[dpdk-dev] [PATCH] librte_eal: Allow combining --no-huge with -m XXX

[dpdk-dev] [PATCH] librte_eal: Allow combining --no-huge with -m XXX

[Simon Kagstrom]

Useful to run applications in usermode via a test driver.

Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
---
Not sure if there are other implications of this, so please check!

 lib/librte_eal/common/eal_common_options.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/librte_eal/common/eal_common_options.c b/lib/librte_eal/common/eal_common_options.c
index 4319549..af865f5 100644
--- a/lib/librte_eal/common/eal_common_options.c
+++ b/lib/librte_eal/common/eal_common_options.c
@@ -850,7 +850,7 @@ eal_check_common_options(struct internal_config *internal_cfg)
 		return -1;
 	}
 	if (internal_cfg->no_hugetlbfs &&
-			(mem_parsed || internal_cfg->force_sockets == 1)) {
+			(internal_cfg->force_sockets == 1)) {
 		RTE_LOG(ERR, EAL, "Options -m or --"OPT_SOCKET_MEM" cannot "
 			"be specified together with --"OPT_NO_HUGE"\n");
 		return -1;
-- 
1.9.1

Re: [dpdk-dev] [PATCH] librte_eal: Allow combining --no-huge with -m XXX

[David Marchand]

Hello Simon,

On Mon, Mar 23, 2015 at 3:33 PM, Simon Kagstrom <
simon.kagstrom@netinsight.net> wrote:

> Useful to run applications in usermode via a test driver.
>

I would add that this commit removes a constraint on no-huge mode being
limited to 64M.


> Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
> ---
> Not sure if there are other implications of this, so please check!
>

Looked around, I can find no issue with this change.


>  lib/librte_eal/common/eal_common_options.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/librte_eal/common/eal_common_options.c
> b/lib/librte_eal/common/eal_common_options.c
> index 4319549..af865f5 100644
> --- a/lib/librte_eal/common/eal_common_options.c
> +++ b/lib/librte_eal/common/eal_common_options.c
> @@ -850,7 +850,7 @@ eal_check_common_options(struct internal_config
> *internal_cfg)
>                 return -1;
>         }
>         if (internal_cfg->no_hugetlbfs &&
> -                       (mem_parsed || internal_cfg->force_sockets == 1)) {
> +                       (internal_cfg->force_sockets == 1)) {
>                 RTE_LOG(ERR, EAL, "Options -m or --"OPT_SOCKET_MEM" cannot
> "
>                         "be specified together with --"OPT_NO_HUGE"\n");
>                 return -1;
>

We still can't use  --socket-mem in no-huge mode since we can't pin memory
to socket at the moment, so the test on force_sockets can remain.

However, you missed reference to -m in the log message. Can you remove it ?


In which situation did you encounter this limitation and how did you test
this change ?
This should go in the commit log.


Thanks.

-- 
David Marchand