Author: cem
Date: Fri Dec 13 04:48:20 2019
New Revision: 355694
URL: https://svnweb.freebsd.org/changeset/base/355694
Log:
kern/subr_unit: Rip srandomdev, random(3) out of dead code
The simulation cannot be reproduced, so the value of using a deterministic
PRNG
like random(3) is dubious. The number of repitions used in the sample isn't a
problem for the Chacha implementation of arc4random we have today. (Also, no
one actually runs this code; it was provided as an example of the work the
author did validating the implementation. It's not even test code.)
Modified:
head/sys/kern/subr_unit.c
Modified: head/sys/kern/subr_unit.c
==============================================================================
--- head/sys/kern/subr_unit.c Fri Dec 13 04:37:39 2019 (r355693)
+++ head/sys/kern/subr_unit.c Fri Dec 13 04:48:20 2019 (r355694)
@@ -1021,7 +1021,7 @@ main(int argc, char **argv)
long count = 10000; /* Number of unrs to test */
long reps = 1, m;
int ch;
- u_int i, j;
+ u_int i;
verbose = false;
@@ -1055,20 +1055,18 @@ main(int argc, char **argv)
a = calloc(count, sizeof(char));
if (a == NULL)
err(1, "calloc failed");
- srandomdev();
printf("sizeof(struct unr) %zu\n", sizeof(struct unr));
printf("sizeof(struct unrb) %zu\n", sizeof(struct unrb));
printf("sizeof(struct unrhdr) %zu\n", sizeof(struct unrhdr));
printf("NBITS %lu\n", (unsigned long)NBITS);
for (m = 0; m < count * reps; m++) {
- j = random();
- i = (j >> 1) % count;
+ i = arc4random_uniform(count);
#if 0
if (a[i] && (j & 1))
continue;
#endif
- if ((random() & 1) != 0)
+ if ((arc4random() & 1) != 0)
test_alloc_unr(uh, i, a);
else
test_alloc_unr_specific(uh, i, a);
_______________________________________________
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
