Hello,
On Tue, Jul 4, 2023 at 11:49 PM Mattias Rönnblom
<mattias.ronnb...@ericsson.com> wrote:
>
> Calling rte_exit() from a worker lcore thread causes a deadlock in
> rte_service_finalize().
>
> This patch makes rte_service_finalize() deadlock-free by avoiding the
> need to synchronize with service lcore threads, which in turn is
> achieved by moving service and per-lcore state from the heap to being
> statically allocated.
>
> The BSS segment increases with ~156 kB (on x86_64 with default
> RTE_MAX_LCORE and RTE_SERVICE_NUM_MAX).
>
> According to the service perf autotest, this change also results in a
> slight reduction of service framework overhead.
>
> Fixes: 33666b448f15 ("service: fix crash on exit")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Mattias Rönnblom <mattias.ronnb...@ericsson.com>
> Acked-by: Harry van Haaren <harry.van.haa...@intel.com>
Sorry, this fell through the cracks.
CI was complaining about this patch and it was probably a false
positive, but we don't have the (full) logs anymore.
Could you resend it please?
--
David Marchand
