linux-kernel - Re: [PATCH] net: lan78xx: fix "enabled interrupts" warninig

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: <20190408155947.3efcb1eb@xhacker.debian>
Date:   Mon, 8 Apr 2019 08:07:43 +0000
From:   Jisheng Zhang <Jisheng.Zhang@...aptics.com>
To:     "RaghuramChary.Jallipalli@...rochip.com" 
        <RaghuramChary.Jallipalli@...rochip.com>,
        Marc Zyngier <marc.zyngier@....com>
CC:     "Woojung.Huh@...rochip.com" <Woojung.Huh@...rochip.com>,
        "UNGLinuxDriver@...rochip.com" <UNGLinuxDriver@...rochip.com>,
        "davem@...emloft.net" <davem@...emloft.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] net: lan78xx: fix "enabled interrupts" warninig

On Mon, 8 Apr 2019 07:46:03 +0000 wrote:

> 
> 
> >
> > Per my understanding, the proper handling of PHY irq is to make use of
> > PHY_IGNORE_INTERRUPT then call phy_mac_interrupt when
> > INT_ENP_PHY_INT is triggered.
> >  
> 
> Hi Jisheng,

Hi

> Thanks for the changes.
> Is this warning specific to any linux version? 

In theory, no. I only tested 5.0, 4.20, both can reproduce this warning.

> Why do you think PHY irq domain handling is not proper?

+ Marc

The warning comes from calling generic_handle_irq() in usb tasklet context.
This is not correct.

Per my understanding, if there's chained irq, we could introduce extra
irqdomain. E.g

GIC <--> another irqchip controller <--> HW device

But in lan78xx, this is not the case. There's no chained irq at all.
In fact, there's a bit INT_ENP_PHY_INT in MAC's Interrupt Endpoint status
word to indicate this is PHY interrupt. So this is the case
PHY_IGNORE_INTERRUPT/phy_mac_interrupt intend for.

irq experts knows irqdomain etc better, maybe they can provide more inputs

> Maybe we can fix that rather removing complete IRQ domain code changes.
> Also, Can you pls let us know how this changes fixed your warning.

The patch removes the generic_handle_irq() calling from invalid context.

Thanks