linux-kernel - Re: [PATCH 2/2] soc: qcom: smp2p: Add support for smp2p v2

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: <CAOdFzchG34Eq-Px--ii3s+3yUf4ZNc=+0msr2bB86xMUExsanw@mail.gmail.com>
Date: Tue, 28 Oct 2025 17:35:45 -0700
From: Christopher Lew <christopher.lew@....qualcomm.com>
To: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
Cc: Deepak Kumar Singh <quic_deesin@...cinc.com>,
        Jingyi Wang <jingyi.wang@....qualcomm.com>,
        Bjorn Andersson <andersson@...nel.org>,
        Konrad Dybcio <konradybcio@...nel.org>, linux-arm-msm@...r.kernel.org,
        linux-kernel@...r.kernel.org, Chris Lew <chris.lew@....qualcomm.com>
Subject: Re: [PATCH 2/2] soc: qcom: smp2p: Add support for smp2p v2

On Tue, Oct 21, 2025 at 2:39 AM Konrad Dybcio
<konrad.dybcio@....qualcomm.com> wrote:
>
> On 10/21/25 10:23 AM, Deepak Kumar Singh wrote:
> >
> >
> > On 9/24/2025 8:27 PM, Konrad Dybcio wrote:
> >> On 9/24/25 6:18 AM, Jingyi Wang wrote:
> >>> From: Chris Lew <chris.lew@....qualcomm.com>
> >>>
> >>> Some remoteproc need smp2p v2 support, mirror the version written by the
> >>> remote if the remote supports v2. This is needed if the remote does not
> >>> have backwards compatibility with v1. And reset entry last value on SSR for
> >>> smp2p v2.
> >>>
> >>> Signed-off-by: Chris Lew <chris.lew@....qualcomm.com>
> >>> Co-developed-by: Jingyi Wang <jingyi.wang@....qualcomm.com>
> >>> Signed-off-by: Jingyi Wang <jingyi.wang@....qualcomm.com>
> >>> ---
> >>
> >> [...]
> >>
> >>> +static int qcom_smp2p_in_version(struct qcom_smp2p *smp2p)
> >>> +{
> >>> +    unsigned int smem_id = smp2p->smem_items[SMP2P_INBOUND];
> >>> +    unsigned int pid = smp2p->remote_pid;
> >>> +    struct smp2p_smem_item *in;
> >>> +    size_t size;
> >>> +
> >>> +    in = qcom_smem_get(pid, smem_id, &size);
> >>> +    if (IS_ERR(in))
> >>> +        return 0;
> >>> +
> >>> +    return in->version;
> >>
> >> are in and out versions supposed to be out of sync in case of
> >> error?
> >>
> > I think that should be ok. If we return error smp2p connection will be completely broken. With default version 1 partial features can be supported even if remote is using version 2. Some features like smp2p connection after subsystem restart may be affected by this.>> +}
>
> Perhaps a different question is in order.. do we ever expect smem_get to
> fail under normal conditions?
>

We would, this new flow gets executed for all the smp2p device probes.
For remoteprocs booted by HLOS, I would expect this call to
qcom_smem_get() to fail during smp2p probe time, so I have a silent
error. The in item will be found when we get the first ipcc interrupt
from the remote. I would only expect qcom_smem_get() to succeed here
on smp2p devices for early boot processors.

> [...]
>
> >>>       /*
> >>>        * Make sure the rest of the header is written before we validate the
> >>>        * item by writing a valid version number.
> >>>        */
> >>>       wmb();
> >>> -    out->version = 1;
> >>> +    out->version = (in_version) ? in_version : 1;
> >>
> >> = in_version ?: 1
> >>
> >> Konrad
> >>
> > We want to assign in_version when value is 1/2 and 1 if value is 0 i.e. error case.
>
> That's what this syntax does, with less characters
>
> Konrad