Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1330606 > unrolled thread

Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read request

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read  request Mark Brown <broonie@kernel.org> - 2016-02-09 20:40 +0100
    Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read  request Vignesh R <vigneshr@ti.com> - 2016-02-11 06:40 +0100
      Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read  request Mark Brown <broonie@kernel.org> - 2016-02-12 23:40 +0100
        Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read  request Vignesh R <vigneshr@ti.com> - 2016-02-16 09:10 +0100
          Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read  request Mark Brown <broonie@kernel.org> - 2016-02-16 13:40 +0100
            Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read  request "R, Vignesh" <vigneshr@ti.com> - 2016-02-17 17:20 +0100

#1330606 — Re: [PATCH v5 3/5] mtd: devices: m25p80: add support for mmap read request

[Multipart message — attachments visible in raw view] — view raw

On Fri, Dec 11, 2015 at 09:39:58AM +0530, Vignesh R wrote:

> +	if (spi_flash_read_supported(spi)) {
> +		struct spi_flash_read_message msg;
> +		int ret;
> +
> +		msg.buf = buf;
> +		msg.from = from;
> +		msg.len = len;
> +		msg.read_opcode = nor->read_opcode;
> +		msg.addr_width = nor->addr_width;
> +		msg.dummy_bytes = dummy;
> +		/* TODO: Support other combinations */
> +		msg.opcode_nbits = SPI_NBITS_SINGLE;
> +		msg.addr_nbits = SPI_NBITS_SINGLE;
> +		msg.data_nbits = m25p80_rx_nbits(nor);
> +
> +		ret = spi_flash_read(spi, &msg);
> +		*retlen = msg.retlen;
> +		return ret;

Looking at this I can't help but think that spi_flash_read() ought to
have the stub in rather than the caller.  But given that we're pretty
much only ever expecting one user I'm not 100% sure it actually matters.
Anyway, I applied the first two patches.

[toc] | [next] | [standalone]

#1331698

On 02/10/2016 01:06 AM, Mark Brown wrote:
> On Fri, Dec 11, 2015 at 09:39:58AM +0530, Vignesh R wrote:
> 
>> +	if (spi_flash_read_supported(spi)) {
>> +		struct spi_flash_read_message msg;
>> +		int ret;
>> +
>> +		msg.buf = buf;
>> +		msg.from = from;
>> +		msg.len = len;
>> +		msg.read_opcode = nor->read_opcode;
>> +		msg.addr_width = nor->addr_width;
>> +		msg.dummy_bytes = dummy;
>> +		/* TODO: Support other combinations */
>> +		msg.opcode_nbits = SPI_NBITS_SINGLE;
>> +		msg.addr_nbits = SPI_NBITS_SINGLE;
>> +		msg.data_nbits = m25p80_rx_nbits(nor);
>> +
>> +		ret = spi_flash_read(spi, &msg);
>> +		*retlen = msg.retlen;
>> +		return ret;
> 
> Looking at this I can't help but think that spi_flash_read() ought to
> have the stub in rather than the caller.  But given that we're pretty
> much only ever expecting one user I'm not 100% sure it actually matters.

Well, my initial patch set passed long list of arguments to
spi_flash_read(), but Brian suggested to use struct[1] in order to avoid
unnecessary churn when things need changed in the API.


[1] https://lkml.org/lkml/2015/11/11/454

-- 
Regards
Vignesh

[toc] | [prev] | [next] | [standalone]

#1333183

[Multipart message — attachments visible in raw view] — view raw

On Thu, Feb 11, 2016 at 11:03:50AM +0530, Vignesh R wrote:
> On 02/10/2016 01:06 AM, Mark Brown wrote:
> > On Fri, Dec 11, 2015 at 09:39:58AM +0530, Vignesh R wrote:

> >> +	if (spi_flash_read_supported(spi)) {
> >> +		struct spi_flash_read_message msg;
> >> +		int ret;

> > Looking at this I can't help but think that spi_flash_read() ought to
> > have the stub in rather than the caller.  But given that we're pretty
> > much only ever expecting one user I'm not 100% sure it actually matters.

> Well, my initial patch set passed long list of arguments to
> spi_flash_read(), but Brian suggested to use struct[1] in order to avoid
> unnecessary churn when things need changed in the API.

