* There has a backport bug between v5.10.79 and v5.10.80 when run bpf selftest "test_sockmap" on 5.10 lts kernel
@ 2024-02-02 3:40 Lucien Wang
2024-02-02 15:03 ` Greg KH
0 siblings, 1 reply; 3+ messages in thread
From: Lucien Wang @ 2024-02-02 3:40 UTC (permalink / raw)
To: bpf; +Cc: ast, daniel, john.fastabend, gregkh
Kernel version:16ad71c250c1 (HEAD -> linux-5.10.y, tag: v5.10.209,
origin/linux-5.10.y) Linux 5.10.209
Bug reproduced steps:
1. cd (kernel source tree root)/tools/testing/selftests/bpf
2. make test_sockmap ; make test_progs
3. ./test_sockmap
# 1/ 6 sockmap::txmsg test passthrough:OK
# 2/ 6 sockmap::txmsg test redirect:OK
# 3/ 6 sockmap::txmsg test drop:OK
# 4/ 6 sockmap::txmsg test ingress redirect:OK
After "# 4/ 6 sockmap::txmsg test ingress redirect:OK" display from
terminal, the main process stucks and sends nothing.
4. In other terminal run " ps fax |grep sockmap " ,below is output
13076 pts/0 S+ 0:00 | \_ ./test_sockmap
13129 pts/0 S+ 0:00 | \_ ./test_sockmap
13130 pts/0 Z+ 0:00 | \_ [test_sockmap] <defunct>
13237 pts/1 S+ 0:00 \_ grep --color=auto sockmap
Obversely, because of child process 13129 sleep, so the main process is stuck.
My research:
I use Bisection method to find the bug patch " c842a4c4ae7f bpf:
sockmap, strparser, and tls are reusing qdisc_skb_cb and colliding
"(on linux-5.10.y branch), it backport from v5.16-rc1 ,
It must due to merge high patches incompletely, Please take a few
moment for this.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: There has a backport bug between v5.10.79 and v5.10.80 when run bpf selftest "test_sockmap" on 5.10 lts kernel
2024-02-02 3:40 There has a backport bug between v5.10.79 and v5.10.80 when run bpf selftest "test_sockmap" on 5.10 lts kernel Lucien Wang
@ 2024-02-02 15:03 ` Greg KH
2024-02-04 0:35 ` Lucien Wang
0 siblings, 1 reply; 3+ messages in thread
From: Greg KH @ 2024-02-02 15:03 UTC (permalink / raw)
To: Lucien Wang; +Cc: bpf, ast, daniel, john.fastabend
On Fri, Feb 02, 2024 at 11:40:38AM +0800, Lucien Wang wrote:
> Kernel version:16ad71c250c1 (HEAD -> linux-5.10.y, tag: v5.10.209,
> origin/linux-5.10.y) Linux 5.10.209
>
> Bug reproduced steps:
> 1. cd (kernel source tree root)/tools/testing/selftests/bpf
> 2. make test_sockmap ; make test_progs
> 3. ./test_sockmap
> # 1/ 6 sockmap::txmsg test passthrough:OK
> # 2/ 6 sockmap::txmsg test redirect:OK
> # 3/ 6 sockmap::txmsg test drop:OK
> # 4/ 6 sockmap::txmsg test ingress redirect:OK
>
> After "# 4/ 6 sockmap::txmsg test ingress redirect:OK" display from
> terminal, the main process stucks and sends nothing.
> 4. In other terminal run " ps fax |grep sockmap " ,below is output
> 13076 pts/0 S+ 0:00 | \_ ./test_sockmap
> 13129 pts/0 S+ 0:00 | \_ ./test_sockmap
> 13130 pts/0 Z+ 0:00 | \_ [test_sockmap] <defunct>
> 13237 pts/1 S+ 0:00 \_ grep --color=auto sockmap
> Obversely, because of child process 13129 sleep, so the main process is stuck.
>
> My research:
> I use Bisection method to find the bug patch " c842a4c4ae7f bpf:
> sockmap, strparser, and tls are reusing qdisc_skb_cb and colliding
> "(on linux-5.10.y branch), it backport from v5.16-rc1 ,
> It must due to merge high patches incompletely, Please take a few
> moment for this.
I do not understand, sorry, what exactly do you want us to do here?
confused,
greg k-h
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: There has a backport bug between v5.10.79 and v5.10.80 when run bpf selftest "test_sockmap" on 5.10 lts kernel
2024-02-02 15:03 ` Greg KH
@ 2024-02-04 0:35 ` Lucien Wang
0 siblings, 0 replies; 3+ messages in thread
From: Lucien Wang @ 2024-02-04 0:35 UTC (permalink / raw)
To: Greg KH; +Cc: bpf, ast, daniel, john.fastabend
There has a bug when I run bpf selftest "test_sockmap" on Linux
5.10.209, and I found it introduced by "c842a4c4ae7f bpf: sockmap,
strparser, and tls are reusing qdisc_skb_cb and colliding".
On Fri, Feb 2, 2024 at 11:03 PM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> On Fri, Feb 02, 2024 at 11:40:38AM +0800, Lucien Wang wrote:
> > Kernel version:16ad71c250c1 (HEAD -> linux-5.10.y, tag: v5.10.209,
> > origin/linux-5.10.y) Linux 5.10.209
> >
> > Bug reproduced steps:
> > 1. cd (kernel source tree root)/tools/testing/selftests/bpf
> > 2. make test_sockmap ; make test_progs
> > 3. ./test_sockmap
> > # 1/ 6 sockmap::txmsg test passthrough:OK
> > # 2/ 6 sockmap::txmsg test redirect:OK
> > # 3/ 6 sockmap::txmsg test drop:OK
> > # 4/ 6 sockmap::txmsg test ingress redirect:OK
> >
> > After "# 4/ 6 sockmap::txmsg test ingress redirect:OK" display from
> > terminal, the main process stucks and sends nothing.
> > 4. In other terminal run " ps fax |grep sockmap " ,below is output
> > 13076 pts/0 S+ 0:00 | \_ ./test_sockmap
> > 13129 pts/0 S+ 0:00 | \_ ./test_sockmap
> > 13130 pts/0 Z+ 0:00 | \_ [test_sockmap] <defunct>
> > 13237 pts/1 S+ 0:00 \_ grep --color=auto sockmap
> > Obversely, because of child process 13129 sleep, so the main process is stuck.
> >
> > My research:
> > I use Bisection method to find the bug patch " c842a4c4ae7f bpf:
> > sockmap, strparser, and tls are reusing qdisc_skb_cb and colliding
> > "(on linux-5.10.y branch), it backport from v5.16-rc1 ,
> > It must due to merge high patches incompletely, Please take a few
> > moment for this.
>
> I do not understand, sorry, what exactly do you want us to do here?
>
> confused,
>
> greg k-h
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2024-02-04 0:36 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-02-02 3:40 There has a backport bug between v5.10.79 and v5.10.80 when run bpf selftest "test_sockmap" on 5.10 lts kernel Lucien Wang
2024-02-02 15:03 ` Greg KH
2024-02-04 0:35 ` Lucien Wang
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox