qemu-arm
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-arm] [Qemu-devel] [PATCH 18/19] aspeed/smc: inject errors in D


From: Cédric Le Goater
Subject: Re: [Qemu-arm] [Qemu-devel] [PATCH 18/19] aspeed/smc: inject errors in DMA checksum
Date: Fri, 14 Jun 2019 14:02:04 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.7.0

On 13/06/2019 16:31, Philippe Mathieu-Daudé wrote:
> Hi Cédric,
> 
> On 5/25/19 5:12 PM, Cédric Le Goater wrote:
>> Emulate read errors in the DMA Checksum Register for high frequencies
>> and optimistic settings of the Read Timing Compensation Register. This
>> will help in tuning the SPI timing calibration algorithm.
>>
>> The values below are those to expect from the first flash device of
>> the FMC controller of a palmetto-bmc machine.
>>
>> Signed-off-by: Cédric Le Goater <address@hidden>
>> ---
>>  hw/ssi/aspeed_smc.c | 29 +++++++++++++++++++++++++++++
>>  1 file changed, 29 insertions(+)
>>
>> diff --git a/hw/ssi/aspeed_smc.c b/hw/ssi/aspeed_smc.c
>> index 406c30c60b3f..4c162912cf62 100644
>> --- a/hw/ssi/aspeed_smc.c
>> +++ b/hw/ssi/aspeed_smc.c
>> @@ -866,6 +866,30 @@ static void aspeed_smc_dma_calibration(AspeedSMCState 
>> *s)
>>      s->regs[s->r_ctrl0 + cs] |= CE_CTRL_CLOCK_FREQ(hclk_div);
>>  }
>>  
> 
> Can you add a comment (like the patch description) here?

yes. done.
 
>> +static bool aspeed_smc_inject_read_failure(AspeedSMCState *s)
>> +{
>> +    uint8_t delay =
>> +        (s->regs[R_DMA_CTRL] >> DMA_CTRL_DELAY_SHIFT) & DMA_CTRL_DELAY_MASK;
>> +    uint8_t hclk_mask =
>> +        (s->regs[R_DMA_CTRL] >> DMA_CTRL_FREQ_SHIFT) & DMA_CTRL_FREQ_MASK;
>> +
>> +    /*
>> +     * Typical values of a palmetto-bmc machine.
>> +     */
>> +    switch (aspeed_smc_hclk_divisor(hclk_mask)) {
>> +    case 4 ... 16:
>> +        return false;
>> +    case 3: /* at least one HCLK cycle delay */
>> +        return (delay & 0x7) < 1;
>> +    case 2: /* at least two HCLK cycle delay */
>> +        return (delay & 0x7) < 2;
>> +    case 1: /* (> 100MHz) is above the max freq of the controller */
>> +        return true;
>> +    default:
>> +        g_assert_not_reached();
>> +    }
>> +}
>> +
>>  /*
>>   * Accumulate the result of the reads to provide a checksum that will
>>   * be used to validate the read timing settings.
>> @@ -903,6 +927,11 @@ static void aspeed_smc_dma_checksum(AspeedSMCState *s)
>>          s->regs[R_DMA_FLASH_ADDR] += 4;
>>          s->regs[R_DMA_LEN] -= 4;
>>      }
>> +
>> +    if (aspeed_smc_inject_read_failure(s)) {
> 
> So this model real world where noise eventually triggers errors. Don't
> we want this to be enable by the user (or a QMP command)?

I can add a property at the device model level to trigger this behavior.
Such as :

   -global driver=aspeed.smc,property=timing,value=true

timing if defined would provide the maximum clock and delay settings.

Are there any other examples in QEMU ? 

Thanks,

C.

> 
>> +        s->regs[R_DMA_CHECKSUM] = 0xbadc0de;
>> +    }
>> +
>>  }
>>  
>>  static void aspeed_smc_dma_rw(AspeedSMCState *s)
>>




reply via email to

[Prev in Thread] Current Thread [Next in Thread]