linux-kernel - Re: [V5] tty: serial: qcom-geni-serial: Fix get_clk_div_rate() which otherwise could return a sub-optimal clock rate.

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]

Date:   Thu, 14 Jul 2022 18:35:17 +0530
From:   Vijaya Krishna Nivarthi <quic_vnivarth@...cinc.com>
To:     Jiri Slaby <jirislaby@...nel.org>, <agross@...nel.org>,
        <bjorn.andersson@...aro.org>, <konrad.dybcio@...ainline.org>,
        <gregkh@...uxfoundation.org>, <linux-arm-msm@...r.kernel.org>,
        <linux-serial@...r.kernel.org>, <linux-kernel@...r.kernel.org>
CC:     <quic_msavaliy@...cinc.com>, <dianders@...omium.org>,
        <mka@...omium.org>, <swboyd@...omium.org>
Subject: Re: [V5] tty: serial: qcom-geni-serial: Fix get_clk_div_rate() which
 otherwise could return a sub-optimal clock rate.


On 7/14/2022 4:16 PM, Jiri Slaby wrote:
> On 12. 07. 22, 14:45, Vijaya Krishna Nivarthi wrote:
> ...
>> +static unsigned long get_clk_div_rate(struct clk *clk, struct device 
>> *dev,
>> +        unsigned int baud, unsigned int sampling_rate, unsigned int 
>> *clk_div)
>> +{
>> +    unsigned long ser_clk;
>> +    unsigned long desired_clk;
>> +
>> +    desired_clk = baud * sampling_rate;
>> +    if (!desired_clk) {
>> +        dev_dbg(dev, "Invalid frequency\n");
>> +        return 0;
>>       }
>>   -    *clk_div = ser_clk / desired_clk;
>> -    if (!(*clk_div))
>> -        *clk_div = 1;
>> +    /*
>> +     * try to find a clock rate within 2% tolerance, then within
>
> "then within" ... "5" is missing, perhaps?


Yes :(

Will change.


>
>> +     */
>> +    ser_clk = find_clk_rate_in_tol(clk, desired_clk, clk_div, 2);
>> +    if (!ser_clk)
>> +        ser_clk = find_clk_rate_in_tol(clk, desired_clk, clk_div, 5);
>> +
>> +    if (!ser_clk)
>> +        dev_err(dev, "Couldn't find suitable clock rate for %lu\n", 
>> desired_clk);
>> +    else
>> +        dev_dbg(dev, "desired_clk-%lu, ser_clk-%lu, clk_div-%u\n",
>> +                desired_clk, ser_clk, *clk_div);
>>         return ser_clk;
>>   }
>> @@ -1021,8 +1040,8 @@ static void qcom_geni_serial_set_termios(struct 
>> uart_port *uport,
>>       if (ver >= QUP_SE_VERSION_2_5)
>>           sampling_rate /= 2;
>>   -    clk_rate = get_clk_div_rate(port->se.clk, baud,
>> -        sampling_rate, &clk_div);
>> +    clk_rate = get_clk_div_rate(port->se.clk, port->se.dev, baud,
>
> Maybe worth passing whole geni_se (port->se) then?


but then geni_se is a struct; shall we pass port instead?

Alternately move logging to caller?

Thank you.


>
>> +                    sampling_rate, &clk_div);
>>       if (!clk_rate)
>>           goto out_restart_rx;
>
> thanks,