public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: "Eelco Chaudron" <echaudro@redhat.com>
To: "Jakub Kicinski" <kuba@kernel.org>
Cc: netdev@vger.kernel.org, davem@davemloft.net, dev@openvswitch.org,
	pabeni@redhat.com, pshelar@ovn.org, fw@strlen.de
Subject: Re: [PATCH net-next v2 2/2] net: openvswitch: make masks cache size configurable
Date: Fri, 24 Jul 2020 10:00:16 +0200	[thread overview]
Message-ID: <5A55F49F-24E5-4149-A212-8F00589D6777@redhat.com> (raw)
In-Reply-To: <20200723094236.04d82921@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>



On 23 Jul 2020, at 18:42, Jakub Kicinski wrote:

> On Thu, 23 Jul 2020 14:58:31 +0200 Eelco Chaudron wrote:
>> +	if ((size & (size - 1)) != 0 ||
>
> is_power_of_2() ?

Was not aware of this macro, will replace…
>
>> +	    (size * sizeof(struct mask_cache_entry)) > PCPU_MIN_UNIT_SIZE)
>> +		return NULL;
>
>> +	new->cache_size = size;
>> +	if (new->cache_size > 0) {
>> +		cache = __alloc_percpu(sizeof(struct mask_cache_entry) *
>> +				       new->cache_size,
>
> array_size() ?

Will change to:

		cache = __alloc_percpu(array_size(
					       sizeof(struct mask_cache_entry),
					       new->cache_size),
				       __alignof__(struct mask_cache_entry));
		if (!cache) {
>
>> +				       __alignof__(struct mask_cache_entry));
>> +
>
> No need for the new line here

Will remove

>> +		if (!cache) {
>> +			kfree(new);
>> +			return NULL;
>> +		}
>> +	}
>
>> +	if (size == mc->cache_size || (size & (size - 1)) != 0)
>
> why check "is power of 2" twice?

Will remove as it will always call tbl_mask_cache_alloc() which has the 
check above.

>
>> @@ -454,7 +516,7 @@ void ovs_flow_tbl_destroy(struct flow_table 
>> *table)
>>  	struct table_instance *ti = rcu_dereference_raw(table->ti);
>>  	struct table_instance *ufid_ti = 
>> rcu_dereference_raw(table->ufid_ti);
>>
>> -	free_percpu(table->mask_cache);
>> +	call_rcu(&table->mask_cache->rcu, mask_cache_rcu_cb);
>>  	call_rcu(&table->mask_array->rcu, mask_array_rcu_cb);
>>  	table_instance_destroy(table, ti, ufid_ti, false);
>>  }
>
> This adds a new warning :(
>
> net/openvswitch/flow_table.c:519:24: warning: incorrect type in 
> argument 1 (different address spaces)
> net/openvswitch/flow_table.c:519:24:    expected struct callback_head 
> *head
> net/openvswitch/flow_table.c:519:24:    got struct callback_head 
> [noderef] __rcu *

:( Yes my diff with older did not work… Sure I’ve fixed them all now 
:)


Will sent out a V3 after some testing…


      reply	other threads:[~2020-07-24  8:00 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-23 12:58 [PATCH net-next v2 0/2] net: openvswitch: masks cache enhancements Eelco Chaudron
2020-07-23 12:58 ` [PATCH net-next v2 1/2] net: openvswitch: add masks cache hit counter Eelco Chaudron
2020-07-24  2:37   ` [ovs-dev] " Tonghao Zhang
2020-07-23 12:58 ` [PATCH net-next v2 2/2] net: openvswitch: make masks cache size configurable Eelco Chaudron
2020-07-23 16:42   ` Jakub Kicinski
2020-07-24  8:00     ` Eelco Chaudron [this message]

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=5A55F49F-24E5-4149-A212-8F00589D6777@redhat.com \
    --to=echaudro@redhat.com \
    --cc=davem@davemloft.net \
    --cc=dev@openvswitch.org \
    --cc=fw@strlen.de \
    --cc=kuba@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=pshelar@ovn.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