ntp: Fixup adjtimex freq validation on 32bit systems
Subsystems (2)
PATCH 1/1 ntp: Fixup adjtimex freq validation on 32bit systems
For tip/timers/urgent.
Additional validation of adjtimex freq values to avoid potential multiplication overflows were added in commit 5e5aeb4367b (time: adjtimex: Validate the ADJ_FREQUENCY values)
Unfortunately the patch used LONG_MAX/MIN instead of LLONG_MAX/MIN, which was fine on 64bit systems, but being much smaller on 32bit systems caused false positives resulting in most direct frequency adjustments to fail w/ EINVAL.
ntpd only does direct frequency adjustments at startup, so the issue was not as easily observed there, but other time sync applications like ptpd and chrony were more effected by the bug.
See bugs:
https://bugzilla.kernel.org/show_bug.cgi?id=92481
https://bugzilla.redhat.com/show_bug.cgi?id=1188074
---
One note:
0day kbuild bot complains about
Quoted 2 lines
Quoted 2 lines
kernel/time/ntp.c:637: warning: comparison is always false due to limited range of data typekernel/time/ntp.c:639: warning: comparison is always false due to limited range of data type
We could fix this via adding an extra (BITS_PER_LONG == 64) case before we check these to avoid it, but that seemed a bit too ugly to me. Thoughts?
kernel/time/ntp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
kernel/time/ntp.c +2 -2
@@ -634,9 +634,9 @@ int ntp_validate_timex(struct timex *txc)return -EPERM ;if (txc->modes &ADJ_FREQUENCY ) { -if (LONG_MIN /PPM_SCALE > txc->freq) +if (LLONG_MIN /PPM_SCALE > txc->freq)return -EINVAL ; -if (LONG_MAX /PPM_SCALE < txc->freq) +if (LLONG_MAX /PPM_SCALE < txc->freq)return -EINVAL ; }
Signature · 1 line
1.9.1
Quoted 5 lines · John Stultz
Unfortunately the patch used LONG_MAX/MIN instead of LLONG_MAX/MIN, which was fine on 64bit systems, but being much smaller on 32bit systems caused false positives resulting in most direct frequency adjustments to fail w/ EINVAL.
...
Quoted 8 lines
One note:
0day kbuild bot complains aboutQuoted 2 lines
Quoted 2 lines
kernel/time/ntp.c:637: warning: comparison is always false due to limited range of data typekernel/time/ntp.c:639: warning: comparison is always false due to limited range of data typeWe could fix this via adding an extra (
BITS_PER_LONG== 64) case before we check these to avoid it, but that seemed a bit too ugly to me. Thoughts?
So the check is 64 bit only? Might as well mark it like that explicitly to avoid needless head scratching of 32 bit people.
Thanks,
Richard
PATCH v2 1/1 ntp: Fixup adjtimex freq validation on 32bit systems
For tip/timers/urgent.
Additional validation of adjtimex freq values to avoid potential multiplication overflows were added in commit 5e5aeb4367b (time: adjtimex: Validate the ADJ_FREQUENCY values)
Unfortunately the patch used LONG_MAX/MIN instead of LLONG_MAX/MIN, which was fine on 64bit systems, but being much smaller on 32bit systems caused false positives resulting in most direct frequency adjustments to fail w/ EINVAL.
ntpd only does direct frequency adjustments at startup, so the issue was not as easily observed there, but other time sync applications like ptpd and chrony were more effected by the bug.
See bugs:
https://bugzilla.kernel.org/show_bug.cgi?id=92481
https://bugzilla.redhat.com/show_bug.cgi?id=1188074
This patch changes the checks to use LLONG_MAX for clarity, and additionally the checks are disabled on 32bit systems since LLONG_MAX/PPM_SCALE is always larger then the 32bit long freq value, so mult overflows aren't possible there.
---
v2: Included the BITS_PER_LONG check to disable the check on 32bit systems, since it generates build warnings in some settings.
kernel/time/ntp.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
kernel/time/ntp.c +7 -3
@@ -633,10 +633,14 @@ int ntp_validate_timex(struct timex *txc)if ((txc->modes &ADJ_SETOFFSET ) && (!capable (CAP_SYS_TIME )))return -EPERM ; -if (txc->modes &ADJ_FREQUENCY ) { -if (LONG_MIN /PPM_SCALE > txc->freq) +/* +* Check for potential mult overflows that can +* only happen on 64bit systems. +*/ +if ((txc->modes &ADJ_FREQUENCY ) && (BITS_PER_LONG ==64 )) { +if (LLONG_MIN /PPM_SCALE > txc->freq)return -EINVAL ; -if (LONG_MAX /PPM_SCALE < txc->freq) +if (LLONG_MAX /PPM_SCALE < txc->freq)return -EINVAL ; }
Signature · 1 line
1.9.1