On 11-Oct-17 18:26, Jens Axboe wrote:
On 10/11/2017 04:18 PM, Sitsofe Wheeler wrote:
Hi Jens,
On 10 October 2017 at 22:27, Andrzej Jakowski
<andrzej.jakowski@xxxxxxxxx> wrote:
This patch fixes percentile calculation for Windows where overflow
on 32 bit type happened while calculating actual percentiles.
Signed-off-by: Andrzej Jakowski <andrzej.jakowski@xxxxxxxxx>
---
stat.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/stat.c b/stat.c
index c5a68ad5..5c758680 100644
--- a/stat.c
+++ b/stat.c
@@ -139,7 +139,7 @@ unsigned int calc_clat_percentiles(unsigned int *io_u_plat, unsigned long nr,
fio_fp64_t *plist, unsigned long long **output,
unsigned long long *maxv, unsigned long long *minv)
{
- unsigned long sum = 0;
+ unsigned long long sum = 0;
unsigned int len, i, j = 0;
unsigned int oval_len = 0;
unsigned long long *ovals = NULL;
--
2.13.6
Just in case you didn't see it, the above looks sensible/necessary...
Thanks for pointing it my way, I don't always see the mailing list
posts quickly.
Looks good to me, I'll get it applied. Thanks Andrzej!
Andrzej, does this actually resolve the problem you observed? unsigned
long nr in the parameter list of calc_clat_percentiles() and
show_clat_percentiles() also looks like it will overflow when you have
more than 2^32 IOs.
--
To unsubscribe from this list: send the line "unsubscribe fio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
