[lvc-project] [PATCH RFC net] net: sockmap: avoid race between sock_map_destroy() and sk_psock_put()

Cong Wang xiyou.wangcong at gmail.com
Wed Sep 11 19:45:54 MSK 2024


On Wed, Sep 11, 2024 at 12:51:04PM +0300, Dmitry Antipov wrote:
> On 9/11/24 7:32 AM, Cong Wang wrote:
> 
> > I never tested the RDS code (hence why I didn't post it). But for the warning
> > itself, actually disabling CONFIG_RDS made it disappear on my side, yet
> > another reason why I suspect it is RDS related.
> 
> OTOH sockmap code depends from CONFIG_BPF_SYSCALL. So I'm pretty sure that
> there are more sockmap users beyond RDS and turning off CONFIG_RDS by itself
> is not too useful for further investigations of this case.
> 

I guess you totally misunderstand my point. As a significant sockmap
contributor, I am certainly aware of sockmap users. My point is that I
needed to narrow down the problem to CONFIG_RDS when I was debugging it.

So, please let me know if you can still reproduce this after disabling
CONFIG_RDS, because I could not reproduce it any more. If you can,
please kindly share the stack trace without rds_* functions.

Thanks.



More information about the lvc-project mailing list