netdev - Re: [PATCH v2 net-next 2/3] net: phy: ensure phylib state machine is stopped after calling phy

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: <76ba0e6e-ead4-31f8-1235-d7bc59e95e16@gmail.com>
Date:   Thu, 17 Jan 2019 07:10:35 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     Florian Fainelli <f.fainelli@...il.com>,
        David Miller <davem@...emloft.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH v2 net-next 2/3] net: phy: ensure phylib state machine is
 stopped after calling phy_stop

On 16.01.2019 22:59, Andrew Lunn wrote:
> On Wed, Jan 16, 2019 at 09:25:54PM +0100, Heiner Kallweit wrote:
>> The call to the phylib state machine in phy_stop() just ensures that
>> the state machine isn't re-triggered, but a state machine call may
>> be scheduled already. So lets's call phy_stop_machine().
>> This also allows to get rid of the call to phy_stop_machine() in
>> phy_disconnect().
>>
>> Signed-off-by: Heiner Kallweit <hkallweit1@...il.com>
> 
> Reviewed-by: Andrew Lunn <andrew@...n.ch>
> 
> At some point it would be good to audit the code and see if anything
> is still used in interrupt context. PHY interrupt handling is now done
> in a threaded interrupt handler. So i think it is just callers to
> phy_mac_interrupt() that need to be checked. It could be the work
> queue can be removed and everything done synchronous.
> 
Yes, I think few calls to phy_trigger_machine() could be changed to be
synchronous. But I think the workqueue will still  be needed:
- for phy_mac_interrupt() which is typically called from hard irq
  context
- for polling link status (delayed work)

>     Andrew
> 
Heiner