Commit Graph

88 Commits (8b46a72583c072a809149f05e8bdc3dd5c6e0d47)

Author SHA1 Message Date
Florian Fainelli 88864e2aa1 avoid namespace clash on ar7_gpio_init
SVN-Revision: 23917
2010-11-07 22:25:25 +00:00
Florian Fainelli 00e15a3abf add 2.6.35 support
SVN-Revision: 22812
2010-08-26 09:13:44 +00:00
Florian Fainelli dad3df8582 move files to file-2.6.30
SVN-Revision: 19518
2010-02-04 14:35:21 +00:00
Florian Fainelli 2239295892 add UPF_IOREMAP flag to the uart port definition (#6532)
SVN-Revision: 19374
2010-01-28 21:17:21 +00:00
Florian Fainelli 5a481b16ad add support for titan variant (WRTP54G) (#6185)
SVN-Revision: 19022
2010-01-03 20:19:32 +00:00
Florian Fainelli cf176f82db add experimental support for 2.6.30
SVN-Revision: 17097
2009-08-02 13:01:18 +00:00
Florian Fainelli c739ce1888 remove old 2.6.24 code
SVN-Revision: 16240
2009-06-01 08:04:56 +00:00
Florian Fainelli fc72d90103 revert part of r16112, the setup.c part needs more testing
SVN-Revision: 16113
2009-05-27 16:22:01 +00:00
Florian Fainelli 33f6145c51 include missing atm-tools to setup pppoe/a links
SVN-Revision: 16112
2009-05-27 16:20:21 +00:00
Florian Fainelli 3b92b4de00 I recently came across an ar7 device which has the vlynq hardwired so that the clocks are always generated by the remote device instead of the local one.
Upon initialization the current version of vlynq driver disables
remote clock generation and causes the entire bus to hang on my
device.

This patch adds support for detecting which device (local or remote)
is responsible of clock generation and implements clock
initialization based on detection result.

Signed-off-by: Antti Seppala <a.seppala at gmail.com>

SVN-Revision: 16049
2009-05-25 13:13:10 +00:00
Florian Fainelli 6239eb2c7b update to 2.6.27 (#4122)
SVN-Revision: 15665
2009-05-07 12:17:56 +00:00
Florian Fainelli 086b7e33f1 do not patch the local copy of ar7part.c to be exactly like what is mainline already
SVN-Revision: 14625
2009-02-22 18:48:55 +00:00
Felix Fietkau c3ae8337f4 AR7 watchdog supports magic close
This patch adds the magic close flag (stopping the watchdog on an expected close
by wrting 'V' to the device before closing it) to the watchdogs list, which the
AR7 watchdog code actually supports.

Signed-off-by: Axel Gembe <ago@bastart.eu.org>

SVN-Revision: 12666
2008-09-23 15:43:30 +00:00
Florian Fainelli 0fdecd35bd Treat vlynq external divisor just like automatic, fix comment about it, thanks sn9
SVN-Revision: 12467
2008-09-01 19:29:25 +00:00
Florian Fainelli 3ab8360af0 Fix divisor calculation and configuration from previous commit, thanks sn9
SVN-Revision: 12454
2008-09-01 11:49:57 +00:00
Florian Fainelli 3abe371efd Fix divisor settings for external devices like wireless devices, thanks sn9
SVN-Revision: 12443
2008-08-31 11:32:26 +00:00
Gabor Juhos 936b0e3cc7 use the righ value in the DSP clock calculation
SVN-Revision: 11123
2008-05-12 11:56:21 +00:00
Matteo Croce 4a85d8edac ar7: remove volatiles definitely
SVN-Revision: 10762
2008-04-07 15:22:31 +00:00
Matteo Croce f9f4afa8ff vlynq: revert volatiles until we find a good way to remove them
SVN-Revision: 10760
2008-04-07 14:30:20 +00:00
Matteo Croce d84f27f264 ar7: remove useless volatiles
SVN-Revision: 10757
2008-04-07 13:49:44 +00:00
Matteo Croce 78e07d75cd ar7: remove unneeded packed and array initialization
SVN-Revision: 10752
2008-04-07 01:30:07 +00:00
Matteo Croce 67dc179e05 vlynq: small fixes
SVN-Revision: 10711
2008-04-02 14:50:19 +00:00
Matteo Croce af4c2c041f add proper email addresses to the comment headers
SVN-Revision: 10709
2008-04-02 11:45:18 +00:00
Matteo Croce c96a4d1808 Let authors holds copyright of the AR7 code (closes #2369)
SVN-Revision: 10708
2008-04-02 11:18:00 +00:00
Matteo Croce 9b93a436c7 vlynq: probe for an external clock first, needed to enable acx on the Leonardo board
SVN-Revision: 10707
2008-04-02 11:06:50 +00:00
Matteo Croce b2b102fdcc Update kernel to 2.6.24
SVN-Revision: 10421
2008-02-08 17:17:45 +00:00
Matteo Croce fb5be81fe7 make current AR7 code 2.6.24 aware, for future use
SVN-Revision: 10415
2008-02-08 00:21:08 +00:00
Felix Fietkau 0ae7379775 DG834G Power LED fix
This patch enables the new "default-on" trigger for the AR7 target and uses it to ensure the "power" LED on the DG834G is initialised in the ON state.

