forked from Minki/linux
drm/i915: prevent FIFO calculation overflows on 32 bits with high dotclocks
A very high dotclock (e.g. 229500kHz as reported by Anton) can cause the entries_required variable to overflow, potentially leading to a FIFO watermark value that's too low to support the given mode. Split the division across the calculation to avoid this. Cc: stable@kernel.org Reported-by: Anton Khirnov <wyskas@gmail.com> Tested-by: Anton Khirnov <wyskas@gmail.com> Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
This commit is contained in:
parent
decbbcda29
commit
d660467c3f
@ -2119,7 +2119,14 @@ static unsigned long intel_calculate_wm(unsigned long clock_in_khz,
|
||||
{
|
||||
long entries_required, wm_size;
|
||||
|
||||
entries_required = (clock_in_khz * pixel_size * latency_ns) / 1000000;
|
||||
/*
|
||||
* Note: we need to make sure we don't overflow for various clock &
|
||||
* latency values.
|
||||
* clocks go from a few thousand to several hundred thousand.
|
||||
* latency is usually a few thousand
|
||||
*/
|
||||
entries_required = ((clock_in_khz / 1000) * pixel_size * latency_ns) /
|
||||
1000;
|
||||
entries_required /= wm->cacheline_size;
|
||||
|
||||
DRM_DEBUG("FIFO entries required for mode: %d\n", entries_required);
|
||||
|
Loading…
Reference in New Issue
Block a user