|SUCCESS| pw169219 [PATCH v5 08/17] eal: warn if thread name is truncated

[0-day Robot <robot@bytheb.org>]

Test-Label: 0day Robot - Claude SONNET
Test-Status: SUCCESS
http://dpdk.org/patch/159219

_AI Review Comments Available_

Looking at this patch, I have a few questions about the consistency and 
implementation of the thread name truncation handling.

Git SHA: 20251207191242.165619-9-stephen@networkplumber.org
Author: Stephen Hemminger <stephen@networkplumber.org>
Subject: eal: warn if thread name is truncated

This patch adds warning messages when thread names get truncated due to 
length limits, specifically for the dpdk-worker-%d pattern that can 
overflow with more than 9999 worker cores.

> diff --git a/lib/eal/freebsd/eal.c b/lib/eal/freebsd/eal.c
> index 6215245ad5..beda7e0c5c 100644
> --- a/lib/eal/freebsd/eal.c
> +++ b/lib/eal/freebsd/eal.c
> @@ -690,8 +690,10 @@ rte_eal_init(int argc, char **argv)
>  			rte_panic("Cannot create thread\n");
> 
>  		/* Set thread_name for aid in debugging. */
> -		snprintf(thread_name, sizeof(thread_name),
> -				"dpdk-worker%d", i);
> +		ret = snprintf(thread_name, sizeof(thread_name), "dpdk-worker%d", i);
> +		if (ret >= (int)sizeof(thread_name))
> +			EAL_LOG(INFO, "Worker thread name %s truncated", thread_name);
> +
>  		rte_thread_set_name(lcore_config[i].thread_id, thread_name);

[ ... ]

> diff --git a/lib/eal/linux/eal.c b/lib/eal/linux/eal.c
> index b12f325ddd..d848de03d8 100644
> --- a/lib/eal/linux/eal.c
> +++ b/lib/eal/linux/eal.c
> @@ -863,8 +863,10 @@ rte_eal_init(int argc, char **argv)
>  			rte_panic("Cannot create thread\n");
> 
>  		/* Set thread_name for aid in debugging. */
> -		snprintf(thread_name, sizeof(thread_name),
> -			"dpdk-worker%d", i);
> +		ret = snprintf(thread_name, sizeof(thread_name), "dpdk-worker%d", i);
> +		if (ret >= RTE_THREAD_NAME_SIZE)
> +			EAL_LOG(INFO, "Worker thread name %s truncated", thread_name);
> +
>  		rte_thread_set_name(lcore_config[i].thread_id, thread_name);

[ ... ]

> diff --git a/lib/eal/windows/eal.c b/lib/eal/windows/eal.c
> index 21fe7cb1d9..8b539f8601 100644
> --- a/lib/eal/windows/eal.c
> +++ b/lib/eal/windows/eal.c
> @@ -378,8 +378,10 @@ rte_eal_init(int argc, char **argv)
>  			rte_panic("Cannot create thread\n");
> 
>  		/* Set thread name for aid in debugging. */
> -		snprintf(thread_name, sizeof(thread_name),
> -			"dpdk-worker%d", i);
> +		ret = snprintf(thread_name, sizeof(thread_name), "dpdk-worker%d", i);
> +		if (ret >= (int)sizeof(thread_name))
> +			EAL_LOG(INFO, "Worker thread name %s truncated", thread_name);
> +
>  		rte_thread_set_name(lcore_config[i].thread_id, thread_name);

Why does the Linux implementation check against RTE_THREAD_NAME_SIZE while 
FreeBSD and Windows use sizeof(thread_name)? Should all platforms use the 
same constant for consistency?

Does the logic make sense to log thread_name after truncation? The logged 
name will already be truncated, so it won't show what the intended full 
name was supposed to be.

Could this create confusion in debugging since you're logging the 
truncated name rather than the original intended name?