[LTP] [PATCH v2 4/4] readahead02: reset timeout based the IO test elapsed

[Li Wang]

Note: Patch 4/4 need to be rewirten.

If go with Li Wang's .runtime thought, just ignore this patch should be ok.

If go with Cryil's suggestion, need changes to:

--- a/testcases/kernel/syscalls/readahead/readahead02.c
+++ b/testcases/kernel/syscalls/readahead/readahead02.c
@@ -318,19 +318,6 @@ static void test_readahead(unsigned int n)
                tst_res(TCONF, "Page cache on your system is too small "
                        "to hold whole testfile.");
        }
-
-       /*
-        * The time consuming of readahead quite depending on the platform
IO
-        * speed, sometime test timeout when the default runtime is used up.
-        *
-        *  readahead02.c:221: TINFO: Test #2: POSIX_FADV_WILLNEED on file
-        *  readahead02.c:285: TINFO: read_testfile(0) took: 26317623 usec
-        *  readahead02.c:286: TINFO: read_testfile(1) took: 26101484 usec
-        *
-        * Here raise the maximum runtime dynamically.
-        */
-       if ((tc+1)->readahead)
-               tst_set_runtime(test.runtime + (usec + usec_ra) / 1000000);
 }


@@ -422,7 +409,7 @@ static struct tst_test test = {
        },
        .test = test_readahead,
        .tcnt = ARRAY_SIZE(tcases),
-       .runtime = 30,
+       .timeout = 60,   // or another value measured on slow HW
        .tags = (const struct tst_tag[]) {
                {"linux-git", "b833a3660394"},
                {"linux-git", "5b910bd615ba"},


-- 
Regards,
Li Wang