Return with the PORT_STATUS register value instead of
writing that directly into the corresponding register.
Also rename the function to ar8327_get_port_init_status.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 36043
The presence of the platform data is already verified in
ar8327_hw_init, and the driver does not start without that
anyway.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 36042
Setting this bit stops BCM53125 (bgmac actually) from receiving any
packets. This bit is cleared conditionally in b53_switch_reset and it
seems the same is done in bcmrobo.c which never sets that bit again.
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 35723
Currently the switch gets registered when an ethernet
driver connects to a PHY of the switch. This method does
not work with the ethernet drivers which are connecting
to the PHY from their ndo_open callback. With those
ethernet drivers, the driver tries to register the switch
each time when the etherned device is opened and this causes
a deadlock.
Move the switch registration into the probe routine to fix
this problem.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 35602
The switches are accessed via an MDIO bus. Set the alias
to the name of the MDIO bus, and show that in the message
along with the name of the switch switch device.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 35560
The private data of the switch is already
allocated in ar8216_priv, assign that to
each PHY on the same MDIO bus. Also remove
the redundant code from ar8216_config_init.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 35559
Sujith says:
This commit breaks the WAN port on my AP96 - DHCP fails.
Reverting it fixes the issue.
commit b67cc3a0cdd02973610d4d5a63226d1c44841e94
Author: juhosg <juhosg@3c298f89-4303-0410-b956-a3cf2f4a3e73>
Date: Fri Feb 8 09:13:18 2013 +0000
generic: ar8216: simplify phy features setup
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/trunk@355183c298f89-4303-0410-b956-a3cf2f4a3e73
This change restores the previous behaviour and moves
the code into the ar8216_probe function.
Reported-by: Sujith Manoharan <sujith@msujith.org>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Tested-by: Sujith Manoharan <sujith@msujith.org>
SVN-Revision: 35549
b53_switch_detect returns value returned by b53_read8, which is 0 for
success. So fail (and return error) only if b53_switch_detect returned
something else than 0. This fixes supported and advertising being zeros
for MDIO access.
Cc: Jonas Gorski <jogo@openwrt.org>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
SVN-Revision: 35534
It is enough to sleep once after ANEG is enabled on
all PHYs. This reduces boot time on boards with
AR8316 switch.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 35532
Use the detach callback to disable packet mangling
to avoid possible NULL pointer dereference.
Compile tested only.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 35512
Add swconfig switch driver for Broadcom BCM53XX switch chips. Supports
switches connected through MDIO, SPI or memory mapped registers, and
supports BCM5325, BCM539x, BCM531x5 and the BCM63XX internal switch
chips.
Tested are BCM5325 trough MDIO, BCM53115 through SPI, and BCM6328.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 35305
Only assign a pvid if the current pvid is 0. Fixes using mixed tagged
and untagged traffic on a port and the untagged vlan isn't the last one
of which the port is a member.
Tested on RTL8366S and RTL8367R.
Fixes#8501.
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 35227
CC drivers/net/phy/rtl8366rb.o
In file included from drivers/net/phy/rtl8366s.c:16:0:
include/linux/of_platform.h:106:13: warning: 'struct device' declared inside parameter list [enabled by default]
include/linux/of_platform.h:106:13: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
drivers/net/phy/rtl8366s.c: In function 'rtl8366s_probe':
drivers/net/phy/rtl8366s.c:1094:2: warning: label 'err_out' defined but not used [-Wunused-label]
drivers/net/phy/rtl8366s.c:1055:32: warning: unused variable 'pdata'[-Wunused-variable]
CC drivers/net/phy/rtl8366rb.o
In file included from drivers/net/phy/rtl8366rb.c:17:0:
include/linux/of_platform.h:106:13: warning: 'struct device' declared inside parameter list [enabled by default]
include/linux/of_platform.h:106:13: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
drivers/net/phy/rtl8366rb.c: In function 'rtl8366rb_probe':
drivers/net/phy/rtl8366rb.c:1214:2: warning: label 'err_out' defined but not used [-Wunused-label]
drivers/net/phy/rtl8366rb.c:1175:32: warning: unused variable 'pdata'[-Wunused-variable]
CC drivers/net/phy/rtl8367.o
In file included from drivers/net/phy/rtl8367.c:15:0:
include/linux/of_platform.h:106:13: warning: 'struct device' declared inside parameter list [enabled by default]
include/linux/of_platform.h:106:13: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
drivers/net/phy/rtl8367.c: In function 'rtl8367_probe':
drivers/net/phy/rtl8367.c:1712:2: warning: label 'err_out' defined but not used [-Wunused-label]
drivers/net/phy/rtl8367.c:1677:32: warning: unused variable 'pdata' [-Wunused-variable]
CC drivers/net/phy/rtl8367b.o
In file included from drivers/net/phy/rtl8367b.c:15:0:
include/linux/of_platform.h:106:13: warning: 'struct device' declared inside parameter list [enabled by default]
include/linux/of_platform.h:106:13: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
drivers/net/phy/rtl8367b.c: In function 'rtl8367b_probe':
drivers/net/phy/rtl8367b.c:1494:2: warning: label 'err_out' defined but not used [-Wunused-label]
drivers/net/phy/rtl8367b.c:1459:32: warning: unused variable 'pdata' [-Wunused-variable]
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 34731
fixes the following errors caused by r34682:
CC [M] drivers/net/phy/rtl8366_smi.o
In file included from drivers/net/phy/rtl8366_smi.c:26:0:
drivers/net/phy/rtl8366_smi.h:149:46: warning: 'struct platform_device' declared inside parameter list [enabled by default]
drivers/net/phy/rtl8366_smi.h:149:46: warning: its scope is only this definition or declaration, which is probably not what you want [enabled by default]
drivers/net/phy/rtl8366_smi.c:1398:65: warning: 'struct platform_device' declared inside parameter list [enabled by default]
drivers/net/phy/rtl8366_smi.c: In function 'rtl8366_smi_probe_plat':
drivers/net/phy/rtl8366_smi.c:1400:44: error: dereferencing pointer to incomplete type
drivers/net/phy/rtl8366_smi.c:1402:11: error: dereferencing pointer to incomplete type
drivers/net/phy/rtl8366_smi.c:1403:16: error: dereferencing pointer to incomplete type
drivers/net/phy/rtl8366_smi.c: At top level:
drivers/net/phy/rtl8366_smi.c:1415:46: warning: 'struct platform_device' declared inside parameter list [enabled by default]
drivers/net/phy/rtl8366_smi.c:1415:21: error: conflicting types for 'rtl8366_smi_probe'
drivers/net/phy/rtl8366_smi.h:149:21: note: previous declaration of 'rtl8366_smi_probe' was here
drivers/net/phy/rtl8366_smi.c: In function 'rtl8366_smi_probe':
drivers/net/phy/rtl8366_smi.c:1420:31: error: dereferencing pointer to incomplete type
drivers/net/phy/rtl8366_smi.c:1424:10: error: dereferencing pointer to incomplete type
drivers/net/phy/rtl8366_smi.c:1425:3: warning: passing argument 1 of 'rtl8366_smi_probe_of' from incompatible pointer type [enabled by default]
drivers/net/phy/rtl8366_smi.c:1392:19: note: expected 'struct device_node *' but argument is of type 'struct platform_device *'
drivers/net/phy/rtl8366_smi.c:1427:3: warning: passing argument 1 of 'rtl8366_smi_probe_plat' from incompatible pointer type [enabled by default]
drivers/net/phy/rtl8366_smi.c:1398:5: note: expected 'struct platform_device *' but argument is of type 'struct platform_device *'
drivers/net/phy/rtl8366_smi.c: At top level:
drivers/net/phy/rtl8366_smi.c:1438:1: error: conflicting types for 'rtl8366_smi_probe'
drivers/net/phy/rtl8366_smi.h:149:21: note: previous declaration of 'rtl8366_smi_probe' was here
make[7]: *** [drivers/net/phy/rtl8366_smi.o] Error 1
make[6]: *** [drivers/net/phy] Error 2
make[5]: *** [drivers/net] Error 2
make[4]: *** [drivers] Error 2
Signed-off-by: Jonas Gorski <jogo@openwrt.org>
SVN-Revision: 34703
The ar8216_mib_{capture,flush} functions are similar.
