netdev - Re: [PATCH net-next v2 6/6] enetc: Add adaptive interrupt coalescing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <71e698b5-5e52-44fb-48e6-ed9ce94cd978@gmail.com>
Date:   Sat, 18 Jul 2020 20:20:18 +0300
From:   Claudiu Manoil <claudiu.manoil@...il.com>
To:     Jakub Kicinski <kuba@...nel.org>,
        Claudiu Manoil <claudiu.manoil@....com>
Cc:     "David S . Miller" <davem@...emloft.net>, netdev@...r.kernel.org
Subject: Re: [PATCH net-next v2 6/6] enetc: Add adaptive interrupt coalescing

On 17.07.2020 22:30, Jakub Kicinski wrote:
> On Fri, 17 Jul 2020 18:37:04 +0300 Claudiu Manoil wrote:
>> +	if (tx_ictt == ENETC_TXIC_TIMETHR)
>> +		ic_mode |= ENETC_IC_TX_OPTIMAL;
> 
> Doesn't seem you ever read/check the ENETC_IC_TX_OPTIMAL flag?
> 

It's used implicitly though ;), as it signals a state change when
the user changes the default value of the tx time threshold,
triggering the device recofiguration with the new value. True that
the said reconfiguration could be also performed in the 'MANUAL' state.
I added the extra state called 'OPTIMAL' to make the code more easier to 
follow actually. I mean, it's easy to follow that the tx coalescing 
state starts in the "OPTIMAL" mode, w/ the preconfigured "optimal" 
value. Then if the user changes the value, doing some manual tuning of 
tx-usecs, it moves into the 'MANUAL' mode, returning to the 'OPTIMAL' 
mode if the user goes back to the optimal value.
This handling could also be done in the 'MANUAL' mode alone, so if you 
want me to make this change pls let me know.

Thanks,
Claudiu