Commit f4065395 authored Aug 14, 2015 by Arun Menon

msm: vidc: Avoid integer overflow for supported bitrate calculation



Supported bitrate is calculated as 40% of clock frequency.
However table[i].freq is an unsigned 32 bit integer. Multiplying
table[i].freq by 40 would cause an integer overflow and hence the
resulting supported_bitrate calculated for that frequency would
be lower. This would result in the frequency being bumped up
higher to meeting the bitrate requirement.

Change-Id: Idbdfd2301a6ad5b1d7ba0b7374375f9856f3e118
Signed-off-by: Arun Menon <avmenon@codeaurora.org>

parent 65bdd3d2

drivers/media/platform/msm/vidc/venus_hfi.c

+1 −1

Original line number	Diff line number	Diff line
		@@ -1359,7 +1359,7 @@ static unsigned long __get_clock_rate_with_bitrate(struct clock_info *clock,
		continue;

		freq = table[i].freq;
		supported_bitrate = table[i].freq * 40/100;
		supported_bitrate = freq * 40/100;

		if (table[i].freq >= base_freq &&
		supported_bitrate >= instant_bitrate)