linux-kernel - Re: [PATCH] cdx: add MSI support for CDX bus

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: <877ct9a6f4.ffs@tglx>
Date:   Mon, 15 May 2023 18:46:55 +0200
From:   Thomas Gleixner <tglx@...utronix.de>
To:     Nipun Gupta <nipun.gupta@....com>,
        "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>,
        "maz@...nel.org" <maz@...nel.org>, "jgg@...pe.ca" <jgg@...pe.ca>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Cc:     "git (AMD-Xilinx)" <git@....com>,
        "Anand, Harpreet" <harpreet.anand@....com>,
        "Jansen Van Vuuren, Pieter" <pieter.jansen-van-vuuren@....com>,
        "Agarwal, Nikhil" <nikhil.agarwal@....com>,
        "Simek, Michal" <michal.simek@....com>,
        "Gangurde, Abhijit" <abhijit.gangurde@....com>,
        "Cascon, Pablo" <pablo.cascon@....com>
Subject: Re: [PATCH] cdx: add MSI support for CDX bus

Nipun!

On Mon, May 15 2023 at 18:39, Nipun Gupta wrote:
> On 5/12/2023 11:45 PM, Thomas Gleixner wrote:
>> On Fri, May 12 2023 at 19:50, Nipun Gupta wrote:
>>>
>>> As per your suggestion, we can add Firmware interaction code in the
>>> irq_bus_sync_xx APIs. Another option is to change the
>>> cdx_mcdi_rpc_async() API to atomic synchronous API.
>> 
>> I'm not a great fan of that. Depending on how long this update takes the
>> CPU will busy wait for it to complete with interrupts disabled and locks
>> held.
>
> Agree. we are also inclined towards using irq_bus_sync_xx APIs. This 
> would definitely solve the issue (#1 and #2) for the setup_irq which you 
> mentioned.
>
> For MSI affinity, since GIC-ITS always return IRQ_SET_MASK_OK_DONE, the 
> irq_chip_write_msi_msg does not get called.
>
> msi_domain_set_affinity(...)
>      ret = parent->chip->irq_set_affinity(...);
>      // For GIC ITS it always return IRQ_SET_MASK_OK_DONE
>      if (ret >= 0 && ret != IRQ_SET_MASK_OK_DONE) {
>          irq_chip_write_msi_msg(...);
>      }
> Since CDX bus is specific to ARM and uses GIC ITS, it seems we do not 
> need to do anything here. Can you please share your opinion on this?

That makes sense.

Thanks,

        tglx