Re: [PATCH] usb: renesas-xhci: Fix External ROM access timeouts

[Marek Vasut <marek.vasut@mailbox.org>]

On 7/29/25 5:20 PM, Mathias Nyman wrote:

Hi,

>> diff --git a/drivers/usb/host/xhci-pci-renesas.c b/drivers/usb/host/ 
>> xhci-pci-renesas.c
>> index 620f8f0febb8..86df80399c9f 100644
>> --- a/drivers/usb/host/xhci-pci-renesas.c
>> +++ b/drivers/usb/host/xhci-pci-renesas.c
>> @@ -47,8 +47,9 @@
>>   #define RENESAS_ROM_ERASE_MAGIC                0x5A65726F
>>   #define RENESAS_ROM_WRITE_MAGIC                0x53524F4D
>> -#define RENESAS_RETRY    10000
>> -#define RENESAS_DELAY    10
>> +#define RENESAS_RETRY            50000    /* 50000 * RENESAS_DELAY ~= 
>> 500ms */
>> +#define RENESAS_CHIP_ERASE_RETRY    500000    /* 500000 * 
>> RENESAS_DELAY ~= 5s */
>> +#define RENESAS_DELAY            10
> 
> No objection, just making sure author is aware that RENESAS_RETRY is 
> used in
> _seven_ for loops, and this change will increase the timeout five-fold for
> all of them.

Yes, I am aware this increases the timeout for all SPI EEPROM status 
polling loops in this driver.

The longer retry count would only have an impact in case something bad 
happened during SPI EEPROM programming. On user system, that should 
happen never -- user hardware should ship with already programmed SPI 
EEPROM, so this programming code is skipped. In factory, this might 
happen, but then it is likely a hardware defect and that hardware never 
reaches users.

-- 
Best regards,
Marek Vasut