On Wed, 15 Nov 2023 12:23:13 +0100 Morten Brørup <[email protected]> wrote:
> > From: Stephen Hemminger [mailto:[email protected]] > > Sent: Wednesday, 15 November 2023 01.23 > > > > On Tue, 14 Nov 2023 16:24:58 +0800 > > Dengdui Huang <[email protected]> wrote: > > > > > The function strerror() is insecure in a multi-thread environment. > > > This patch uses rte_strerror() to replace it. > > > > > > Cc: [email protected] > > > > > > Signed-off-by: Dengdui Huang <[email protected]> > > > Acked-by: Chengwen Feng <[email protected]> > > > --- > > > > Look at implementation of rte_strerror(), it still has issues if > > used from non-DPDK registered threads. > > The macro names RTE_DEFINE_PER_LCORE() and RTE_PER_LCORE() strongly indicate > that the variables are instantiated per lcore. > > But they are not; they are instantiated per thread [1]: So rte_strerror() is thread safe (good). It make sense to use it, take back my original negative comments. We could improve documentation of rte_strerror() to help.
