app/testpmd: report invalid command line parameter
Checks
Commit Message
We currently do not check that a non option string has been passed to
testpmd.
Example:
$ ./master/app/testpmd --no-huge -m 512 --vdev net_null0 \
--vdev net_null1 -- -i nb-cores=2 --total-num-mbuf 2048
[...]
testpmd> show config fwd
io packet forwarding - ports=2 - cores=1 - streams=2 - NUMA support
enabled, MP allocation mode: native
Logical Core 1 (socket 0) forwards packets on 2 streams:
RX P=0/Q=0 (socket 0) -> TX P=1/Q=0 (socket 0) peer=02:00:00:00:00:01
RX P=1/Q=0 (socket 0) -> TX P=0/Q=0 (socket 0) peer=02:00:00:00:00:00
Here nb-cores=2 is just ignored, while the (probably sleepy) user did not
notice this.
Validate that all strings passed to testpmd are part of a known option.
After this patch:
$ ./master/app/testpmd --no-huge -m 512 --vdev net_null0 \
--vdev net_null1 -- -i -nb-cores=2 --total-num-mbuf 2048
[...]
Invalid parameter: nb-cores=2
EAL: Error - exiting with code: 1
Cause: Command line incorrect
While at it, when passing an unknown option, print the string that gets
refused by getopt_long to help the user.
Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
---
This seems a bit dangerous to take this kind of change.
Some "working fine" scripts might now report failures from testpmd because
of garbage in the command line.
Sending the patch anyway to see what others think about it.
---
app/test-pmd/parameters.c | 7 +++++++
1 file changed, 7 insertions(+)
Comments
On Sun, Nov 17, 2019 at 6:10 PM David Marchand
<david.marchand@redhat.com> wrote:
> After this patch:
> $ ./master/app/testpmd --no-huge -m 512 --vdev net_null0 \
> --vdev net_null1 -- -i -nb-cores=2 --total-num-mbuf 2048
Sorry wrong example, s/-nb-cores/nb-cores/
> [...]
> Invalid parameter: nb-cores=2
> EAL: Error - exiting with code: 1
> Cause: Command line incorrect
But the output was correct.
On 11/17/2019 5:10 PM, David Marchand wrote:
> We currently do not check that a non option string has been passed to
> testpmd.
>
> Example:
> $ ./master/app/testpmd --no-huge -m 512 --vdev net_null0 \
> --vdev net_null1 -- -i nb-cores=2 --total-num-mbuf 2048
> [...]
> testpmd> show config fwd
> io packet forwarding - ports=2 - cores=1 - streams=2 - NUMA support
> enabled, MP allocation mode: native
> Logical Core 1 (socket 0) forwards packets on 2 streams:
> RX P=0/Q=0 (socket 0) -> TX P=1/Q=0 (socket 0) peer=02:00:00:00:00:01
> RX P=1/Q=0 (socket 0) -> TX P=0/Q=0 (socket 0) peer=02:00:00:00:00:00
>
> Here nb-cores=2 is just ignored, while the (probably sleepy) user did not
> notice this.
>
> Validate that all strings passed to testpmd are part of a known option.
>
> After this patch:
> $ ./master/app/testpmd --no-huge -m 512 --vdev net_null0 \
> --vdev net_null1 -- -i -nb-cores=2 --total-num-mbuf 2048
> [...]
> Invalid parameter: nb-cores=2
> EAL: Error - exiting with code: 1
> Cause: Command line incorrect
>
> While at it, when passing an unknown option, print the string that gets
> refused by getopt_long to help the user.
>
> Fixes: af75078fece3 ("first public release")
> Cc: stable@dpdk.org
>
> Signed-off-by: David Marchand <david.marchand@redhat.com>
> ---
> This seems a bit dangerous to take this kind of change.
> Some "working fine" scripts might now report failures from testpmd because
> of garbage in the command line.
>
> Sending the patch anyway to see what others think about it.
+1 to the patch, I think better to fail in this case.
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
@@ -1363,12 +1363,19 @@ launch_args_parse(int argc, char** argv)
break;
default:
usage(argv[0]);
+ printf("Invalid option: %s\n", argv[optind]);
rte_exit(EXIT_FAILURE,
"Command line is incomplete or incorrect\n");
break;
}
}
+ if (optind != argc) {
+ usage(argv[0]);
+ printf("Invalid parameter: %s\n", argv[optind]);
+ rte_exit(EXIT_FAILURE, "Command line is incorrect\n");
+ }
+
/* Set offload configuration from command line parameters. */
rx_mode.offloads = rx_offloads;
tx_mode.offloads = tx_offloads;