V4L/DVB fixes

Linus,

Please pull from:
ssh://master.kernel.org/pub/scm/linux/kernel/git/mchehab/v4l-dvb.git
master

For two driver regression fixes:
- bttv: VIDIOCGMBUF ioctl hangs on bttv driver
- ivtv: udelay has to be changed *after* the eeprom was read, not before

Cheers,
Mauro.

---

 drivers/media/video/bt8xx/bttv-driver.c |    4 +---
 drivers/media/video/ivtv/ivtv-driver.c  |    4 ++++
 drivers/media/video/ivtv/ivtv-i2c.c     |    5 +----
 3 files changed, 6 insertions(+), 7 deletions(-)

Gregor Jasny (1):
V4L/DVB (6944a): Fix Regression VIDIOCGMBUF ioctl hangs on bttv driver

Hans Verkuil (1):
V4L/DVB (6916): ivtv: udelay has to be changed *after* the eeprom was
read, not before

---------------------------------------------------
V4L/DVB development is hosted at http://linuxtv.org

Am Montag, 7. Januar 2008 schrieb Mauro Carvalho Chehab:

boils down to this patch:

89dab3573aa1d95fd222ee4551f964bfa4c16823

 drivers/media/video/ivtv/ivtv-driver.c |    4 ++++
 drivers/media/video/ivtv/ivtv-i2c.c    |    5 +----
 2 files changed, 5 insertions(+), 4 deletions(-)

b/drivers/media/video/ivtv/ivtv-driver.c
index 6d2dd87..10d6faf 100644

@@ -1076,6 +1076,10 @@ static int __devinit ivtv_probe(struct pci_dev *dev,
 		ivtv_process_eeprom(itv);
 	}
 
+	/* The mspx4xx chips need a longer delay for some reason */
+	if (!(itv->hw_flags & IVTV_HW_MSP34XX))
+		itv->i2c_algo.udelay = 5;
+
 	if (itv->std == 0) {
 		itv->std = V4L2_STD_NTSC_M;
 	}

b/drivers/media/video/ivtv/ivtv-i2c.c
index 44678fe..36e54f7 100644

@@ -541,7 +541,7 @@ static const struct i2c_algo_bit_data 

ivtv_i2c_algo_template = {
.setscl = ivtv_setscl_old,
.getsda = ivtv_getsda_old,
.getscl = ivtv_getscl_old,
- .udelay = 5,
+ .udelay = 10,
.timeout = 200,
};

@@ -718,9 +718,6 @@ int init_ivtv_i2c(struct ivtv *itv)
sizeof(struct i2c_adapter));
memcpy(&itv->i2c_algo, &ivtv_i2c_algo_template,
sizeof(struct i2c_algo_bit_data));
- /* The mspx4xx chips need a longer delay for some reason */
- if (itv->hw_flags & IVTV_HW_MSP34XX)
- itv->i2c_algo.udelay = 10;
itv->i2c_algo.data = itv; itv->i2c_adap.algo_data = &itv->i2c_algo;
}

where the logic in hunk #1 was switched, resulting in a now misleading comment over there.

How about something in the line of:

	/* 
	 * We started with a bigger udelay in order to fulfill the needs of the
	 * mspx4xx chips: cut it down here for all other members of the family.
	 */

Cheers,
Pete

If you notice the ! at the above line, and remind that the original value for udelay is 10, you can understand the comment ;)

Your suggestion seems clearer to my eyes. Still the other possibilities are not from Micronas msp34xx family, but audio decoder chips from other vendors - as weel as other kind of i2c devices, like IR chips, video encoders, etc.

So, IMO, a better comment would be something like:

/*
* We started with a bigger udelay in order to avoid troubles with some
* msp34xx chips. Boards without msp34xx are known to work with the full

	 * i2c 100 kHz speed.
	 */

I'm C/C the patch author (Hans) for his comments about this, since he is the one who faced with the problems with msp34xx and ivtv.

Cheers,
Mauro