I don't see what that has to do with my point?

[toc] | [prev] | [next] | [standalone]

#1335125

On 02/13/2016 04:07 AM, Mark Brown wrote:
> On Thu, Feb 11, 2016 at 11:03:50AM +0530, Vignesh R wrote:
>> On 02/10/2016 01:06 AM, Mark Brown wrote:
>>> On Fri, Dec 11, 2015 at 09:39:58AM +0530, Vignesh R wrote:
> 
>>>> +	if (spi_flash_read_supported(spi)) {
>>>> +		struct spi_flash_read_message msg;
>>>> +		int ret;
> 
>>> Looking at this I can't help but think that spi_flash_read() ought to
>>> have the stub in rather than the caller.  But given that we're pretty
>>> much only ever expecting one user I'm not 100% sure it actually matters.
> 
>> Well, my initial patch set passed long list of arguments to
>> spi_flash_read(), but Brian suggested to use struct[1] in order to avoid
>> unnecessary churn when things need changed in the API.
> 
> I don't see what that has to do with my point?
> 

AFAIU, your previous comment was to move initialization of
spi_flash_read_message struct to spi_flash_read(). This would mean
sending long list of arguments to spi_flash_read() which needs to be
updated whenever an argument needs to be added/deleted (in future).
Instead passing around a struct would be much easier in case of
adding/removing parameters.
Please correct me if I misunderstood your comment?

-- 
Regards
Vignesh

[toc] | [prev] | [next] | [standalone]

#1335330

[Multipart message — attachments visible in raw view] — view raw

On Tue, Feb 16, 2016 at 01:30:49PM +0530, Vignesh R wrote:
> On 02/13/2016 04:07 AM, Mark Brown wrote:
> > On Thu, Feb 11, 2016 at 11:03:50AM +0530, Vignesh R wrote:
> >> On 02/10/2016 01:06 AM, Mark Brown wrote:

> >>> Looking at this I can't help but think that spi_flash_read() ought to
> >>> have the stub in rather than the caller.  But given that we're pretty
> >>> much only ever expecting one user I'm not 100% sure it actually matters.

> >> Well, my initial patch set passed long list of arguments to
> >> spi_flash_read(), but Brian suggested to use struct[1] in order to avoid
> >> unnecessary churn when things need changed in the API.

> > I don't see what that has to do with my point?

> AFAIU, your previous comment was to move initialization of
> spi_flash_read_message struct to spi_flash_read(). This would mean

No, not at all.  I'm talking about how we handle the case where we don't
have hardware support for this and need to implement it in software -
currently that's in a separate place to the place where we call the
driver.

[toc] | [prev] | [next] | [standalone]

#1336551

On 02/16/2016 06:08 PM, Mark Brown wrote:
> On Tue, Feb 16, 2016 at 01:30:49PM +0530, Vignesh R wrote:
>> On 02/13/2016 04:07 AM, Mark Brown wrote:
>>> On Thu, Feb 11, 2016 at 11:03:50AM +0530, Vignesh R wrote:
>>>> On 02/10/2016 01:06 AM, Mark Brown wrote:
> 
>>>>> Looking at this I can't help but think that spi_flash_read() ought to
>>>>> have the stub in rather than the caller.  But given that we're pretty
>>>>> much only ever expecting one user I'm not 100% sure it actually matters.
> 
>>>> Well, my initial patch set passed long list of arguments to
>>>> spi_flash_read(), but Brian suggested to use struct[1] in order to avoid
>>>> unnecessary churn when things need changed in the API.
> 
>>> I don't see what that has to do with my point?
> 
>> AFAIU, your previous comment was to move initialization of
>> spi_flash_read_message struct to spi_flash_read(). This would mean
> 
> No, not at all.  I'm talking about how we handle the case where we don't
> have hardware support for this and need to implement it in software -
> currently that's in a separate place to the place where we call the
> driver.
> 

Yeah, but AFAIK, hardware accelerated read support is applicable for
m25p80 flashes only, I doubt whether spi_flash_read() will be used by
other types. I felt keeping the software implementation in m25p80_read()
will be consistent with m25p80_write().


-- 
Regards
Vignesh

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web