[U-Boot] [PATCH] net/designware: Add-on: Consecutive writes must have delay

[U-Boot] [PATCH] net/designware: Add-on: Consecutive writes must have delay

[Dinh Nguyen]

This commit is an add-on to f6c4191f. There are a few other registers where
consecutive writes must have a delay.

Signed-off-by: Dinh Nguyen <dinguyen@altera.com>
---
 drivers/net/designware.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/designware.c b/drivers/net/designware.c
index e8e669b..34952c0 100644
--- a/drivers/net/designware.c
+++ b/drivers/net/designware.c
@@ -163,8 +163,8 @@ static int dw_eth_init(struct eth_device *dev, bd_t *bis)
 	writel(FIXEDBURST | PRIORXTX_41 | BURST_16,
 			&dma_p->busmode);
 
-	writel(FLUSHTXFIFO | readl(&dma_p->opmode), &dma_p->opmode);
-	writel(STOREFORWARD | TXSECONDFRAME, &dma_p->opmode);
+	writel(readl(&dma_p->opmode) | FLUSHTXFIFO | STOREFORWARD | \
+		TXSECONDFRAME, &dma_p->opmode);
 
 	conf = FRAMEBURSTENABLE | DISABLERXOWN;
 
-- 
1.7.9.5

[U-Boot] [PATCH] net/designware: Add-on: Consecutive writes must have delay

[Wolfgang Denk]

Dear Dinh,

In message <1339104480-6191-1-git-send-email-dinguyen@altera.com> you wrote:
> This commit is an add-on to f6c4191f. There are a few other registers where
> consecutive writes must have a delay.

Sorry, but this commit message is misleading - I was expecting to see
something like udelay() in the code, but there wasn't any...

> Signed-off-by: Dinh Nguyen <dinguyen@altera.com>
> ---
>  drivers/net/designware.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/designware.c b/drivers/net/designware.c
> index e8e669b..34952c0 100644
> --- a/drivers/net/designware.c
> +++ b/drivers/net/designware.c
> @@ -163,8 +163,8 @@ static int dw_eth_init(struct eth_device *dev, bd_t *bis)
>  	writel(FIXEDBURST | PRIORXTX_41 | BURST_16,
>  			&dma_p->busmode);
>  
> -	writel(FLUSHTXFIFO | readl(&dma_p->opmode), &dma_p->opmode);
> -	writel(STOREFORWARD | TXSECONDFRAME, &dma_p->opmode);
> +	writel(readl(&dma_p->opmode) | FLUSHTXFIFO | STOREFORWARD | \
> +		TXSECONDFRAME, &dma_p->opmode);

There is no need for the continuation line here; please drop the
backslash.

Hm... Should we not rather use something like setbits_le32() here?


Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
You said you  didn't  want  to  use  CGI.pm,  but  methinks  you  are
needlessly reinventing the wheel, one spoke at a time. Either you are
masochistic,  or  you  just haven't seen enough of what CGI.pm can do
for you. -- Randal L. Schwartz in <8cyb81rg81.fsf@gadget.cscaper.com>

[U-Boot] [PATCH] net/designware: Add-on: Consecutive writes must have delay

[Wolfgang Denk]

Dear Dinh,

In message <71B37E0559AC6849A68C5BA94C509FB458298D3C24@SJ-ITMSG02.altera.priv.altera.com> you wrote:
> 
> > Sorry, but this commit message is misleading - I was expecting to see
> > something like udelay() in the code, but there wasn't any...
> 
> Combining the 2 individual writes into a single write will also work.
> I'm just keeping in sync with commit # f6c4191f.

I understand this - but the commit message should match the code.

> Again, just keeping the code in sync with the previous commit that address this issue. If you like, I can change it to setbits_le32().

I think it would be cleaner, but if the rest of the code looks like
that, I will not insist.

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"...one of the main causes of the fall of the Roman Empire was  that,
lacking  zero,  they had no way to indicate successful termination of
their C programs."                                     - Robert Firth