Re: [bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

On Tue, Jul 28, 2020 at 05:44:19PM -0700, John Fastabend wrote: > Martin KaFai Lau wrote: > > On Tue, Jul 28, 2020 at 01:55:22PM -0700, John Fastabend wrote: > > > Martin KaFai Lau wrote: > > > > On Tue, Jul 28, 2020 at 08:43:46AM -0700, John Fastabend wrote: > > > > > I had a sockmap program that

Re: [bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

Martin KaFai Lau wrote: > On Tue, Jul 28, 2020 at 01:55:22PM -0700, John Fastabend wrote: > > Martin KaFai Lau wrote: > > > On Tue, Jul 28, 2020 at 08:43:46AM -0700, John Fastabend wrote: > > > > I had a sockmap program that after doing some refactoring started > > > > spewing > > > > this splat a

Re: [bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

On Tue, Jul 28, 2020 at 01:55:22PM -0700, John Fastabend wrote: > Martin KaFai Lau wrote: > > On Tue, Jul 28, 2020 at 08:43:46AM -0700, John Fastabend wrote: > > > I had a sockmap program that after doing some refactoring started spewing > > > this splat at me: > > > > > > [18610.807284] BUG: unab

Re: [bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

On 7/28/20 10:55 PM, John Fastabend wrote: Martin KaFai Lau wrote: On Tue, Jul 28, 2020 at 08:43:46AM -0700, John Fastabend wrote: I had a sockmap program that after doing some refactoring started spewing this splat at me: [18610.807284] BUG: unable to handle kernel NULL pointer dereference at

Re: [bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

Martin KaFai Lau wrote: > On Tue, Jul 28, 2020 at 08:43:46AM -0700, John Fastabend wrote: > > I had a sockmap program that after doing some refactoring started spewing > > this splat at me: > > > > [18610.807284] BUG: unable to handle kernel NULL pointer dereference at > > 0001 > > [.

Re: [bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

On Tue, Jul 28, 2020 at 08:43:46AM -0700, John Fastabend wrote: > I had a sockmap program that after doing some refactoring started spewing > this splat at me: > > [18610.807284] BUG: unable to handle kernel NULL pointer dereference at > 0001 > [...] > [18610.807359] Call Trace: > [18

[bpf PATCH 1/3] bpf: sock_ops ctx access may stomp registers in corner case

I had a sockmap program that after doing some refactoring started spewing this splat at me: [18610.807284] BUG: unable to handle kernel NULL pointer dereference at 0001 [...] [18610.807359] Call Trace: [18610.807370] ? 0xc114d0d5 [18610.807382] __cgroup_bpf_run_filter_sock_o