Date: Sun, 11 Aug 2019 21:21:32 +0000 (UTC) From: Ian Lepore <ian@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-12@freebsd.org Subject: svn commit: r350870 - stable/12/sys/dev/iicbus Message-ID: <201908112121.x7BLLWi7089301@repo.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: ian Date: Sun Aug 11 21:21:32 2019 New Revision: 350870 URL: https://svnweb.freebsd.org/changeset/base/350870 Log: MFC r349839, r349850 r349839: Call device_unbusy() on the error exit path, because if iicbus_request_bus() returns an error, iicbus_release_bus() is not going to be called. r349850: Restore the ability for i2c slave devices to do IO from their probe method. r348164 added code to iicbus_request_bus/iicbus_release_bus to automatically call device_busy()/device_unbusy() as part of aquiring exclusive use of the bus (so modules can't be unloaded while the bus is exclusively owned and/or IO is in progress). That broke the ability to do i2c IO from a slave device probe method, because the slave isn't attached yet, so calling device_busy() triggers a sanity-check panic for trying to busy a non-attached device. Now we check whether the device status is < DS_ATTACHING, and if so we busy the iicbus rather than the slave device. I think this leaves a small window where a module could be unloaded while probing is in progress. But I think that's true of all devices, and probably should be fixed by introducing a DS_PROBING state for devices, and handling that at various points in the newbus code. Modified: stable/12/sys/dev/iicbus/iicbus.h stable/12/sys/dev/iicbus/iiconf.c Directory Properties: stable/12/ (props changed) Modified: stable/12/sys/dev/iicbus/iicbus.h ============================================================================== --- stable/12/sys/dev/iicbus/iicbus.h Sun Aug 11 21:15:30 2019 (r350869) +++ stable/12/sys/dev/iicbus/iicbus.h Sun Aug 11 21:21:32 2019 (r350870) @@ -41,6 +41,7 @@ struct iicbus_softc { device_t dev; /* Myself */ device_t owner; /* iicbus owner device structure */ + device_t busydev; /* iicbus_release_bus calls unbusy on this */ u_int owncount; /* iicbus ownership nesting count */ u_char started; /* address of the 'started' slave * 0 if no start condition succeeded */ Modified: stable/12/sys/dev/iicbus/iiconf.c ============================================================================== --- stable/12/sys/dev/iicbus/iiconf.c Sun Aug 11 21:15:30 2019 (r350869) +++ stable/12/sys/dev/iicbus/iiconf.c Sun Aug 11 21:21:32 2019 (r350870) @@ -131,9 +131,15 @@ iicbus_request_bus(device_t bus, device_t dev, int how /* * Mark the device busy while it owns the bus, to * prevent detaching the device, bus, or hardware - * controller, until ownership is relinquished. + * controller, until ownership is relinquished. If the + * device is doing IO from its probe method before + * attaching, it cannot be busied; mark the bus busy. */ - device_busy(dev); + if (device_get_state(dev) < DS_ATTACHING) + sc->busydev = bus; + else + sc->busydev = dev; + device_busy(sc->busydev); /* * Drop the lock around the call to the bus driver, it * should be allowed to sleep in the IIC_WAIT case. @@ -150,6 +156,7 @@ iicbus_request_bus(device_t bus, device_t dev, int how sc->owner = NULL; sc->owncount = 0; wakeup_one(sc); + device_unbusy(sc->busydev); } } } @@ -183,7 +190,7 @@ iicbus_release_bus(device_t bus, device_t dev) IICBUS_LOCK(sc); sc->owner = NULL; wakeup_one(sc); - device_unbusy(dev); + device_unbusy(sc->busydev); } IICBUS_UNLOCK(sc); return (0);
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201908112121.x7BLLWi7089301>