projects / ipfire-2.x.git / blame
| Commit | Line | Data |
|---|---|---|
| c5c12c9c AF |
1 | From 5bb6f6e1d44aa91323857715dfddb63337f8307b Mon Sep 17 00:00:00 2001 |
| 2 | From: =?UTF-8?q?Bj=C3=B8rn=20Mork?= <bjorn@mork.no> | |
| 3 | Date: Sun, 3 Jul 2016 22:24:50 +0200 | |
| 4 | Subject: cdc_ncm: workaround for EM7455 "silent" data interface | |
| 5 | MIME-Version: 1.0 | |
| 6 | Content-Type: text/plain; charset=UTF-8 | |
| 7 | Content-Transfer-Encoding: 8bit | |
| 8 | ||
| 9 | [ Upstream commit c086e7096170390594c425114d98172bc9aceb8a ] | |
| 10 | ||
| 11 | Several Lenovo users have reported problems with their Sierra | |
| 12 | Wireless EM7455 modem. The driver has loaded successfully and | |
| 13 | the MBIM management channel has appeared to work, including | |
| 14 | establishing a connection to the mobile network. But no frames | |
| 15 | have been received over the data interface. | |
| 16 | ||
| 17 | The problem affects all EM7455 and MC7455, and is assumed to | |
| 18 | affect other modems based on the same Qualcomm chipset and | |
| 19 | baseband firmware. | |
| 20 | ||
| 21 | Testing narrowed the problem down to what seems to be a | |
| 22 | firmware timing bug during initialization. Adding a short sleep | |
| 23 | while probing is sufficient to make the problem disappear. | |
| 24 | Experiments have shown that 1-2 ms is too little to have any | |
| 25 | effect, while 10-20 ms is enough to reliably succeed. | |
| 26 | ||
| 27 | Reported-by: Stefan Armbruster <ml001@armbruster-it.de> | |
| 28 | Reported-by: Ralph Plawetzki <ralph@purejava.org> | |
| 29 | Reported-by: Andreas Fett <andreas.fett@secunet.com> | |
| 30 | Reported-by: Rasmus Lerdorf <rasmus@lerdorf.com> | |
| 31 | Reported-by: Samo Ratnik <samo.ratnik@gmail.com> | |
| 32 | Reported-and-tested-by: Aleksander Morgado <aleksander@aleksander.es> | |
| 33 | Signed-off-by: Bjørn Mork <bjorn@mork.no> | |
| 34 | Signed-off-by: David S. Miller <davem@davemloft.net> | |
| 35 | Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> | |
| 36 | --- | |
| 37 | drivers/net/usb/cdc_ncm.c | 7 +++++++ | |
| 38 | 1 file changed, 7 insertions(+) | |
| 39 | ||
| 40 | diff --git a/drivers/net/usb/cdc_ncm.c b/drivers/net/usb/cdc_ncm.c | |
| 41 | index c663722..584504e 100644 | |
| 42 | --- a/drivers/net/usb/cdc_ncm.c | |
| 43 | +++ b/drivers/net/usb/cdc_ncm.c | |
| 44 | @@ -438,6 +438,13 @@ advance: | |
| 45 | if (cdc_ncm_setup(dev)) | |
| 46 | goto error2; | |
| 47 | ||
| 48 | + /* Some firmwares need a pause here or they will silently fail | |
| 49 | + * to set up the interface properly. This value was decided | |
| 50 | + * empirically on a Sierra Wireless MC7455 running 02.08.02.00 | |
| 51 | + * firmware. | |
| 52 | + */ | |
| 53 | + usleep_range(10000, 20000); | |
| 54 | + | |
| 55 | /* configure data interface */ | |
| 56 | temp = usb_set_interface(dev->udev, iface_no, data_altsetting); | |
| 57 | if (temp) { | |
| 58 | -- | |
| 59 | cgit v0.12 | |
| 60 |