src/sys/dev/mii, branch release/5.3.0_cvs FreeBSD source tree This commit was manufactured by cvs2svn to create tag 2004-11-04T19:12:42+00:00 cvs2svn cvs2svn@FreeBSD.org 2004-11-04T19:12:42+00:00 3f86d8a2ea3f3265afaa1fd263b0004c5c000e69 'RELENG_5_3_0_RELEASE'. This commit was manufactured to restore the state of the 5.3-RELEASE image. 
'RELENG_5_3_0_RELEASE'.

This commit was manufactured to restore the state of the 5.3-RELEASE image.
Merge the vge(4) driver from the HEAD. Also add nge(4) and lge(4) to 2004-09-14T21:22:34+00:00 Bill Paul wpaul@FreeBSD.org 2004-09-14T21:22:34+00:00 8257b2d6bd3d2f9b4febf9eced9e4e2cd3bb2d0b GENERIC for x86. Approved by: re (scottl) 
GENERIC for x86.

Approved by:	re (scottl)
This commit was manufactured by cvs2svn to create branch 'RELENG_5'. 2004-09-14T01:07:20+00:00 cvs2svn cvs2svn@FreeBSD.org 2004-09-14T01:07:20+00:00 4d991369186d62ecda4db6ce2f74213b1ff3454e 






Add locking to the kqueue subsystem.  This also makes the kqueue subsystem
2004-08-15T06:24:42+00:00

John-Mark Gurney
jmg@FreeBSD.org

2004-08-15T06:24:42+00:00

ad3b9257c2f09a6f1c0e56c1100aa4e43fb95e24

a more complete subsystem, and removes the knowlege of how things are
implemented from the drivers.  Include locking around filter ops, so a
module like aio will know when not to be unloaded if there are outstanding
knotes using it's filter ops.

Currently, it uses the MTX_DUPOK even though it is not always safe to
aquire duplicate locks.  Witness currently doesn't support the ability
to discover if a dup lock is ok (in some cases).

Reviewed by:	green, rwatson (both earlier versions)





a more complete subsystem, and removes the knowlege of how things are
implemented from the drivers.  Include locking around filter ops, so a
module like aio will know when not to be unloaded if there are outstanding
knotes using it's filter ops.

Currently, it uses the MTX_DUPOK even though it is not always safe to
aquire duplicate locks.  Witness currently doesn't support the ability
to discover if a dup lock is ok (in some cases).

Reviewed by:	green, rwatson (both earlier versions)







Add some missing <sys/module.h> includes which are masked by the
2004-05-30T17:57:46+00:00

Poul-Henning Kamp
phk@FreeBSD.org

2004-05-30T17:57:46+00:00

41ee9f1c6918a36d2feca1640bde21b420618948

one on death-row in <sys/kernel.h>





one on death-row in <sys/kernel.h>







DP83840A on hme(4) don't advertise their media capabilities themselves
2004-05-29T18:23:26+00:00

Marius Strobl
marius@FreeBSD.org

2004-05-29T18:23:26+00:00

f0e531572d9dd5057c120928b19b989a4e7e5c32

properly. This causes the autonegotiation to e.g. never establish a
100baseTX full-duplex link. The solution to this problem is to manually
write the capabilities from the BMSR to the ANAR every time a media
change occurs, even when already in autonegotiation mode.
The NetBSD way of doing this is to set their MIIF_FORCEANEG flag in the
NIC driver. This causes mii_phy_setmedia() to call mii_phy_auto() (which
will set the ANAR according to the BMSR) even when the PHY alread is in
autonegotiation mode. However, while doing the same on FreeBSD (which
involves porting the MIIF_FORCEANEG flag and converting nsphy.c to use
mii_phy_setmedia()) fixes autonegotiation, using mii_phy_setmedia()
causes this driver to no longer work properly in the other modes.
Another drawback of that approach is that this will also force writing
the ANAR on other PHYs whose drivers use mii_phy_setmedia() and which
are used with a NIC whose driver sets MIIF_FORCEANEG (e.g. hme(4) is
known to be used together with 3 different PHYs while only the DP83840A
require this workaround).
So instead of moving to MIIF_FORCEANEG, just call mii_phy_auto() in
nsphy_service() unconditionally when hanging off of a hme(4) and serving
a media change

This is part 1/2 of fixing autonegotiation on hme(4) using DP83840A PHYs.





properly. This causes the autonegotiation to e.g. never establish a
100baseTX full-duplex link. The solution to this problem is to manually
write the capabilities from the BMSR to the ANAR every time a media
change occurs, even when already in autonegotiation mode.
The NetBSD way of doing this is to set their MIIF_FORCEANEG flag in the
NIC driver. This causes mii_phy_setmedia() to call mii_phy_auto() (which
will set the ANAR according to the BMSR) even when the PHY alread is in
autonegotiation mode. However, while doing the same on FreeBSD (which
involves porting the MIIF_FORCEANEG flag and converting nsphy.c to use
mii_phy_setmedia()) fixes autonegotiation, using mii_phy_setmedia()
causes this driver to no longer work properly in the other modes.
Another drawback of that approach is that this will also force writing
the ANAR on other PHYs whose drivers use mii_phy_setmedia() and which
are used with a NIC whose driver sets MIIF_FORCEANEG (e.g. hme(4) is
known to be used together with 3 different PHYs while only the DP83840A
require this workaround).
So instead of moving to MIIF_FORCEANEG, just call mii_phy_auto() in
nsphy_service() unconditionally when hanging off of a hme(4) and serving
a media change

This is part 1/2 of fixing autonegotiation on hme(4) using DP83840A PHYs.







Spelling fix in a comment.
2004-05-29T18:13:44+00:00

Marius Strobl
marius@FreeBSD.org

2004-05-29T18:13:44+00:00

b78791d37ed5ec9b41612054ba1e94314af752aa












Remove double __FBSDID and move the remaining one into a common place after
2004-05-29T18:09:10+00:00

Marius Strobl
marius@FreeBSD.org

2004-05-29T18:09:10+00:00

50aa106108cc9a610e4bdee7c7fe614a8c425c18

the license(s) and before the driver comment (the latter only in drivers not
having __FBSDID at that location).





the license(s) and before the driver comment (the latter only in drivers not
having __FBSDID at that location).







Move __FBSDID out from under a comment.
2004-05-29T17:45:45+00:00

Marius Strobl
marius@FreeBSD.org

2004-05-29T17:45:45+00:00

c4386506c7c9cec6c0a8e7a9407d3265e4a3e330












Spelling fix in a comment.
2004-05-29T16:54:59+00:00

Marius Strobl
marius@FreeBSD.org

2004-05-29T16:54:59+00:00

5aa54809198b2bb201ec3a9555e204d158fe68ac