linux-kernel - Re: [PATCH 1/1] ishtp: Add support for Intel ishtp eclite driver

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] [day] [month] [year] [list]

Message-ID: <a447321d-79fc-6d2c-4a07-d4a24223c5a5@redhat.com>
Date:   Thu, 3 Jun 2021 19:05:33 +0200
From:   Hans de Goede <hdegoede@...hat.com>
To:     "K Naduvalath, Sumesh" <sumesh.k.naduvalath@...el.com>,
        "mgross@...ux.intel.com" <mgross@...ux.intel.com>,
        "srinivas.pandruvada@...ux.intel.com" 
        <srinivas.pandruvada@...ux.intel.com>
Cc:     "Pandruvada, Srinivas" <srinivas.pandruvada@...el.com>,
        "platform-driver-x86@...r.kernel.org" 
        <platform-driver-x86@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "Chinnu, Ganapathi" <ganapathi.chinnu@...el.com>,
        "Kumar, Nachiketa" <nachiketa.kumar@...el.com>
Subject: Re: [PATCH 1/1] ishtp: Add support for Intel ishtp eclite driver

Hi,

On 6/3/21 6:48 PM, K Naduvalath, Sumesh wrote:
> Thank you Hans for the scrutiny. Please find my comments inline.

<snip>

>>> ISH Firmware (the PSE subsystem) can boot up/reinitialize Every Sx
>>> based on usecase or sometimes PSE is always-on. So the resume path is
>>> asynchronous and unpredictable in this case. Re-initialization and
>>> clean up required if PSE also boot up every Sx and might take good
>>> amount of time (Host can come alive before PSE comes up). Thus using
>> asynchronous reset notification.
>>
>> Ok, so I guess we need to live with the ugly deregister + re-register OpRegion
>> dance for the devices where the PSE is shutdown during suspend.
>>
>> You also write: "sometimes PSE is always-on", what about that case, I assume
>> in this case there will be no reset after resume? So then unregistering the
>> OpRegion handler at suspend (and setting link_ready=false) will be wrong
>> since without the reset the OpRegion handler will never get reinstalled ?
> 
> reset is called from bus after every resume. When PSE is always-on, the notification
> comes quicky, where as there reset comes bit later in case of new boot.

Ok, in that case it makes sense to treat the PSE always-on scenario the same
as the scenario where it gets turned off during suspend.

Regards,

Hans