| Message ID | 1302824827-7348-1-git-send-email-alexdeucher@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Hi Alex, On Thu, 14 Apr 2011 19:47:06 -0400, Alex Deucher wrote: > Apparently some distros set i2c-algo-bit.bit_test to 1 by > default. In some cases this causes i2c_bit_add_bus > to fail and prevents the i2c bus from being added. In the > radeon case, we fail to add the ddc i2c buses which prevents > the driver from being able to detect attached monitors. > The i2c bus works fine even if bit_test fails. This is likely > due to gpio switching that is required and handled in the > pre/post_xfer hooks, so call the pre/post_xfer hooks in the > bit test as well. > > Fixes: > https://bugs.freedesktop.org/show_bug.cgi?id=36221 > > Cc: Jean Delvare <khali@linux-fr.org> > Signed-off-by: Alex Deucher <alexdeucher@gmail.com> Good catch, applied, thanks. I'll also push this to the stable kernel trees (from .38 down to .34.) > --- > drivers/i2c/algos/i2c-algo-bit.c | 21 ++++++++++++++++++--- > 1 files changed, 18 insertions(+), 3 deletions(-) > > diff --git a/drivers/i2c/algos/i2c-algo-bit.c b/drivers/i2c/algos/i2c-algo-bit.c > index 38319a6..e2740e6 100644 > --- a/drivers/i2c/algos/i2c-algo-bit.c > +++ b/drivers/i2c/algos/i2c-algo-bit.c > @@ -232,9 +232,16 @@ static int i2c_inb(struct i2c_adapter *i2c_adap) > * Sanity check for the adapter hardware - check the reaction of > * the bus lines only if it seems to be idle. > */ > -static int test_bus(struct i2c_algo_bit_data *adap, char *name) > +static int test_bus(struct i2c_adapter *i2c_adap, char *name) > { > - int scl, sda; > + struct i2c_algo_bit_data *adap = i2c_adap->algo_data; > + int scl, sda, ret; > + > + if (adap->pre_xfer) { > + ret = adap->pre_xfer(i2c_adap); > + if (ret < 0) > + return -ENODEV; > + } > > if (adap->getscl == NULL) > pr_info("%s: Testing SDA only, SCL is not readable\n", name); > @@ -297,11 +304,19 @@ static int test_bus(struct i2c_algo_bit_data *adap, char *name) > "while pulling SCL high!\n", name); > goto bailout; > } > + > + if (adap->post_xfer) > + adap->post_xfer(i2c_adap); > + > pr_info("%s: Test OK\n", name); > return 0; > bailout: > sdahi(adap); > sclhi(adap); > + > + if (adap->post_xfer) > + adap->post_xfer(i2c_adap); > + > return -ENODEV; > } > > @@ -607,7 +622,7 @@ static int __i2c_bit_add_bus(struct i2c_adapter *adap, > int ret; > > if (bit_test) { > - ret = test_bus(bit_adap, adap->name); > + ret = test_bus(adap, adap->name); > if (ret < 0) > return -ENODEV; > }
diff --git a/drivers/i2c/algos/i2c-algo-bit.c b/drivers/i2c/algos/i2c-algo-bit.c index 38319a6..e2740e6 100644 --- a/drivers/i2c/algos/i2c-algo-bit.c +++ b/drivers/i2c/algos/i2c-algo-bit.c @@ -232,9 +232,16 @@ static int i2c_inb(struct i2c_adapter *i2c_adap) * Sanity check for the adapter hardware - check the reaction of * the bus lines only if it seems to be idle. */ -static int test_bus(struct i2c_algo_bit_data *adap, char *name) +static int test_bus(struct i2c_adapter *i2c_adap, char *name) { - int scl, sda; + struct i2c_algo_bit_data *adap = i2c_adap->algo_data; + int scl, sda, ret; + + if (adap->pre_xfer) { + ret = adap->pre_xfer(i2c_adap); + if (ret < 0) + return -ENODEV; + } if (adap->getscl == NULL) pr_info("%s: Testing SDA only, SCL is not readable\n", name); @@ -297,11 +304,19 @@ static int test_bus(struct i2c_algo_bit_data *adap, char *name) "while pulling SCL high!\n", name); goto bailout; } + + if (adap->post_xfer) + adap->post_xfer(i2c_adap); + pr_info("%s: Test OK\n", name); return 0; bailout: sdahi(adap); sclhi(adap); + + if (adap->post_xfer) + adap->post_xfer(i2c_adap); + return -ENODEV; } @@ -607,7 +622,7 @@ static int __i2c_bit_add_bus(struct i2c_adapter *adap, int ret; if (bit_test) { - ret = test_bus(bit_adap, adap->name); + ret = test_bus(adap, adap->name); if (ret < 0) return -ENODEV; }
Apparently some distros set i2c-algo-bit.bit_test to 1 by default. In some cases this causes i2c_bit_add_bus to fail and prevents the i2c bus from being added. In the radeon case, we fail to add the ddc i2c buses which prevents the driver from being able to detect attached monitors. The i2c bus works fine even if bit_test fails. This is likely due to gpio switching that is required and handled in the pre/post_xfer hooks, so call the pre/post_xfer hooks in the bit test as well. Fixes: https://bugs.freedesktop.org/show_bug.cgi?id=36221 Cc: Jean Delvare <khali@linux-fr.org> Signed-off-by: Alex Deucher <alexdeucher@gmail.com> --- drivers/i2c/algos/i2c-algo-bit.c | 21 ++++++++++++++++++--- 1 files changed, 18 insertions(+), 3 deletions(-)