[alsa-devel] [PATCH] ALSA: emux_hwdep: Fix potential Spectre v1 vulnerabilities
info.mode and info.port are indirectly controlled by user-space, hence leading to a potential exploitation of the Spectre variant 1 vulnerability.
These issues were detected with the help of Smatch:
sound/synth/emux/emux_hwdep.c:72 snd_emux_hwdep_misc_mode() warn: potential spectre issue 'emu->portptrs[i]->ctrls' [w] (local cap) sound/synth/emux/emux_hwdep.c:75 snd_emux_hwdep_misc_mode() warn: potential spectre issue 'emu->portptrs' [w] (local cap) sound/synth/emux/emux_hwdep.c:75 snd_emux_hwdep_misc_mode() warn: potential spectre issue 'emu->portptrs[info.port]->ctrls' [w] (local cap)
Fix this by sanitizing both info.mode and info.port before using them to index emu->portptrs[i]->ctrls, emu->portptrs[info.port]->ctrls and emu->portptrs.
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 --- sound/synth/emux/emux_hwdep.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/sound/synth/emux/emux_hwdep.c b/sound/synth/emux/emux_hwdep.c index e557946718a9..a0d5cd99572f 100644 --- a/sound/synth/emux/emux_hwdep.c +++ b/sound/synth/emux/emux_hwdep.c @@ -24,6 +24,7 @@ #include <linux/uaccess.h> #include "emux_voice.h"
+#include <linux/nospec.h>
#define TMP_CLIENT_ID 0x1001
@@ -66,13 +67,16 @@ snd_emux_hwdep_misc_mode(struct snd_emux *emu, void __user *arg) return -EFAULT; if (info.mode < 0 || info.mode >= EMUX_MD_END) return -EINVAL; + info.mode = array_index_nospec(info.mode, EMUX_MD_END);
if (info.port < 0) { for (i = 0; i < emu->num_ports; i++) emu->portptrs[i]->ctrls[info.mode] = info.value; } else { - if (info.port < emu->num_ports) + if (info.port < emu->num_ports) { + info.port = array_index_nospec(info.port, emu->num_ports); emu->portptrs[info.port]->ctrls[info.mode] = info.value; + } } return 0; }
Hi,
[This is an automated email]
This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all
The bot has tested the following trees: v4.19.8, v4.14.87, v4.9.144, v4.4.166, v3.18.128,
v4.19.8: Build OK! v4.14.87: Build OK! v4.9.144: Build OK! v4.4.166: Build OK! v3.18.128: Build failed! Errors: sound/synth/emux/emux_hwdep.c:27:10: fatal error: linux/nospec.h: No such file or directory
How should we proceed with this patch?
-- Thanks, Sasha
On Wed, 12 Dec 2018 18:20:49 +0100, Gustavo A. R. Silva wrote:
info.mode and info.port are indirectly controlled by user-space, hence leading to a potential exploitation of the Spectre variant 1 vulnerability.
These issues were detected with the help of Smatch:
sound/synth/emux/emux_hwdep.c:72 snd_emux_hwdep_misc_mode() warn: potential spectre issue 'emu->portptrs[i]->ctrls' [w] (local cap) sound/synth/emux/emux_hwdep.c:75 snd_emux_hwdep_misc_mode() warn: potential spectre issue 'emu->portptrs' [w] (local cap) sound/synth/emux/emux_hwdep.c:75 snd_emux_hwdep_misc_mode() warn: potential spectre issue 'emu->portptrs[info.port]->ctrls' [w] (local cap)
Fix this by sanitizing both info.mode and info.port before using them to index emu->portptrs[i]->ctrls, emu->portptrs[info.port]->ctrls and emu->portptrs.
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 now (with moving the linux/nospec.h in a more appropriate line).
thanks,
Takashi
participants (3)
-
Gustavo A. R. Silva
-
Sasha Levin
-
Takashi Iwai