On Fri, 21 Dec 2018 17:24:01 -0800, Alexei Starovoitov wrote:
> On Fri, Dec 21, 2018 at 03:46:32PM -0800, Jakub Kicinski wrote:
> > On Thu, 20 Dec 2018 07:19:06 +, Yonghong Song wrote:
> > > > I think this will break func_info, since it's not adjusted here.
> > > > Also iirc line_info is relyin
On Fri, Dec 21, 2018 at 03:46:32PM -0800, Jakub Kicinski wrote:
> On Thu, 20 Dec 2018 07:19:06 +, Yonghong Song wrote:
> > > I think this will break func_info, since it's not adjusted here.
> > > Also iirc line_info is relying on first insn to always have line_info.
> > > If first insn is dead,
On Fri, Dec 21, 2018 at 03:46:32PM -0800, Jakub Kicinski wrote:
> I want to also
> tackle the JIT linfo offsets for offloads while at it and post an RFC
> (unless you're handling that already Martin?)
No. I am not :)
On Thu, 20 Dec 2018 07:19:06 +, Yonghong Song wrote:
> > I think this will break func_info, since it's not adjusted here.
> > Also iirc line_info is relying on first insn to always have line_info.
> > If first insn is dead, second insn might not have a line_info generated
> > and things won't g
On 12/19/18 4:45 PM, Alexei Starovoitov wrote:
> On Wed, Dec 19, 2018 at 10:29:18AM -0800, Jakub Kicinski wrote:
>> Instead of overwriting dead code with jmp -1 instructions
>> remove it completely for root. Adjust verifier state and
>> line info appropriately.
>>
>> Signed-off-by: Jakub Kicinsk
On Wed, Dec 19, 2018 at 10:29:18AM -0800, Jakub Kicinski wrote:
> Instead of overwriting dead code with jmp -1 instructions
> remove it completely for root. Adjust verifier state and
> line info appropriately.
>
> Signed-off-by: Jakub Kicinski
> ---
> include/linux/filter.h | 1 +
> kernel/bp
Instead of overwriting dead code with jmp -1 instructions
remove it completely for root. Adjust verifier state and
line info appropriately.
Signed-off-by: Jakub Kicinski
---
include/linux/filter.h | 1 +
kernel/bpf/core.c | 12
kernel/bpf/verifier.c | 139 +
