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

Groups > linux.kernel > #1418472 > unrolled thread

Re: [PATCH] staging: fbtft: do not allocate huge txbuf

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] staging: fbtft: do not allocate huge txbuf Noralf Trønnes <noralf@tronnes.org> - 2016-06-09 18:30 +0200
    Re: [PATCH] staging: fbtft: do not allocate huge txbuf Michal Suchanek <hramrach@gmail.com> - 2016-06-09 19:20 +0200
      Re: [PATCH] staging: fbtft: do not allocate huge txbuf Noralf Trønnes <noralf@tronnes.org> - 2016-06-09 20:50 +0200

#1418472 — Re: [PATCH] staging: fbtft: do not allocate huge txbuf

Den 09.06.2016 17:08, skrev Michal Suchanek:
> txbuflen can be set to arbitrary value by user and it is also set
> automagically to the maximum transfer size of the SPI master controller.

AFAICT this is a result of your previous patch. Please make a new version of
your previous patch with this fix in it. Also make a note in that thread 
that
a new version is made so Greg doesn't pull it when he sees my ack. He
handles a large volume of patches so let's make it as easy as we can for 
him.

Thanks,
Noralf.

> Do not allocate the buffer when larger than vmem. When my SPI master
> controller reports maximum transfer size 16M the probe of fbtft fails.
>
> Signed-off-by: Michal Suchanek <hramrach@gmail.com>
> ---
>   drivers/staging/fbtft/fbtft-core.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/staging/fbtft/fbtft-core.c b/drivers/staging/fbtft/fbtft-core.c
> index f3bdc8f..b95cf69 100644
> --- a/drivers/staging/fbtft/fbtft-core.c
> +++ b/drivers/staging/fbtft/fbtft-core.c
> @@ -820,6 +820,8 @@ struct fb_info *fbtft_framebuffer_alloc(struct fbtft_display *display,
>   	/* Transmit buffer */
>   	if (txbuflen == -1)
>   		txbuflen = vmem_size + 2; /* add in case startbyte is used */
> +	if (txbuflen >= vmem_size + 2)
> +		txbuflen = 0;
>   
>   #ifdef __LITTLE_ENDIAN
>   	if ((!txbuflen) && (bpp > 8))

[toc] | [next] | [standalone]

#1418505

Hello,

On 9 June 2016 at 18:20, Noralf Trønnes <noralf@tronnes.org> wrote:
>
> Den 09.06.2016 17:08, skrev Michal Suchanek:
>>
>> txbuflen can be set to arbitrary value by user and it is also set
>> automagically to the maximum transfer size of the SPI master controller.
>
>
> AFAICT this is a result of your previous patch. Please make a new version of
> your previous patch with this fix in it. Also make a note in that thread
> that

That patch is pretty much independent and just exposes this problem
under different circumstances. The txbuflen can be also set by the
user.

> a new version is made so Greg doesn't pull it when he sees my ack. He
> handles a large volume of patches so let's make it as easy as we can for
> him.

Which would be letting in the patches separately imho.

Thanks

Michal

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

#1418587

[resending got rejected on the list for html format]

Den 09.06.2016 19:18, skrev Michal Suchanek:
> Hello,
>
> On 9 June 2016 at 18:20, Noralf Trønnes <noralf@tronnes.org> wrote:
>> Den 09.06.2016 17:08, skrev Michal Suchanek:
>>> txbuflen can be set to arbitrary value by user and it is also set
>>> automagically to the maximum transfer size of the SPI master controller.
>>
>> AFAICT this is a result of your previous patch. Please make a new version of
>> your previous patch with this fix in it. Also make a note in that thread
>> that
> That patch is pretty much independent and just exposes this problem
> under different circumstances. The txbuflen can be also set by the
> user.
>
>> a new version is made so Greg doesn't pull it when he sees my ack. He
>> handles a large volume of patches so let's make it as easy as we can for
>> him.
> Which would be letting in the patches separately imho.


Ok,

Acked-by: Noralf Trønnes <noralf@tronnes.org>

> Thanks
>
> Michal
>

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web