Unloading the e6000sw driver with a "fixed" ixgbe (which is doing
MDIO transfers faster than 8ms per) has exposed another fun race
condition - the MII busses were being torn down before the etherswitch
device.
* Modify e6000sw_miiforphy() to return NULL if the mii bus device
isn't setup, which stops the panic
* Change the module order so the e6000sw module is detached first,
before the miibus entries and attached PHYs are destroyed.
This ensures that the miibus entries aren't destroyed outside
of the driver lock, and e6000sw_tick() doesn't try dereferencing
dead miibus device_t's.
Differential Revision: https://reviews.freebsd.org/D50294
Reviewed by: jhb
This adds the minimum support required to probe/attach the 88E6190X.
I've tested this against an AT&T ATT-150 OCP device (Silicom i3000)
with local changes to export MDIO via ixge(4).
Hints are required to probe/attach/configure the switch on amd64,
but with the mentioned diffs, it does work.
Thanks to Stas Alekseev <stas@alekseev.us> for the pull request
and Stas / Jason Hensler <omegadraconis@gmail.com> for chasing
down information about the chipset, linux stuff and AT&T OCP
hardware information.
PR: kern/281211
Pull Request: https://github.com/freebsd/freebsd-src/pull/1408
Differential Revision: https://reviews.freebsd.org/D50044
Reviewed by: imp
Although all the machinery is present, the tick routine only ran once.
It was never rescheduled. So, reschedule it.
However in practice I've discovered that the tick routine is running
whilst a bunch of phys are actually being probe/attached on each per-port
MII bus being created and probed off of the switch itself.
Until that's debugged (and likely the whole PHY management stuff is
cleaned up here), just add a NULL check and thus don't panic.
Differential Revision: https://reviews.freebsd.org/D50031
Reviewed by: imp
Although the tick isn't running every hz right now, when it /is/
running at hz, the shutdown path will race with an existing running
tick routine, causing unpredictable panics.
* Introduce a shutdown flag which will abort doing the tick work if set
* set the shutdown flag and start cancel/draining the taskqueue during
detach.
Differential Revision: https://reviews.freebsd.org/D50030
The driver sleep lock was being held during most of the error paths,
and not unlocking it will panic the kernel during detach. So, fix it.
Differential Revision: https://reviews.freebsd.org/D50029
Reviewed by: imp
This fixes the driver to load correctly as a module when etherswitch
is also a module.
Differential Revision: https://reviews.freebsd.org/D50027
Reviewed by: imp
Call bus_generic_detach first and return any error. Remove no longer
needed individual device_delete_child calls.
Differential Revision: https://reviews.freebsd.org/D47970
While here, check for errors from bus_generic_detach and move it to
the start of detach if necessary.
Differential Revision: https://reviews.freebsd.org/D47969
In some cases, move the call to bus_generic_detach earlier so that any
detach failures from child devices do not leave the parent device
partially detached.
Differential Revision: https://reviews.freebsd.org/D47966
This provides better semantics as a standalone DEVMETHOD for
device_attach as bus drivers should remove child devices they created
as part of detach cleanup. The implementation calls
bus_detach_children() first to permit child devices an opportunity to
veto the detach operation. If that succeeds, device_delete_children()
is used to delete the child devices.
This requires fixing various drivers that were deleting devices
explicitly (via a device_t pointer cached in the softc) after calling
bus_generic_detach to stop doing that and just rely on
bus_generic_detach to remove child devices.
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D47959
Change 4787572d05 made if_alloc_domain() never fail, then also do the
wrappers if_alloc(), if_alloc_dev(), and if_gethandle().
No functional change intended.
Reviewed by: kp, imp, glebius, stevek
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D45740
We've removed kernel option EXT_RESOURCES almost two years ago.
While it was ok to have some code under a common 'extres' subdirectory
at first, we now have a lot of consumer of it and we made it mandatory
so no need to have it under a cryptic name.
Reviewed by: imp
Sponsored by: Beckhoff Automation GmbH & Co. KG
Differential Revision: https://reviews.freebsd.org/D43192
We've removed kernel option EXT_RESOURCES almost two years ago.
While it was ok to have some code under a common 'extres' subdirectory
at first, we now have a lot of consumer of it and we made it mandatory
so no need to have it under a cryptic name.
Reviewed by: mhorne
Sponsored by: Beckhoff Automation GmbH & Co. KG
Differential Revision: https://reviews.freebsd.org/D43191
Apply the following automated changes to try to eliminate
no-longer-needed sys/cdefs.h includes as well as now-empty
blank lines in a row.
Remove /^#if.*\n#endif.*\n#include\s+<sys/cdefs.h>.*\n/
Remove /\n+#include\s+<sys/cdefs.h>.*\n+#if.*\n#endif.*\n+/
Remove /\n+#if.*\n#endif.*\n+/
Remove /^#if.*\n#endif.*\n/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/types.h>/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/param.h>/
Remove /\n+#include\s+<sys/cdefs.h>\n#include\s+<sys/capsicum.h>/
Sponsored by: Netflix
Commit 4692906480 made e6000sw's
implementation of miibus_(read|write)reg assume that the softc lock is
held. I presume that is to avoid lock recursion in e6000sw_attach() ->
e6000sw_attach_miibus() -> mii_attach() -> MIIBUS_READREG().
However, the lock assertion in e6000sw_readphy_locked() can fail if a
different driver uses the interface to probe registers. Work around the
problem by providing implementations which lock the softc if it is not
already locked.
PR: 274795
Fixes: 4692906480 ("e6000sw: add readphy and writephy wrappers")
Reviewed by: kp, imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D42466
With the general removal of MIPS support there's no longer a need to
support these integrated switches.
Approved by: adrian
Differential Revision: https://reviews.freebsd.org/D41394
During porting of etherswitch to NetBSD mistypo was discovered in
Atheros switch version detection.
Reported by: Hiroki Mori yamori813@yahoo.co.jp
MFC after: 1 week
The SPDX folks have obsoleted the BSD-2-Clause-FreeBSD identifier. Catch
up to that fact and revert to their recommended match of BSD-2-Clause.
Discussed with: pfg
MFC After: 3 days
Sponsored by: Netflix
During attach we release the lock to call e6000sw_attach_miibus(), which
calls mii_attach(). The mii_attach() function calls miibus_readreg() /
miibus_writereg(). However, these are set to be
e6000sw_readphy_locked().
That is, the read/write phy functions do not acquire the lock, but
expect to be called while locked. Simply do not unlock and relock while
calling mii_attach().
Reviewed by: Hubert Mazur <hum@semihalf.com>
Sponsored by: Rubicon Communications, LLC ("Netgate")
Differential Revision: https://reviews.freebsd.org/D36117
New functions take lock and give lock after operation.
Removed locking and unlocking within other methods,
to prevent from recursive locking on non recursive lock.
Tested on EspressoBin.
Reviewed by: mw
Obtained from: Semihalf
Differential Revision: https://reviews.freebsd.org/D35656
When accessing a register directly from etherswitchcfg one must specify
a register group(e.g. registers of portN) and the register offset within
the group. The latter is passed as the 5 least significant bits.
Extract the former by dividing the register address by 32, not by 5.
Approved by: mw(mentor)
Obtained from: Semihalf
Sponsored by: Stormshield
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D35672