Move the common code into a helper and use that from
the original functions.
This change also fixes a lockdep warning in
ar8216_mib_flush.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 34297
swconfig does not compile with 3.7 because
struct netlink_notify->pid has been renamed to
struct netlink_notify->portid
and
struct genl_info->snd_pid has been renamed to
struct genl_info->snd_portid
Use the new names and add compatibility patches
to 3.3 and 3.6.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 34288
The names are misleading, those are only valid from
AR8236. The AR8216 chip uses different offsets.
Thanks to Jonas Gorski for the report.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 34244
The 'led_brightness_set' function has
been renamed to 'led_set_brightness'
in recent ernels. Use the new name and
add a compatibility patch for 3.3.
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
SVN-Revision: 33946
Switch reset on AR8316 appears to clobber the MTU configuration register
and possibly other global config registers. Move global configuration
register init writes back to the reset callback.
SVN-Revision: 30951
After issuing a soft reset on the RT8366{S,RB}
switch, waiting for the last acknowlegement fails
in rtl8366_smi_write_reg. Add a _noack version of
the function and use that for issuing a soft reset.
SVN-Revision: 29844
This patch adds the option to enable jumbo frames on the rtl8366s
switch. This is almost exactly the same as the first patch at
https://dev.openwrt.org/ticket/7977 and it has been requested
in a comment there.
I have tested this and it successfully allows devices on the switch
using jumbo frames to communicate.
Signed-off-by: Ash Hughes <ashley.hughes@blueyonder.co.uk>
SVN-Revision: 28267
Also make switches available under a generic name "switch<num>" for device
name agnostic access. The old device name is used as an alias for backward
compatibility.
SVN-Revision: 27800
This patch adds 802.1Q VLAN support for the ADM6996M chip.
The driver is loaded for both the FC and M model. It will detect which of the
two chips is connected. The FC model is initialised, but no further
functionality is offered.
The PHY driver will always report "100 Mbit/s, link up", for both the M and FC
models. This reflects the fact that the link between switch chip and Ethernet
MAC is always on[1].
Further documentation can be found in the kernel's
Documentation/networking/adm6996.txt
Signed-of-By: Peter Lebbing <peter@digitalbrains.com>
SVN-Revision: 26865
SPIN_LOCK_UNLOCKED has been deprecated since 2.6.19 and will get removed
in 2.6.39, so replace it with DEFINE_SPINLOCK().
Signed-off-by: Jonas Gorski <jonas.gorski+openwrt@gmail.com>
SVN-Revision: 26770
Since the PHY driver is only used for the WAN port and there is virtually
no difference between it and the generic PHY driver, we can sefely remove
it.
Signed-off-by: Jonas Gorski <jonas.gorski+openwrt@gmail.com>
SVN-Revision: 26600
Allows to use fids 0 to 7. Default is 0.
Those who will use this don't forget to `set vlan fid` every time
after `set vlan ports` if you want non default behavior.
Minor changes. Version bump.
Signed-off-by: Roman Yeryomin <roman@advem.lv>
SVN-Revision: 24125
The RTL8366S/RB support connecting a second mac to it and using the 5th
port as an independent phy. Therefore we need to handle autonegotiation
changes for phy 4 like on a single phy device.
This fixes#7980, verified on a WNDR3700.
Signed-off-by: Jonas Gorski <jonas.gorski+openwrt@gmail.com>
SVN-Revision: 23127
Tested on ifxmips based board. The PSB6970 is also known as the Infineon Tantos switch.
Signed-off-by: Ithamar R. Adema <ithamar.adema@team-embedded.nl>
SVN-Revision: 22914
The usage of bitfields causes problems in Little-Endian mode. Remove
them and manipulate the bits directly in the appropriate functions. This
change also reduces the size of the binary a bit.
SVN-Revision: 22192
