Hello,
as described in the commit log of the 2nd patch returning an error code from a bus' remove callback doesn't make any difference as the driver core ignores it and still considers the device removed.
So change the remove callback to return void to not give driver authors an incentive to believe they could return an error.
There is only a single isa driver in the tree (assuming I didn't miss any) that has a remove callback that can return a non zero return code. This is "fixed" in the first patch, to make the second patch more obviously correct.
Best regards Uwe
Uwe Kleine-König (2): watchdog: pcwd: drop always-false if from remove callback isa: Make the remove callback for isa drivers return void
drivers/base/isa.c | 2 +- drivers/i2c/busses/i2c-elektor.c | 4 +--- drivers/i2c/busses/i2c-pca-isa.c | 4 +--- drivers/input/touchscreen/htcpen.c | 4 +--- drivers/media/radio/radio-sf16fmr2.c | 4 +--- drivers/net/can/sja1000/tscan1.c | 4 +--- drivers/net/ethernet/3com/3c509.c | 3 +-- drivers/scsi/advansys.c | 3 +-- drivers/scsi/aha1542.c | 3 +-- drivers/scsi/fdomain_isa.c | 3 +-- drivers/scsi/g_NCR5380.c | 3 +-- drivers/watchdog/pcwd.c | 7 +------ include/linux/isa.h | 2 +- sound/isa/ad1848/ad1848.c | 3 +-- sound/isa/adlib.c | 3 +-- sound/isa/cmi8328.c | 3 +-- sound/isa/cmi8330.c | 3 +-- sound/isa/cs423x/cs4231.c | 3 +-- sound/isa/cs423x/cs4236.c | 3 +-- sound/isa/es1688/es1688.c | 3 +-- sound/isa/es18xx.c | 3 +-- sound/isa/galaxy/galaxy.c | 3 +-- sound/isa/gus/gusclassic.c | 3 +-- sound/isa/gus/gusextreme.c | 3 +-- sound/isa/gus/gusmax.c | 3 +-- sound/isa/gus/interwave.c | 3 +-- sound/isa/msnd/msnd_pinnacle.c | 3 +-- sound/isa/opl3sa2.c | 3 +-- sound/isa/opti9xx/miro.c | 3 +-- sound/isa/opti9xx/opti92x-ad1848.c | 3 +-- sound/isa/sb/jazz16.c | 3 +-- sound/isa/sb/sb16.c | 3 +-- sound/isa/sb/sb8.c | 3 +-- sound/isa/sc6000.c | 3 +-- sound/isa/sscape.c | 3 +-- sound/isa/wavefront/wavefront.c | 3 +-- 36 files changed, 36 insertions(+), 79 deletions(-)
base-commit: 5a158981aafa7f29709034b17bd007b15cb29983