Re: [RFC PATCH] intel_pstate: Fix possible overflow complained by Coverity
From: Rafael J. Wysocki <hidden>
Date: 2015-08-05 23:43:58
On Wednesday, July 29, 2015 11:53:10 PM Chen Yu wrote:
Coverity scanning performed on intel_pstate.c shows possible overflow when doing left shifting: val = pstate << 8; since pstate is of type integer, while val is of u64, left shifting pstate might lead to potential loss of upper bits. Say, if pstate equals 0x4000 0000, after pstate << 8 we will get zero assigned to val. Although pstate will not likely be that big, this patch cast the left operand to u64 before performing the left shift, to avoid complaining from Coverity. Reported-by: Coquard, Christophe <redacted> Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Patch queued up for 4.3, thanks!
quoted hunk ↗ jump to hunk
--- drivers/cpufreq/intel_pstate.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c index 15ada47..6effe81 100644 --- a/drivers/cpufreq/intel_pstate.c +++ b/drivers/cpufreq/intel_pstate.c@@ -522,7 +522,7 @@ static void byt_set_pstate(struct cpudata *cpudata, int pstate) int32_t vid_fp; u32 vid; - val = pstate << 8; + val = (u64)pstate << 8; if (limits.no_turbo && !limits.turbo_disabled) val |= (u64)1 << 32;@@ -611,7 +611,7 @@ static void core_set_pstate(struct cpudata *cpudata, int pstate) { u64 val; - val = pstate << 8; + val = (u64)pstate << 8; if (limits.no_turbo && !limits.turbo_disabled) val |= (u64)1 << 32;
-- I speak only for myself. Rafael J. Wysocki, Intel Open Source Technology Center.