linux-kernel - Re: [PATCH 2/2] firmware: arm_scmi: Fix possible frequency truncation when using level indexing mode

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: <ZWoInq9QIuZXpTbc@pluto>
Date:   Fri, 1 Dec 2023 16:23:58 +0000
From:   Cristian Marussi <cristian.marussi@....com>
To:     Sudeep Holla <sudeep.holla@....com>
Cc:     linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        quic_mdtipton@...cinc.com, quic_asartor@...cinc.com,
        quic_lingutla@...cinc.com, Sibi Sankar <quic_sibis@...cinc.com>,
        linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH 2/2] firmware: arm_scmi: Fix possible frequency
 truncation when using level indexing mode

On Thu, Nov 30, 2023 at 08:43:43PM +0000, Sudeep Holla wrote:
> The multiplier is already promoted to u64, however the frequency
> calculations done when using level indexing mode doesn't use the
> multiplier computed. It instead hardcodes the multiplier value of 1000
> at all the usage sites.
> 
> Clean that up by assigning the multiplier value of 1000 when using
> the perf level indexing mode and upadte the frequency calculations to

				   ^update

> use the multiplier instead. It should fix the possible frequency
> truncation for all the values greater than or equal to 4GHz.
> 
> Fixes: 31c7c1397a33 ("firmware: arm_scmi: Add v3.2 perf level indexing mode support")
> Reported-by: Sibi Sankar <quic_sibis@...cinc.com>
> Closes: https://lore.kernel.org/all/20231129065748.19871-3-quic_sibis@quicinc.com/
> Cc: Cristian Marussi <cristian.marussi@....com>
> Signed-off-by: Sudeep Holla <sudeep.holla@....com>
> ---

Other than the typo, LGTM.

Reviewed-by: Cristian Marussi <cristian.marussi@....com>

Thanks,
Cristian

>  drivers/firmware/arm_scmi/perf.c | 12 +++++++-----
>  1 file changed, 7 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/firmware/arm_scmi/perf.c b/drivers/firmware/arm_scmi/perf.c
> index 8ce449922e55..875dcb71bb65 100644
> --- a/drivers/firmware/arm_scmi/perf.c
> +++ b/drivers/firmware/arm_scmi/perf.c
> @@ -268,7 +268,8 @@ scmi_perf_domain_attributes_get(const struct scmi_protocol_handle *ph,
>  		dom_info->sustained_perf_level =
>  					le32_to_cpu(attr->sustained_perf_level);
>  		if (!dom_info->sustained_freq_khz ||
> -		    !dom_info->sustained_perf_level)
> +		    !dom_info->sustained_perf_level ||
> +		    dom_info->level_indexing_mode)
>  			/* CPUFreq converts to kHz, hence default 1000 */
>  			dom_info->mult_factor =	1000;
>  		else
> @@ -806,7 +807,7 @@ static int scmi_dvfs_device_opps_add(const struct scmi_protocol_handle *ph,
>  		if (!dom->level_indexing_mode)
>  			freq = dom->opp[idx].perf * dom->mult_factor;
>  		else
> -			freq = dom->opp[idx].indicative_freq * 1000;
> +			freq = dom->opp[idx].indicative_freq * dom->mult_factor;
> 
>  		data.level = dom->opp[idx].perf;
>  		data.freq = freq;
> @@ -853,7 +854,8 @@ static int scmi_dvfs_freq_set(const struct scmi_protocol_handle *ph, u32 domain,
>  	} else {
>  		struct scmi_opp *opp;
> 
> -		opp = LOOKUP_BY_FREQ(dom->opps_by_freq, freq / 1000);
> +		opp = LOOKUP_BY_FREQ(dom->opps_by_freq,
> +				     freq / dom->mult_factor);
>  		if (!opp)
>  			return -EIO;
> 
> @@ -887,7 +889,7 @@ static int scmi_dvfs_freq_get(const struct scmi_protocol_handle *ph, u32 domain,
>  		if (!opp)
>  			return -EIO;
> 
> -		*freq = opp->indicative_freq * 1000;
> +		*freq = opp->indicative_freq * dom->mult_factor;
>  	}
> 
>  	return ret;
> @@ -910,7 +912,7 @@ static int scmi_dvfs_est_power_get(const struct scmi_protocol_handle *ph,
>  		if (!dom->level_indexing_mode)
>  			opp_freq = opp->perf * dom->mult_factor;
>  		else
> -			opp_freq = opp->indicative_freq * 1000;
> +			opp_freq = opp->indicative_freq * dom->mult_factor;
> 
>  		if (opp_freq < *freq)
>  			continue;
> --
> 2.43.0
>