From: Vladimir V. <vla...@li...> - 2022-12-01 12:14:00
|
Hi, With the 4.3.5 release using the gpib_bitbang.ko driver caused crashes in kernel space which are now seemingly resolved with the modified spinlock handling in the latest trunk r2046. We have updated our hackster guide to use the debian package we generate from that svn trunk revision. * Debian package repository: https://github.com/lightside-instruments/gpib-debian/tree/debian/4.3.5-lsi4 * Hackster guide: https://www.hackster.io/lightside-instruments/the-gpib4pi-gpib-for-raspberry-pi-hat-4b3e9a /Vladimir On 11/8/22 00:35, Vladimir Vassilev wrote: > Dave, > > Yes that was the issue .. gpio27 was swapped with pin27! > > One needs the following patch to workaround the issue. (`modprobe > gpib_bitbang board_id=gpib4pi-1.1`): > > --- a/linux-gpib-kernel/drivers/gpib/gpio/gpib_bitbang.c > +++ b/linux-gpib-kernel/drivers/gpib/gpio/gpib_bitbang.c > @@ -91,6 +91,10 @@ > static int sn7516x_used=1; > module_param(sn7516x_used,int,0660); > > +static char *board_id = "default"; > +module_param(board_id, charp, 0660); > +MODULE_PARM_DESC(board_id, "Valid values: default, gpib4pi-1.1"); > + > /********************************************** > * Signal pairing and pin wiring between the * > * Raspberry-Pi connector and the GPIB bus * > @@ -914,6 +918,12 @@ int bb_attach(gpib_board_t *board, const > gpib_board_config_t *config) > priv->direction = -1; > priv->t1_delay = 2000; > > + if(0==strcmp("gpib4pi-1.1", board_id)) { > + dbg_printk(3,"Using pin mapping for %s\n", board_id); > + > + gpios_vector[12] = 0; /* 27 -> 0 */ > + } > + > if (allocate_gpios()) return -EBUSY; > > if (sn7516x_used) { > > > /Vladimir > > On 11/5/22 11:48, dave penkler wrote: >> Hi Vladimir, >> GPIO0 corresponds to pin 27 on the Pi header usually reserved for i2c >> EERROM. What is it connected to, if anything, on the lsi-gpib4pi-1.1 >> shield ? >> cheers, >> -Dave >> >> |