src/sys/dev/spibus, branch releng/12.4

MFC r348169-r348170, r348172-r348173, r348183-r348184

2019-06-23T16:16:55Z

r348169: Define macros making it easier to define bus-specific pnpinfo for FDT systems. Pnpinfo is bus-specific and requires the bus name. The FDTCOMPAT_PNP_INFO() macro makes it easier to define new FDT-based pnpinfo for busses other than simplebus. Differential Revision: https://reviews.freebsd.org/D20382 r348170: Add pnpinfo for icee(4) on fdt systems. r348172: Use the new FDTCOMPAT_PNP_INFO() macro to define SPIBUS_FDT_PNP_INFO(). Also rename SPIBUS_PNP_INFO -> SPIBUS_FDT_PNP_INFO because there could be other kinds of pnpinfo for other (non-fdt) bus attachments. r348173: Rename IICBUS_FDT_PNPINFO -> IICBUS_FDT_PNP_INFO because all the other existing pnpinfo-related macros right now use PNP_INFO, not PNPINFO. r348183: Add pnpinfo. r348184: Add pnpinfo to all i2c drivers that have FDT compat data.

MFC r344556:

2019-04-22T00:49:13Z

Set maximum bus clock speed from hints when attaching hinted spibus(4) children. Some devices (such as spigen(4)) document that this works, but it appears that the code to implement it never got added.

Reapply, with minor tweaks, r338025, from the original commit:

2018-09-26T17:12:14Z

Remove unused and easy to misuse PNP macro parameter Inspired by r338025, just remove the element size parameter to the MODULE_PNP_INFO macro entirely. The 'table' parameter is now required to have correct pointer (or array) type. Since all invocations of the macro already had this property and the emitted PNP data continues to include the element size, there is no functional change. Mostly done with the coccinelle 'spatch' tool: $ cat modpnpsize0.cocci @normaltables@ identifier b,c; expression a,d,e; declarer MODULE_PNP_INFO; @@ MODULE_PNP_INFO(a,b,c,d, -sizeof(d[0]), e); @singletons@ identifier b,c,d; expression a; declarer MODULE_PNP_INFO; @@ MODULE_PNP_INFO(a,b,c,&d, -sizeof(d), 1); $ rg -l MODULE_PNP_INFO -- sys | \ xargs spatch --in-place --sp-file modpnpsize0.cocci (Note that coccinelle invokes diff(1) via a PATH search and expects diff to tolerate the -B flag, which BSD diff does not. So I had to link gdiff into PATH as diff to use spatch.) Tinderbox'd (-DMAKE_JUST_KERNELS). Approved by: re (glen)

Back out r338035 until Warner is finished churning GSoC PNP patches

2018-08-19T00:46:22Z

I was not aware Warner was making or planning to make forward progress in this area and have since been informed of that. It's easy to apply/reapply when churn dies down.

Remove unused and easy to misuse PNP macro parameter

2018-08-19T00:22:21Z

Inspired by r338025, just remove the element size parameter to the MODULE_PNP_INFO macro entirely. The 'table' parameter is now required to have correct pointer (or array) type. Since all invocations of the macro already had this property and the emitted PNP data continues to include the element size, there is no functional change. Mostly done with the coccinelle 'spatch' tool: $ cat modpnpsize0.cocci @normaltables@ identifier b,c; expression a,d,e; declarer MODULE_PNP_INFO; @@ MODULE_PNP_INFO(a,b,c,d, -sizeof(d[0]), e); @singletons@ identifier b,c,d; expression a; declarer MODULE_PNP_INFO; @@ MODULE_PNP_INFO(a,b,c,&d, -sizeof(d), 1); $ rg -l MODULE_PNP_INFO -- sys | \ xargs spatch --in-place --sp-file modpnpsize0.cocci (Note that coccinelle invokes diff(1) via a PATH search and expects diff to tolerate the -B flag, which BSD diff does not. So I had to link gdiff into PATH as diff to use spatch.) Tinderbox'd (-DMAKE_JUST_KERNELS).

Enhancements and fixes for the spigen(4) driver...

2018-07-11T17:54:41Z

- Resources used by spigen_mmap_single() are now tracked using devfs_set_cdevpriv() rather than in the softc. - Since resources are now tracked per-open-fd, there is no need to try to impose any exclusive-open logic, so flags related to that are removed. - Flags used to track open status to prevent detach() when the device is open are replaced with calls to device_busy()/device_unbusy(). That extends the protection up the hierarchy so that the spibus and hardware controller drivers also can't be detached while the device is open/in use. - Arbitrary limits on the maximum size of a transfer are removed, along with the sysctl variables that allowed the limits to be changed. There is just no reason to limit the size of a spi transfer to the machine's page size. Or to any other arbitrary value, really. - Most of the locking is removed. It was mostly protecting access to flags and fields in the softc that no longer exist. The locking that remains is just to prevent concurrent calls to device_[un]busy(). - The code was calling malloc() with M_WAITOK while holding a mutex in several places. Since most of the locking is gone, that's fixed.

Incorporate bus and chip select numbers into spigen(4) cdev names. Rather

2018-06-21T21:16:26Z

than assigning spigen device names in order of creation, this uses a device name that corresponds to the owning spibus and chip-select index. Example: /dev/spigen0.1 would be a child of spibus0, and use cs = 1 The intent is for systems like Raspberry Pi to have a consistent way of using an SPI interface with a specific cs value from a user application. Otherwise, there is no consistent way of knowing which cs pin will be assigned to a particular spigen device. The alternative is to specify everything in "the right order" in an overlay file, which is less than ideal. Additionally, this duplicates (to some extent) the way Linux handles a similar situation with their 'spidev' device, so it would be somewhat familiar to those who also use Linux. A new kernel config option, SPIGEN_LEGACY_CDEVNAME, causes the driver to also create /dev/spigenN device name aliases, with N incrementing in the order of device instantiation. This is provided to ease the transition for existing systems using the original naming convention (particularly when these changes are MFC'd to stable branches). Differential Revision: https://reviews.freebsd.org/D15301

spi: Add SPIBUS_PNP_INFO macro

2018-06-14T17:20:47Z

The PNP info string is the same as the SIMPLEBUS one but driver should depend on spibus and not simplebus

Allow hinted attachment on FDT-based systems. Instead of returning ENXIO

2018-04-08T17:06:30Z

when the FDT data doesn't enable the device instance, return BUS_PROBE_NOWILDCARD, the same as for non-FDT systems.

Cast the data pointer to the correct type for the data being accessed (as

2018-04-07T22:21:06Z

opposed to one that accidentally worked on the one arch I test-compiled for on my first try). Reported by: np@, O. Hartmann Pointy hat: ian@