From: Jakub Kicinski <jakub.kicinski@netronome.com>
To: Jiong Wang <jiong.wang@netronome.com>
Cc: Alexei Starovoitov <alexei.starovoitov@gmail.com>,
daniel@iogearbox.net, bpf@vger.kernel.org,
netdev@vger.kernel.org, oss-drivers@netronome.com
Subject: Re: [oss-drivers] Re: [PATCH/RFC v2 bpf-next 05/19] bpf: split read liveness into REG_LIVE_READ64 and REG_LIVE_READ32
Date: Thu, 11 Apr 2019 09:44:56 -0700 [thread overview]
Message-ID: <20190411094456.1fabc52d@cakuba.netronome.com> (raw)
In-Reply-To: <3A9F4205-577C-4E64-8400-0D476F08459E@netronome.com>
On Thu, 11 Apr 2019 07:13:03 +0100, Jiong Wang wrote:
> >> @@ -1150,17 +1150,17 @@ static int mark_reg_read(struct bpf_verifier_env *env,
> >> parent->var_off.value, parent->off);
> >> return -EFAULT;
> >> }
> >> - if (parent->live & REG_LIVE_READ)
> >> + if ((parent->live & REG_LIVE_READ) == flags)
> >> /* The parentage chain never changes and
> >> - * this parent was already marked as LIVE_READ.
> >> + * this parent was already marked with all read bits.
> >
> > Do we have to propagate all read bits? Read64 is strictly stronger
> > than read32, as long as read64 is set on the parent we should be good?
>
> We should be good, but I doubt there is value to differentiate on this in this
> kind of HOT function.
The entire if clause is an optimization. I'm saying you can maintain it
as more aggressive.
> >> @@ -6196,12 +6286,19 @@ static int propagate_liveness_reg(struct bpf_verifier_env *env,
> >> struct bpf_reg_state *reg,
> >> struct bpf_reg_state *parent_reg)
> >> {
> >> + u8 parent_bits = parent_reg->live & REG_LIVE_READ;
> >> + u8 bits = reg->live & REG_LIVE_READ;
> >> + u8 bits_diff = parent_bits ^ bits;
> >> + u8 bits_prop = bits_diff & bits;
> >> int err;
> >>
> >> - if (parent_reg->live & REG_LIVE_READ || !(reg->live & REG_LIVE_READ))
> >> + /* "reg" and "parent_reg" has the same read bits, or the bit doesn't
> >> + * belong to "reg".
> >> + */
> >> + if (!bits_diff || !bits_prop)
> >
> > bits_prop is a subset of bits_diff, no? !bits_prop is always true
> > if !bits_diff is true, no need to check both.
>
> Bits_prop is a subset of bits_diff WHEN it comes from “reg", we don’t want to
> do the propagation when the diff comes from “parent_reg”, so, we need to check
> both.
Not sure what you're saying, in this patch:
u8 bits_prop = bits_diff & bits;
Maybe you're talking about some patch down the line..
next prev parent reply other threads:[~2019-04-11 16:45 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-04-10 19:50 [PATCH/RFC v2 bpf-next 00/19] bpf: eliminate zero extensions for sub-register writes Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 01/19] bpf: refactor propagate_liveness to eliminate duplicated for loop Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 02/19] bpf: refactor propagate_liveness to eliminate code redundance Jiong Wang
2019-04-11 2:39 ` [oss-drivers] " Jakub Kicinski
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 03/19] bpf: factor out reg and stack slot propagation into "propagate_liveness_reg" Jiong Wang
2019-04-11 2:39 ` [oss-drivers] " Jakub Kicinski
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 04/19] bpf: refactor "check_reg_arg" to eliminate code redundancy Jiong Wang
2019-04-11 2:40 ` Jakub Kicinski
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 05/19] bpf: split read liveness into REG_LIVE_READ64 and REG_LIVE_READ32 Jiong Wang
2019-04-11 2:52 ` Jakub Kicinski
2019-04-11 6:13 ` Jiong Wang
2019-04-11 16:44 ` Jakub Kicinski [this message]
2019-04-11 16:53 ` [oss-drivers] " Jiong Wang
2019-04-12 16:14 ` Jiong Wang
2019-04-11 17:22 ` Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 06/19] bpf: mark lo32 writes that should be zero extended into hi32 Jiong Wang
2019-04-11 3:13 ` Jakub Kicinski
2019-04-11 6:02 ` Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 07/19] bpf: reduce false alarm by refining helper call arg types Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 08/19] bpf: insert explicit zero extension insn when hardware doesn't do it implicitly Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 09/19] bpf: introduce new bpf prog load flags "BPF_F_TEST_RND_HI32" Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 10/19] bpf: randomize high 32-bit when BPF_F_TEST_RND_HI32 is set Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 11/19] libbpf: new global variable "libbpf_test_mode" Jiong Wang
2019-04-11 3:19 ` Jakub Kicinski
2019-04-11 14:32 ` Jiong Wang
2019-04-11 21:49 ` Jiong Wang
2019-04-12 22:08 ` Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 12/19] selftests: enable hi32 randomization for "test_progs" and "test_verifier" Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 13/19] arm: bpf: eliminate zero extension code-gen Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 14/19] powerpc: " Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 15/19] s390: " Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 16/19] sparc: " Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 17/19] x32: " Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 18/19] riscv: " Jiong Wang
2019-04-10 19:50 ` [PATCH/RFC v2 bpf-next 19/19] nfp: " Jiong Wang
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=20190411094456.1fabc52d@cakuba.netronome.com \
--to=jakub.kicinski@netronome.com \
--cc=alexei.starovoitov@gmail.com \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=jiong.wang@netronome.com \
--cc=netdev@vger.kernel.org \
--cc=oss-drivers@netronome.com \
/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).