2011-06-21 18:31:59

by Hauke Mehrtens

Subject: [PATCH] ssb: fix ssb clock rate according to broadcom source

This fix was done according to si_clock_rate function in broadcom siutils.c
---
drivers/ssb/main.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
index e568664..57b7b64 100644
--- a/drivers/ssb/main.c
+++ b/drivers/ssb/main.c
@@ -1002,8 +1002,8 @@ u32 ssb_calc_clock_rate(u32 plltype, u32 n, u32 m)
switch (plltype) {
case SSB_PLLTYPE_6: /* 100/200 or 120/240 only */
if (m & SSB_CHIPCO_CLK_T6_MMASK)
- return SSB_CHIPCO_CLK_T6_M0;
- return SSB_CHIPCO_CLK_T6_M1;
+ return SSB_CHIPCO_CLK_T6_M1;
+ return SSB_CHIPCO_CLK_T6_M0;
case SSB_PLLTYPE_1: /* 48Mhz base, 3 dividers */
case SSB_PLLTYPE_3: /* 25Mhz, 2 dividers */
case SSB_PLLTYPE_4: /* 48Mhz, 4 dividers */
--
1.7.4.1

2011-06-22 22:06:36

by Larry Finger

Subject: Re: [PATCH v2] ssb: fix ssb clock rate according to broadcom source

On 06/22/2011 02:02 AM, Michael B?sch wrote:
> On Tue, 21 Jun 2011 16:36:38 -0500
> Larry Finger<[email protected]> wrote:
>> This patch sent me looking to find the public source for this routine. After
>> finding it in the source section of Broadcom-wl, I noticed there is some
>> additional code that is not implemented in the ssb version. Should this
>> additional material be implemented, or is it not applicable?
>
> could you quote that function here? That would be way easier to get what were talking about.

The file in question is found from the Broadcom wl driver. When you untar it
into directory "broadcom", the file is broadcom/src/shared/siutils.c.

My initial conclusion that the clock_rate routine from the above file and
drivers/ssb/main.c were different was wrong. When I checked them again, they do
exactly the same thing after this patch is applied. Sorry for the noise.

Larry

2011-06-21 21:36:42

by Larry Finger

Subject: Re: [PATCH v2] ssb: fix ssb clock rate according to broadcom source

On 06/21/2011 01:53 PM, Hauke Mehrtens wrote:
> This fix was done according to si_clock_rate function in broadcom siutils.c
>
> Signed-off-by: Hauke Mehrtens<[email protected]>
> ---
> drivers/ssb/main.c | 4 ++--
> 1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
> index e568664..57b7b64 100644
> --- a/drivers/ssb/main.c
> +++ b/drivers/ssb/main.c
> @@ -1002,8 +1002,8 @@ u32 ssb_calc_clock_rate(u32 plltype, u32 n, u32 m)
> switch (plltype) {
> case SSB_PLLTYPE_6: /* 100/200 or 120/240 only */
> if (m& SSB_CHIPCO_CLK_T6_MMASK)
> - return SSB_CHIPCO_CLK_T6_M0;
> - return SSB_CHIPCO_CLK_T6_M1;
> + return SSB_CHIPCO_CLK_T6_M1;
> + return SSB_CHIPCO_CLK_T6_M0;
> case SSB_PLLTYPE_1: /* 48Mhz base, 3 dividers */
> case SSB_PLLTYPE_3: /* 25Mhz, 2 dividers */
> case SSB_PLLTYPE_4: /* 48Mhz, 4 dividers */

This patch sent me looking to find the public source for this routine. After
finding it in the source section of Broadcom-wl, I noticed there is some
additional code that is not implemented in the ssb version. Should this
additional material be implemented, or is it not applicable?

Larry

2011-06-21 18:53:46

by Hauke Mehrtens

Subject: [PATCH v2] ssb: fix ssb clock rate according to broadcom source

This fix was done according to si_clock_rate function in broadcom siutils.c

Signed-off-by: Hauke Mehrtens <[email protected]>
---
drivers/ssb/main.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
index e568664..57b7b64 100644
--- a/drivers/ssb/main.c
+++ b/drivers/ssb/main.c
@@ -1002,8 +1002,8 @@ u32 ssb_calc_clock_rate(u32 plltype, u32 n, u32 m)
switch (plltype) {
case SSB_PLLTYPE_6: /* 100/200 or 120/240 only */
if (m & SSB_CHIPCO_CLK_T6_MMASK)
- return SSB_CHIPCO_CLK_T6_M0;
- return SSB_CHIPCO_CLK_T6_M1;
+ return SSB_CHIPCO_CLK_T6_M1;
+ return SSB_CHIPCO_CLK_T6_M0;
case SSB_PLLTYPE_1: /* 48Mhz base, 3 dividers */
case SSB_PLLTYPE_3: /* 25Mhz, 2 dividers */
case SSB_PLLTYPE_4: /* 48Mhz, 4 dividers */
--
1.7.4.1

2011-06-22 06:59:15

by Michael Büsch

Subject: Re: [PATCH v2] ssb: fix ssb clock rate according to broadcom source

On Tue, 21 Jun 2011 16:36:38 -0500
Larry Finger <[email protected]> wrote:

> On 06/21/2011 01:53 PM, Hauke Mehrtens wrote:
> > This fix was done according to si_clock_rate function in broadcom siutils.c
> >
> > Signed-off-by: Hauke Mehrtens<[email protected]>
> > ---
> > drivers/ssb/main.c | 4 ++--
> > 1 files changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
> > index e568664..57b7b64 100644
> > --- a/drivers/ssb/main.c
> > +++ b/drivers/ssb/main.c
> > @@ -1002,8 +1002,8 @@ u32 ssb_calc_clock_rate(u32 plltype, u32 n, u32 m)
> > switch (plltype) {
> > case SSB_PLLTYPE_6: /* 100/200 or 120/240 only */
> > if (m& SSB_CHIPCO_CLK_T6_MMASK)
> > - return SSB_CHIPCO_CLK_T6_M0;
> > - return SSB_CHIPCO_CLK_T6_M1;
> > + return SSB_CHIPCO_CLK_T6_M1;
> > + return SSB_CHIPCO_CLK_T6_M0;
> > case SSB_PLLTYPE_1: /* 48Mhz base, 3 dividers */
> > case SSB_PLLTYPE_3: /* 25Mhz, 2 dividers */
> > case SSB_PLLTYPE_4: /* 48Mhz, 4 dividers */
>
> This patch sent me looking to find the public source for this routine. After
> finding it in the source section of Broadcom-wl, I noticed there is some
> additional code that is not implemented in the ssb version. Should this
> additional material be implemented, or is it not applicable?

could you quote that function here? That would be way easier to get what were talking about.