Previously this LED would turn off when the driver loaded and stay off until turned back on from userspace. With this patch it remains on throughout the boot process, but can still be controlled from userspace if required.

Signed-off-by: Nick Forbes <nick.forbes@incepta.com>

SVN-Revision: 10349
2008-02-02 01:21:57 +00:00
Matteo Croce c4a4fff789 fixed war.h path
SVN-Revision: 10316
2008-01-29 20:50:50 +00:00
Matteo Croce 72506a987f added missing war.h
SVN-Revision: 10315
2008-01-29 20:47:34 +00:00
Florian Fainelli 7ef96948af Add support for european DSL-502t leds (#2880)
SVN-Revision: 10210
2008-01-16 11:20:03 +00:00
Matteo Croce af9cf4a7a8 gpio: revert it back, wrong and useless
SVN-Revision: 10181
2008-01-13 20:13:08 +00:00
Matteo Croce 5d249c898d GPIO fix
SVN-Revision: 10180
2008-01-12 17:14:15 +00:00
Matteo Croce fae9e82aad cache GPIO addresses to be more efficent
SVN-Revision: 10179
2008-01-12 16:41:28 +00:00
Florian Fainelli b5ccc2e505 Add dg834 leds support from #2817
SVN-Revision: 10105
2008-01-04 02:59:10 +00:00
Florian Fainelli 80c5643c01 Add LEDs definitions for the D-Link DSL-502T (#2746)
Signed-off-by: Oliver <oliver@opencloud.com>

SVN-Revision: 9776
2007-12-16 13:21:31 +00:00
Felix Fietkau f648fc7bc0 Reimplement rx ring with EOQ markers to avoid reset storms
Signed-off-by: oliver@opencloud.com

SVN-Revision: 9665
2007-12-07 00:03:10 +00:00
Felix Fietkau 6111ac6f94 Fix race between device reset and start_xmit
Signed-off-by: oliver@opencloud.com

SVN-Revision: 9664
2007-12-07 00:03:06 +00:00
Felix Fietkau 5dc134c542 Fix VLYNQ device enable for DG834Gv1
This patch allows VLYNQ devices on the DG834Gv1 to be successfully
enabled.

Currently the "__vlynq_enable_device" function attempts to set the VLYNQ
device clock divisor to values from 1 through 8 until a link is
successfully established. On the DG834Gv1 (but not the DG834Gv2),
setting the VLYNQ device clock divisor to 1 (full rate) results in all
further VLYNQ operations failing (including software reset), so the
device is never enabled. This patches changes the function to only
attempt divisors 2 through 8, and hence the device is successfully
enabled.

Signed-off-by: Nick Forbes <nick.forbes@huntsworth.com>

---------

SVN-Revision: 9656
2007-12-04 12:49:54 +00:00
Matteo Croce 788a010131 ar7_gpio: remove unneeded checks and volatile
SVN-Revision: 9419
2007-10-23 15:45:54 +00:00
Eugene Konev 1e5a1f2aa5 fritzbox leds detection (thanks Axel Gembe)
SVN-Revision: 9357
2007-10-19 03:02:49 +00:00
Eugene Konev 53f219cc2b Fix tx error handling (#2552). Make rx ring size actually configurable through ethtool
SVN-Revision: 9333
2007-10-16 12:53:22 +00:00
Eugene Konev 1a3f45fa87 fix the case when NETDEVICES_MULTIQUEUE is not enabled (#2506)
SVN-Revision: 9282
2007-10-13 02:27:19 +00:00
Eugene Konev 4b23cd0283 remove version checks
SVN-Revision: 9253
2007-10-11 01:44:00 +00:00
Eugene Konev 76fdf7d7b0 we are using leds-gpio now
SVN-Revision: 9252
2007-10-11 01:39:24 +00:00
Eugene Konev 982d054c27 locking rework, drop non-napi rx, multiqueue tx, misc cleanups
SVN-Revision: 9251
2007-10-11 01:21:32 +00:00
Matteo Croce 92378640d2 backport ar7_wdt from linux-2.6-watchdog-mm
SVN-Revision: 9182
2007-10-08 01:44:39 +00:00
Eugene Konev 3c4062a45f cleanup vlynq. drop vlynq-pci
SVN-Revision: 9143
2007-10-05 17:54:36 +00:00
Eugene Konev e10dfa5129 more cpmac cleanups (#2452): * don't use net_device::mem* * use net_device::stats * use per-device rx ring size configurable through ethtool * use more ethernet generics (get_stats, set_mac_address, change_mtu)
SVN-Revision: 9098
2007-10-02 14:53:05 +00:00
Eugene Konev 822b666935 add usb controller definition
SVN-Revision: 9087
2007-10-01 12:35:35 +00:00