@ -24,6 +24,8 @@ #include <linux/slab.h> #include <linux/timekeeping.h> +#include <linux/nospec.h> + #include "ptp_private.h" static int ptp_disable_pinfunc(struct ptp_clock_info *ops, @@ -248,6 +250,7 @@ long ptp_ioctl(struct posix_clock *pc, unsigned int cmd, unsigned long arg) err = -EINVAL; break; } + pin_index = array_index_nospec(pin_index, ops->n_pins); if (mutex_lock_interruptible(&ptp->pincfg_mux)) return -ERESTARTSYS; pd = ops->pin_config[pin_index]; @@ -266,6 +269,7 @@ long ptp_ioctl(struct posix_clock *pc, unsigned int cmd, unsigned long arg) err = -EINVAL; break; } + pin_index = array_index_nospec(pin_index, ops->n_pins); if (mutex_lock_interruptible(&ptp->pincfg_mux)) return -ERESTARTSYS; err = ptp_set_pinfunc(ptp, pin_index, pd.func, pd.chan);
pin_index can be indirectly controlled by user-space, hence leading to a potential exploitation of the Spectre variant 1 vulnerability. This issue was detected with the help of Smatch: drivers/ptp/ptp_chardev.c:253 ptp_ioctl() warn: potential spectre issue 'ops->pin_config' [r] (local cap) Fix this by sanitizing pin_index before using it to index ops->pin_config, and before passing it as an argument to function ptp_set_pinfunc(), in which it is used to index info->pin_config. Notice that given that speculation windows are large, the policy is to kill the speculation on the first load and not worry if it can be completed with a dependent load/store [1]. [1] https://marc.info/?l=linux-kernel&m=152449131114778&w=2 Cc: stable@vger.kernel.org Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com> --- drivers/ptp/ptp_chardev.c | 4 ++++ 1 file changed, 4 insertions(+)