From: Jiong Wang <jiong.wang@netronome.com>
To: Jiong Wang <jiong.wang@netronome.com>
Cc: Alexei Starovoitov <alexei.starovoitov@gmail.com>,
"Naveen N. Rao" <naveen.n.rao@linux.vnet.ibm.com>,
Daniel Borkmann <daniel@iogearbox.net>, bpf <bpf@vger.kernel.org>,
Network Development <netdev@vger.kernel.org>,
Michael Ellerman <mpe@ellerman.id.au>,
Jakub Kicinski <jakub.kicinski@netronome.com>
Subject: Re: [PATCH] bpf: optimize constant blinding
Date: Wed, 12 Jun 2019 16:25:48 +0100 [thread overview]
Message-ID: <87r27y25c3.fsf@netronome.com> (raw)
In-Reply-To: <87sgse26av.fsf@netronome.com>
Jiong Wang writes:
> Alexei Starovoitov writes:
>
>> On Wed, Jun 12, 2019 at 4:32 AM Naveen N. Rao
>> <naveen.n.rao@linux.vnet.ibm.com> wrote:
>>>
>>> Currently, for constant blinding, we re-allocate the bpf program to
>>> account for its new size and adjust all branches to accommodate the
>>> same, for each BPF instruction that needs constant blinding. This is
>>> inefficient and can lead to soft lockup with sufficiently large
>>> programs, such as the new verifier scalability test (ld_dw: xor
>>> semi-random 64 bit imms, test 5 -- with net.core.bpf_jit_harden=2)
>>
>> Slowdown you see is due to patch_insn right?
>> In such case I prefer to fix the scaling issue of patch_insn instead.
>> This specific fix for blinding only is not addressing the core of the problem.
>> Jiong,
>> how is the progress on fixing patch_insn?
And what I have done is I have digested your conversion with Edward, and is
slightly incline to the BB based approach as it also exposes the inserted
insn to later pass in a natural way, then was trying to find a way that
could create BB info in little extra code based on current verifier code,
for example as a side effect of check_subprogs which is doing two insn
traversal already. (I had some such code before in the historical
wip/bpf-loop-detection branch, but feel it might be still too heavy for
just improving insn patching)
>
> I actually was about to reply this email as we have discussed exactly the
> same issue on jit blinding here:
>
> https://www.spinics.net/lists/bpf/msg01836.html
>
> And sorry for the slow progress on fixing patch_insn, please give me one
> more week, I will try to send out a RFC for it.
>
> Regards,
> Jiong
next prev parent reply other threads:[~2019-06-12 15:25 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-06-12 11:32 [PATCH] bpf: optimize constant blinding Naveen N. Rao
2019-06-12 14:47 ` Alexei Starovoitov
2019-06-12 15:04 ` Jiong Wang
2019-06-12 15:25 ` Jiong Wang [this message]
2019-06-12 15:28 ` Alexei Starovoitov
2019-06-14 15:13 ` Jiong Wang
2019-06-14 17:05 ` Alexei Starovoitov
2019-06-14 22:28 ` Andrii Nakryiko
2019-06-17 19:47 ` Edward Cree
2019-06-17 19:59 ` Jiong Wang
2019-06-17 20:11 ` Edward Cree
2019-06-17 20:40 ` Jiong Wang
2019-06-17 20:53 ` Alexei Starovoitov
2019-06-17 21:01 ` Jiong Wang
2019-06-17 21:16 ` Edward Cree
2019-06-19 20:45 ` Jiong Wang
2019-06-14 4:30 ` kbuild test robot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87r27y25c3.fsf@netronome.com \
--to=jiong.wang@netronome.com \
--cc=alexei.starovoitov@gmail.com \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=jakub.kicinski@netronome.com \
--cc=mpe@ellerman.id.au \
--cc=naveen.n.rao@linux.vnet.ibm.com \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).