* [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg
@ 2025-10-17 11:06 Bruce Richardson
2025-10-17 11:06 ` [PATCH 2/2] app/pdump: improve cmdline argument handling Bruce Richardson
2025-11-07 16:10 ` [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg Burakov, Anatoly
0 siblings, 2 replies; 4+ messages in thread
From: Bruce Richardson @ 2025-10-17 11:06 UTC (permalink / raw)
To: dev; +Cc: Bruce Richardson, stable
The pdump main function modifies argc/argv before calling EAL init
function. It needs to do so in order to add the proc-type=secondary
flag, but also adds in "-n4" as parameter. This parameter is unnecessary
and also causes duplicate parameter errors if the user also passes in -n
flag on the commandline. Just remove this flag.
Fixes: caa7028276b8 ("app/pdump: add tool for packet capturing")
Cc: stable@dpdk.org
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
app/pdump/main.c | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)
diff --git a/app/pdump/main.c b/app/pdump/main.c
index fa85859703..f3ac81cb80 100644
--- a/app/pdump/main.c
+++ b/app/pdump/main.c
@@ -982,9 +982,8 @@ main(int argc, char **argv)
int ret;
int i;
- char n_flag[] = "-n4";
char mp_flag[] = "--proc-type=secondary";
- char *argp[argc + 2];
+ char *argp[argc + 1];
/* catch ctrl-c so we can cleanup on exit */
sigemptyset(&action.sa_mask);
@@ -996,13 +995,12 @@ main(int argc, char **argv)
sigaction(SIGHUP, &action, NULL);
argp[0] = argv[0];
- argp[1] = n_flag;
- argp[2] = mp_flag;
+ argp[1] = mp_flag;
for (i = 1; i < argc; i++)
- argp[i + 2] = argv[i];
+ argp[i + 1] = argv[i];
- argc += 2;
+ argc += 1;
diag = rte_eal_init(argc, argp);
if (diag < 0)
@@ -1012,7 +1010,7 @@ main(int argc, char **argv)
rte_exit(EXIT_FAILURE, "No Ethernet ports - bye\n");
argc -= diag;
- argv += (diag - 2);
+ argv += (diag - 1);
/* parse app arguments */
if (argc > 1) {
--
2.48.1
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH 2/2] app/pdump: improve cmdline argument handling
2025-10-17 11:06 [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg Bruce Richardson
@ 2025-10-17 11:06 ` Bruce Richardson
2025-11-07 16:14 ` Burakov, Anatoly
2025-11-07 16:10 ` [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg Burakov, Anatoly
1 sibling, 1 reply; 4+ messages in thread
From: Bruce Richardson @ 2025-10-17 11:06 UTC (permalink / raw)
To: dev; +Cc: Bruce Richardson
Rework the argc/argv handling in pdump a little to improve things.
Issues with the original implementation:
* assumed user would never pass proc-type parameter on the commandline
* did not null-terminate the argv array (generally harmless but not in
spec)
* did not handle case where arg processing in eal_init would reorder
non-flag args to the end to be handled by the app.
Fix these - all-be-it minor issues, by having a separate count value for
the number of arguments we put in the argp array, rather than assuming
that its argc + 1 (for proc-type flag). Properly set the last argv entry
to NULL, and when processing app args, reuse the argp array passed to
eal_init rather than reverting back to the original argv array.
Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
---
app/pdump/main.c | 27 ++++++++++++++-------------
1 file changed, 14 insertions(+), 13 deletions(-)
diff --git a/app/pdump/main.c b/app/pdump/main.c
index f3ac81cb80..89347d70dc 100644
--- a/app/pdump/main.c
+++ b/app/pdump/main.c
@@ -983,7 +983,8 @@ main(int argc, char **argv)
int i;
char mp_flag[] = "--proc-type=secondary";
- char *argp[argc + 1];
+ char *argp[argc + 2]; /* add proc-type, and final NULL entry */
+ int n_argp = 0;
/* catch ctrl-c so we can cleanup on exit */
sigemptyset(&action.sa_mask);
@@ -994,27 +995,27 @@ main(int argc, char **argv)
if (origaction.sa_handler == SIG_DFL)
sigaction(SIGHUP, &action, NULL);
- argp[0] = argv[0];
- argp[1] = mp_flag;
+ argp[n_argp++] = argv[0];
+ argp[n_argp++] = mp_flag;
- for (i = 1; i < argc; i++)
- argp[i + 1] = argv[i];
-
- argc += 1;
+ for (i = 1; i < argc; i++) {
+ argp[n_argp] = argv[i];
+ /* drop any user-provided proc-type to avoid dup flags */
+ if (strncmp(argv[i], mp_flag, strlen("--proc-type")) != 0)
+ n_argp++;
+ }
+ argp[n_argp] = NULL;
- diag = rte_eal_init(argc, argp);
+ diag = rte_eal_init(n_argp, argp);
if (diag < 0)
rte_panic("Cannot init EAL\n");
if (rte_eth_dev_count_avail() == 0)
rte_exit(EXIT_FAILURE, "No Ethernet ports - bye\n");
- argc -= diag;
- argv += (diag - 1);
-
/* parse app arguments */
- if (argc > 1) {
- ret = launch_args_parse(argc, argv, argp[0]);
+ if (n_argp - diag > 1) {
+ ret = launch_args_parse(n_argp - diag, argp + diag, argp[0]);
if (ret < 0)
rte_exit(EXIT_FAILURE, "Invalid argument\n");
}
--
2.48.1
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg
2025-10-17 11:06 [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg Bruce Richardson
2025-10-17 11:06 ` [PATCH 2/2] app/pdump: improve cmdline argument handling Bruce Richardson
@ 2025-11-07 16:10 ` Burakov, Anatoly
1 sibling, 0 replies; 4+ messages in thread
From: Burakov, Anatoly @ 2025-11-07 16:10 UTC (permalink / raw)
To: Bruce Richardson, dev; +Cc: stable
On 10/17/2025 1:06 PM, Bruce Richardson wrote:
> The pdump main function modifies argc/argv before calling EAL init
> function. It needs to do so in order to add the proc-type=secondary
> flag, but also adds in "-n4" as parameter. This parameter is unnecessary
> and also causes duplicate parameter errors if the user also passes in -n
> flag on the commandline. Just remove this flag.
>
> Fixes: caa7028276b8 ("app/pdump: add tool for packet capturing")
> Cc: stable@dpdk.org
>
> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
--
Thanks,
Anatoly
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH 2/2] app/pdump: improve cmdline argument handling
2025-10-17 11:06 ` [PATCH 2/2] app/pdump: improve cmdline argument handling Bruce Richardson
@ 2025-11-07 16:14 ` Burakov, Anatoly
0 siblings, 0 replies; 4+ messages in thread
From: Burakov, Anatoly @ 2025-11-07 16:14 UTC (permalink / raw)
To: Bruce Richardson, dev
On 10/17/2025 1:06 PM, Bruce Richardson wrote:
> Rework the argc/argv handling in pdump a little to improve things.
> Issues with the original implementation:
> * assumed user would never pass proc-type parameter on the commandline
> * did not null-terminate the argv array (generally harmless but not in
> spec)
> * did not handle case where arg processing in eal_init would reorder
> non-flag args to the end to be handled by the app.
>
> Fix these - all-be-it minor issues, by having a separate count value for
> the number of arguments we put in the argp array, rather than assuming
> that its argc + 1 (for proc-type flag). Properly set the last argv entry
> to NULL, and when processing app args, reuse the argp array passed to
> eal_init rather than reverting back to the original argv array.
>
> Signed-off-by: Bruce Richardson <bruce.richardson@intel.com>
> ---
<snip>
>
> - argp[0] = argv[0];
> - argp[1] = mp_flag;
> + argp[n_argp++] = argv[0];
> + argp[n_argp++] = mp_flag;
>
> - for (i = 1; i < argc; i++)
> - argp[i + 1] = argv[i];
> -
> - argc += 1;
> + for (i = 1; i < argc; i++) {
> + argp[n_argp] = argv[i];
> + /* drop any user-provided proc-type to avoid dup flags */
> + if (strncmp(argv[i], mp_flag, strlen("--proc-type")) != 0)
> + n_argp++;
Feels a little odd to assign the value and then overwrite it in next
iteration, but same end result so no biggie
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
--
Thanks,
Anatoly
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2025-11-07 16:14 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-10-17 11:06 [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg Bruce Richardson
2025-10-17 11:06 ` [PATCH 2/2] app/pdump: improve cmdline argument handling Bruce Richardson
2025-11-07 16:14 ` Burakov, Anatoly
2025-11-07 16:10 ` [PATCH 1/2] app/pdump: remove hard-coded memory channels EAL arg Burakov, Anatoly
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).