linux-kernel - Re: [PATCH i2c-next v6] i2c: aspeed: Handle master/slave combined irq events properly

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]

Date:   Thu, 13 Sep 2018 20:48:09 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Cédric Le Goater <clg@...d.org>
Cc:     Jae Hyun Yoo <jae.hyun.yoo@...ux.intel.com>,
        Joel Stanley <joel@....id.au>, linux-aspeed@...ts.ozlabs.org,
        Vernon Mauery <vernon.mauery@...ux.intel.com>,
        OpenBMC Maillist <openbmc@...ts.ozlabs.org>,
        Brendan Higgins <brendanhiggins@...gle.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        linux-i2c@...r.kernel.org, jarkko.nikula@...ux.intel.com,
        Linux ARM <linux-arm-kernel@...ts.infradead.org>,
        James Feist <james.feist@...ux.intel.com>
Subject: Re: [PATCH i2c-next v6] i2c: aspeed: Handle master/slave combined irq
 events properly

On 09/13/2018 09:35 AM, Cédric Le Goater wrote:
> On 09/13/2018 05:57 PM, Guenter Roeck wrote:
>> On Thu, Sep 13, 2018 at 05:48:59PM +0200, Cédric Le Goater wrote:
>>> On 09/13/2018 03:33 PM, Guenter Roeck wrote:
>> [ ... ]
>>>>>>    /*
>>>>>>     * The state machine needs some refinement. It is only used to track
>>>>>>     * invalid STOP commands for the moment.
>>>>>> @@ -188,7 +215,7 @@ static void aspeed_i2c_bus_handle_cmd(AspeedI2CBus *bus, uint64_t value)
>>>>>>    {
>>>>>>        bus->cmd &= ~0xFFFF;
>>>>>>        bus->cmd |= value & 0xFFFF;
>>>>>> -    bus->intr_status = 0;> +    bus->intr_status &= I2CD_INTR_RX_DONE;
>>>>>
>>>>> it deserves a comment to understand which scenario we are trying to handle.
>>>>>     
>>>>
>>>> Ok. FWIW, I wonder if intr_status should be touched here in the first place,
>>>> but I neither have the hardware nor a datasheet, so I don't know if any bits
>>>> are auto-cleared.
>>>
>>> I just pushed a patch on my branch with some more explanation :
>>>
>>> 	https://github.com/legoater/qemu/commits/aspeed-3.1
>>>
>> That seems to suggest that none of the status bits auto-clears, and that
>> the above code clearing intr_status should be removed entirely.
>> Am I missing something ?
> 
> You are right. I just pushed another version of the previous patch with this
> new hunk :
> 
> @@ -188,7 +200,6 @@ static void aspeed_i2c_bus_handle_cmd(As
>   {
>       bus->cmd &= ~0xFFFF;
>       bus->cmd |= value & 0xFFFF;
> -    bus->intr_status = 0;
>   
>       if (bus->cmd & I2CD_M_START_CMD) {
>           uint8_t state = aspeed_i2c_get_state(bus) & I2CD_MACTIVE ?
> 
> 
> The QEMU palmetto and witherspoon machines seem to behave fine. Can you give
> it a try ?
> 

Works fine for me for all affected qemu platforms.

How do you want to proceed with the qemu patches ? I attached my patches
for reference. Maybe you can add them to your tree if they are ok and submit
the entire series together to the qemu mailing list ?

Thanks,
Guenter

View attachment "0001-aspeed-i2c-Handle-receive-command-in-separate-functi.patch" of type "text/x-patch" (2454 bytes)

View attachment "0002-aspeed-i2c-Fix-receive-done-interrupt-handling.patch" of type "text/x-patch" (2349 bytes)