From: David Ahern <dsa@cumulusnetworks•com>
To: Alexander Duyck <alexander.duyck@gmail•com>, netdev@vger•kernel.org
Subject: Re: [PATCH net-next 2/9] net: Remove e_inval label from ip_route_input_slow
Date: Wed, 23 Sep 2015 10:02:31 -0600 [thread overview]
Message-ID: <5602CD17.3030008@cumulusnetworks.com> (raw)
In-Reply-To: <5602C901.2050404@gmail.com>
On 9/23/15 9:45 AM, Alexander Duyck wrote:
> On 09/23/2015 08:15 AM, David Ahern wrote:
>> e_inval has 1 explicit user and 1 fallthrough user --
>> martian_destination.
>> Move setting err to EINVAL before the 2 users and use the goto out label
>> instead of e_inval.
>>
>> Signed-off-by: David Ahern <dsa@cumulusnetworks•com>
>> ---
>> net/ipv4/route.c | 7 +++----
>> 1 file changed, 3 insertions(+), 4 deletions(-)
>>
>> diff --git a/net/ipv4/route.c b/net/ipv4/route.c
>> index ef36dfed24da..3c5000db5972 100644
>> --- a/net/ipv4/route.c
>> +++ b/net/ipv4/route.c
>> @@ -1775,8 +1775,9 @@ static int ip_route_input_slow(struct sk_buff
>> *skb, __be32 daddr, __be32 saddr,
>> out: return err;
>> brd_input:
>> + err = -EINVAL;
>> if (skb->protocol != htons(ETH_P_IP))
>> - goto e_inval;
>> + goto out;
>> if (!ipv4_is_zeronet(saddr)) {
>> err = fib_validate_source(skb, saddr, 0, tos, 0, dev,
>> @@ -1842,15 +1843,13 @@ out: return err;
>> * Do not cache martian addresses: they should be logged
>> (RFC1812)
>> */
>> martian_destination:
>> + err = -EINVAL;
>> RT_CACHE_STAT_INC(in_martian_dst);
>> #ifdef CONFIG_IP_ROUTE_VERBOSE
>> if (IN_DEV_LOG_MARTIANS(in_dev))
>> net_warn_ratelimited("martian destination %pI4 from %pI4,
>> dev %s\n",
>> &daddr, &saddr, dev->name);
>> #endif
>> -
>> -e_inval:
>> - err = -EINVAL;
>> goto out;
>> e_nobufs:
>
> This is adding unnecessary bloat to the code. I really think if you
> want to simplify this just replace "goto e_inval;" with "return -EINVAL;".
Not sure why you think this is bloating the code. Before this patch:
$ size kbuild/net/ipv4/route.o
text data bss dec hex filename
19778 2321 32 22131 5673 kbuild/net/ipv4/route.o
With this patch:
$ size kbuild/net/ipv4/route.o
text data bss dec hex filename
19778 2321 32 22131 5673 kbuild/net/ipv4/route.o
So no difference at all.
Please consider this is an intermediate step; the code goes away in
later ones.
>
> Also why is it you end up leaving the out label through all of these
> patches? It seems like that would be one of the first places to start
> in terms of removing the labels.
jump to return is normal usage. Why would I start there and end up with
N exit points? There are a lot of functions in the route code where
doing that will burn you.
next prev parent reply other threads:[~2015-09-23 16:02 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-23 15:15 [PATCH net-next 0/9 v2] net: Refactor ip_route_input_slow David Ahern
2015-09-23 15:15 ` [PATCH net-next 1/9] net: Remove martian_source_keep_err goto label David Ahern
2015-09-23 16:19 ` Tom Herbert
2015-09-23 16:22 ` David Ahern
2015-09-23 15:15 ` [PATCH net-next 2/9] net: Remove e_inval label from ip_route_input_slow David Ahern
2015-09-23 15:45 ` Alexander Duyck
2015-09-23 16:02 ` David Ahern [this message]
2015-09-23 16:31 ` Alexander Duyck
2015-09-23 17:13 ` David Ahern
2015-09-23 17:38 ` Alexander Duyck
2015-09-23 18:03 ` David Ahern
2015-09-23 18:19 ` David Miller
2015-09-23 22:26 ` Alexander Duyck
2015-09-23 18:14 ` David Miller
2015-09-23 15:15 ` [PATCH net-next 3/9] net: Remove e_nobufs " David Ahern
2015-09-23 15:15 ` [PATCH net-next 4/9] net: Move rth handling from ip_route_input_slow to helper David Ahern
2015-09-23 15:15 ` [PATCH net-next 5/9] net: Move martian_destination " David Ahern
2015-09-23 15:15 ` [PATCH net-next 6/9] net: Remove martian_source goto David Ahern
2015-09-23 15:15 ` [PATCH net-next 7/9] net: Remove martian_destination label David Ahern
2015-09-23 17:27 ` Eric Dumazet
2015-09-23 15:15 ` [PATCH net-next 8/9] net: Remove local_input label David Ahern
2015-09-23 15:15 ` [PATCH net-next 9/9] net: Remove no_route label David Ahern
2015-09-23 15:30 ` [PATCH net-next 0/9 v2] net: Refactor ip_route_input_slow Eric Dumazet
2015-09-23 16:08 ` David Ahern
2015-09-23 17:33 ` Eric Dumazet
-- strict thread matches above, loose matches on Subject: below --
2015-09-22 22:55 [PATCH net-next 0/9] " David Ahern
2015-09-22 22:55 ` [PATCH net-next 2/9] net: Remove e_inval label from ip_route_input_slow David Ahern
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=5602CD17.3030008@cumulusnetworks.com \
--to=dsa@cumulusnetworks$(echo .)com \
--cc=alexander.duyck@gmail$(echo .)com \
--cc=netdev@vger$(echo .)kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
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