Skip to contentSkip to conversation

DVB: BANDWIDTH_TO_KHZ strangeness

DISCUSSION Adrian Bunk 3 participants 5 messages Oct 14, 2007 – Oct 22, 2007

lore
Share & tools
lore
#0 lore raw

drivers/media/dvb/frontends/dibx000_common.h contains:

<-- snip -->

...

#define BANDWIDTH_TO_KHZ(v) ( (v) == BANDWIDTH_8_MHZ  ? 8000 : \
                             (v) == BANDWIDTH_7_MHZ  ? 7000 : \
                             (v) == BANDWIDTH_6_MHZ  ? 6000 : 8000 )

...

<-- snip -->

Commit b6884a17fc70e979ef34e4b5560988b522bb50a0 added to both of drivers/media/dvb/frontends/dib7000{m,p}.c:

<-- snip -->

...

        factor = BANDWIDTH_TO_KHZ(ch->u.ofdm.bandwidth);
        if (factor >= 5000)
                factor = 1;
        else
                factor = 6;

...

<-- snip -->

factor < 5000 is obviously never possible.

drivers/media/dvb/frontends/dib0070.c contains a similar assumption that BANDWIDTH_TO_KHZ() could result in values other than {6,7,8}000 (I haven't checked whether there are more such assumptions in other places).

Spotted by the Coverity checker.

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#1 lore raw

Hi all,

This coverity checker is great. He even finds all the things which are there because I'm lazy ;) .

The dibXXXX-drivers are all generated from our internal drivers to have at least something for Linux. Because of no time there is no dedicated source for the kernel but just the processed one.

The particular issue here is, that the internal architecture supports a variable bandwidth while the linux-dvb architecture only has 3 fixed values.

I know that in the future the linux-dvb-API will also support other bandwidths so I'm begging for mercy for those 3 things here to not get too much out-of-sync with our internal code.

thanks,
Patrick.

Quoted 55 lines · Adrian Bunk

drivers/media/dvb/frontends/dibx000_common.h contains:

<-- snip -->

...

#define BANDWIDTH_TO_KHZ(v) ( (v) == BANDWIDTH_8_MHZ  ? 8000 : \
                             (v) == BANDWIDTH_7_MHZ  ? 7000 : \
                             (v) == BANDWIDTH_6_MHZ  ? 6000 : 8000 )

...

<-- snip -->

Commit b6884a17fc70e979ef34e4b5560988b522bb50a0 added to both of drivers/media/dvb/frontends/dib7000{m,p}.c:

<-- snip -->

...

        factor = BANDWIDTH_TO_KHZ(ch->u.ofdm.bandwidth);
        if (factor >= 5000)
                factor = 1;
        else
                factor = 6;

...

<-- snip -->

factor < 5000 is obviously never possible.

drivers/media/dvb/frontends/dib0070.c contains a similar assumption that BANDWIDTH_TO_KHZ() could result in values other than {6,7,8}000 (I haven't checked whether there are more such assumptions in other places).

Spotted by the Coverity checker.

cu
Adrian

Signature · 10 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed


_______________________________________________
v4l-dvb-maintainer mailing list
[email protected]
http://www.linuxtv.org/cgi-bin/mailman/listinfo/v4l-dvb-maintainer
#2 lore raw
Quoted 4 lines

I know that in the future the linux-dvb-API will also support other bandwidths so I'm begging for mercy for those 3 things here to not get too much out-of-sync with our internal code.

I don't see much problem on keeping this for a while.

However, if not causing to much troubles for you to manage, I would to this, instead:

#if 0
	/* Currently, DVB API allows only bandwidths starting from 5 GHz */
        factor = BANDWIDTH_TO_KHZ(ch->u.ofdm.bandwidth);
        if (factor >= 5000)
                factor = 1;
        else
                factor = 6;
#else
	factor = 6;

#endif

With the above code, gentree.pl scripts will automatically remove the dead code from the Kernel, while keeping it defined at the development environment.

If you want, you may also replace the #if 0 by something like:

#ifdef API_SUPPORTS_LOW_BANDWIDTH

In this case, by adding API_SUPPORTS_LOW_BANDWIDTH to gentree.pl, the same effect of eliminating the dead code from kernel can be produced, since gentree.pl is capable of evaluating cpp macros like the above to 0 (or 1) for the constants declared on a table inside its code.

Signature · 2 lines
Cheers,
Mauro
#3 lore raw
Quoted 18 lines · Mauro Carvalho Chehab
Quoted 4 lines

I know that in the future the linux-dvb-API will also support other bandwidths so I'm begging for mercy for those 3 things here to not get too much out-of-sync with our internal code.

I don't see much problem on keeping this for a while.

However, if not causing to much troubles for you to manage, I would to this, instead:

#if 0
	/* Currently, DVB API allows only bandwidths starting from 5 GHz */
        factor = BANDWIDTH_TO_KHZ(ch->u.ofdm.bandwidth);
        if (factor >= 5000)
                factor = 1;
        else
                factor = 6;
#else
	factor = 6;
Quoted 6 lines

#endif

With the above code, gentree.pl scripts will automatically remove the dead code from the Kernel, while keeping it defined at the development environment.
...

Good compilers like gcc generate the same code for both cases [1], so there's no reason for changing anything.

The reason for my email was that it looked strange, but since it's intended I'd say it's OK.

Quoted 2 lines

Cheers,
Mauro

cu
Adrian

[1] except that your #else case contains the wrong value ;-)

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#4 lore raw

Em Seg, 2007-10-22 às 21:54 +0200, Adrian Bunk escreveu:

Quoted 1 line

[1] except that your #else case contains the wrong value ;-)

This is what happens when you try to write "quick" fixes ;)

/me is needing a day with 48 hours...

Signature · 2 lines
Cheers,
Mauro