[dpdk-dev] [PATCH] test: test zero socket-mem as valid

[Ilya Maximets]

On 25.01.2019 16:48, Burakov, Anatoly wrote:
> On 25-Jan-19 9:53 AM, David Marchand wrote:
>>
>>
>> On Fri, Jan 25, 2019 at 9:06 AM Ilya Maximets <i.maximets at samsung.com <mailto:i.maximets at samsung.com>> wrote:
>>
>>     On 25.01.2019 10:55, Ilya Maximets wrote:
>>      > Dynamic memory mode allowes zero socket-mem because all the
>>      > required memory could be allocated on demand.
>>      >
>>      > Fixes: 339c2244b4f1 ("eal: fix parsing zero socket memory and
>>     limits")
>>      > Cc: stable at dpdk.org <mailto:stable at dpdk.org>
>>      >
>>
>>     Reported-by: Shuai Zhu <shuaix.zhu at intel.com
>>     <mailto:shuaix.zhu at intel.com>>
>>
>>      > Signed-off-by: Ilya Maximets <i.maximets at samsung.com
>>     <mailto:i.maximets at samsung.com>>
>>
>>      > ---
>>      >  test/test/test_eal_flags.c | 6 +++---
>>      >  1 file changed, 3 insertions(+), 3 deletions(-)
>>      >
>>      > diff --git a/test/test/test_eal_flags.c b/test/test/test_eal_flags.c
>>      > index e3a60c7ae..81e345b87 100644
>>      > --- a/test/test/test_eal_flags.c
>>      > +++ b/test/test/test_eal_flags.c
>>      > @@ -1158,7 +1158,7 @@ test_memory_flags(void)
>>      >       const char *argv1[] = {prgname, "-c", "10", "-n", "2",
>>      >                       "--file-prefix=" memtest, "-m",
>>     DEFAULT_MEM_SIZE};
>>      >
>>      > -     /* invalid (zero) --socket-mem flag */
>>      > +     /* valid (zero) --socket-mem flag */
>>      >       const char *argv2[] = {prgname, "-c", "10", "-n", "2",
>>      >                       "--file-prefix=" memtest,
>>     "--socket-mem=0,0,0,0"};
>>      >
>>      > @@ -1256,8 +1256,8 @@ test_memory_flags(void)
>>      >               printf("Error - process failed with valid -m flag!\n");
>>      >               return -1;
>>      >       }
>>      > -     if (launch_proc(argv2) == 0) {
>>      > -             printf("Error - process run ok with invalid (zero)
>>     --socket-mem!\n");
>>      > +     if (launch_proc(argv2) != 0) {
>>      > +             printf("Error - process failed with valid (zero)
>>     --socket-mem!\n");
>>      >               return -1;
>>      >       }
>>      >
>>      >
>>
>>
>> Reviewed-by: David Marchand <david.marchand at redhat.com <mailto:david.marchand at redhat.com>>
>>
>>
>> -- 
>> David Marchand
> 
> Now that i think of it, maybe it's not that simple.
> 
> --socket-mem/-m flag with zero is still an invalid value *if* --legacy-mem is involved. However, it is a valid value in non-legacy mode.
> 
> So maybe the test should reflect this, and the previous fix should have instead added a check for legacy mode rather than disabling the zero check outright.
> 

I don't think that it's a big deal, because "--socket-mem=0 --legacy-mem"
quickly fails with clear:

  EAL: WARNING: Master core has no memory on local socket!

IMHO, It's actually more informative than previous:

  EAL: invalid parameters for --socket-limit

I agree that we could add a test for a legacy-mem cases, but that's a bit
different task.