The branch main has been updated by kp:

URL: 
https://cgit.FreeBSD.org/src/commit/?id=7647626d3380d17096b2690eb25f9d8fa18d0f83

commit 7647626d3380d17096b2690eb25f9d8fa18d0f83
Author:     Kristof Provost <k...@freebsd.org>
AuthorDate: 2022-08-10 13:57:05 +0000
Commit:     Kristof Provost <k...@freebsd.org>
CommitDate: 2022-08-19 21:09:11 +0000

    e6000sw: fix incorrect locking
    
    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 <h...@semihalf.com>
    Sponsored by:   Rubicon Communications, LLC ("Netgate")
    Differential Revision:  https://reviews.freebsd.org/D36117
---
 sys/dev/etherswitch/e6000sw/e6000sw.c | 9 ---------
 1 file changed, 9 deletions(-)

diff --git a/sys/dev/etherswitch/e6000sw/e6000sw.c 
b/sys/dev/etherswitch/e6000sw/e6000sw.c
index d096c31b349a..f3d46ff2f926 100644
--- a/sys/dev/etherswitch/e6000sw/e6000sw.c
+++ b/sys/dev/etherswitch/e6000sw/e6000sw.c
@@ -529,20 +529,11 @@ e6000sw_attach(device_t dev)
                if (!e6000sw_is_phyport(sc, port))
                        continue;
 
-               /*
-                * It's necessary to unlock mutex, because e6000sw_attach_miibus
-                * calls functions, which try to lock mutex.That leads
-                * to recursive lock on non recursive mutex.
-                */
-               E6000SW_UNLOCK(sc);
-
                err = e6000sw_attach_miibus(sc, port);
                if (err != 0) {
                        device_printf(sc->dev, "failed to attach miibus\n");
                        goto out_fail;
                }
-
-               E6000SW_LOCK(sc);
        }
 
        etherswitch_info.es_nports = sc->num_ports;

Reply via email to