[PATCH V2 1/2] mtd: bcm47xxpart: move TRX parsing code to separated function

[Marek Vasut]

On 02/09/2017 08:11 PM, Boris Brezillon wrote:
> On Thu, 9 Feb 2017 19:21:03 +0100
> Marek Vasut <marek.vasut at gmail.com> wrote:
> 
>> On 02/09/2017 07:05 PM, Rafał Miłecki wrote:
>>> On 2017-02-09 18:28, Marek Vasut wrote:  
>>>> On 01/10/2017 11:15 PM, Rafał Miłecki wrote:  
>>>>> From: Rafał Miłecki <rafal at milecki.pl>
>>>>>
>>>>> This change simplifies main parsing loop logic a bit. In future it may
>>>>> be useful for moving TRX support to separated module / parser (if we
>>>>> implement support for them at some point).
>>>>> Finally parsing TRX at the end puts us in a better position as we have
>>>>> better flash layout knowledge. It may be useful e.g. if it appears there
>>>>> is more than 1 TRX partition.
>>>>>
>>>>> Signed-off-by: Rafał Miłecki <rafal at milecki.pl>
>>>>> ---
>>>>> V2: Keep using bcm47xxpart_add_part in new parsing function
>>>>> ---
>>>>>  drivers/mtd/bcm47xxpart.c | 121
>>>>> ++++++++++++++++++++++++++++------------------
>>>>>  1 file changed, 74 insertions(+), 47 deletions(-)
>>>>>
>>>>> diff --git a/drivers/mtd/bcm47xxpart.c b/drivers/mtd/bcm47xxpart.c
>>>>> index 283ff7e..1093025 100644
>>>>> --- a/drivers/mtd/bcm47xxpart.c
>>>>> +++ b/drivers/mtd/bcm47xxpart.c
>>>>> @@ -83,6 +83,67 @@ static const char
>>>>> *bcm47xxpart_trx_data_part_name(struct mtd_info *master,
>>>>>      return "rootfs";
>>>>>  }
>>>>>
>>>>> +static int bcm47xxpart_parse_trx(struct mtd_info *master,
>>>>> +                 struct mtd_partition *trx,
>>>>> +                 struct mtd_partition *parts,
>>>>> +                 size_t parts_len)
>>>>> +{
>>>>> +    struct trx_header header;
>>>>> +    size_t bytes_read;
>>>>> +    int curr_part = 0;
>>>>> +    int i, err;
>>>>> +
>>>>> +    if (parts_len < 3) {
>>>>> +        pr_warn("No enough space to add TRX partitions!\n");
>>>>> +        return -ENOMEM;
>>>>> +    }
>>>>> +
>>>>> +    err = mtd_read(master, trx->offset, sizeof(header), &bytes_read,
>>>>> +               (uint8_t *)&header);
>>>>> +    if (err && !mtd_is_bitflip(err)) {
>>>>> +        pr_err("mtd_read error while reading TRX header: %d\n", err);  
>>>>
>>>> Minor nit -- Can we make this dev_err() ?  
>>>
>>> The problem is parsers receive struct mtd_info (as an argument) with has
>>> empty
>>> dev. I just tried
>>> dev_info(&master->dev, "TEST TEST TEST :)\n");
>>> and got this:
>>> [    0.491576]  (null): TEST TEST TEST :)
>>>
>>> I guess that's the reason why parse_mtd_partitions and
>>> mtd_device_parse_register
>>> also use pr_* functions.
>>>
>>> So this should be improved at mtd core subsystem first, then we can use
>>> it in
>>> drivers as well.  
>>
>> Isn't this patch fixing just that?
>>  [PATCH] mtd: Add partition device node to mtd partition devices
>>
> 
> 
> No, this patch is only adding a pointer to a device_node object (DT
> node).
> To solve the problem, we need to have the master dev name set before we
> register its partitions, and this is currently only the case if you
> enable CONFIG_MTD_PARTITIONED_MASTER.
> 
Ah hmmmm , then leave the pr_err() there , this can be fixed later.

Acked-by: Marek Vasut <marek.vasut at gmail.com>

-- 
Best regards,
Marek Vasut