[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <910F9D9E13B84F4C8FA771DC9BDE99F326D4B940@szxeml546-mbs.china.huawei.com>
Date: Tue, 17 Jul 2012 03:53:19 +0000
From: "Fangxiaozhi (Franko)" <fangxiaozhi@...wei.com>
To: Bjørn Mork <bjorn@...k.no>
CC: "linux-usb@...r.kernel.org" <linux-usb@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Xueguiying (Zihan)" <zihan.xue@...wei.com>,
"greg@...ah.com" <greg@...ah.com>,
"smurf@...rf.noris.de" <smurf@...rf.noris.de>,
"Linlei (Lei Lin)" <lei.lin@...wei.com>,
"Liuqifeng (Qifeng)" <qifeng.liuqifeng@...wei.com>,
Wangyeqi <wangyeqi@...wei.com>,
Wangchangliang <wangchangliang@...wei.com>,
Zhuxinbo <zhuxinbo@...wei.com>,
Gustavo Padovan <gustavo@...ovan.org>
Subject: 答复: [PATCH 7/13] USB:support the new interfaces of Huawei Data Card devices in option driver
Dear bjorn:
"Fangxiaozhi (Franko)" <fangxiaozhi@...wei.com> writes:
> From: fangxiaozhi <huananhu@...wei.com>
> 1. This patch is based on the kernel of 3.5-rc6
> 2. In this patch, we add new micro for matching the series USB devices with vendor ID and interface information.
> 3. In this patch, we add new declarations into option.c to support the new interfaces of Huawei Data Card devices.
> Signed-off-by: fangxiaozhi <huananhu@...wei.com>
> -----------------------
> --- ../test/linux-3.5-rc6/include/linux/usb.h 2012-07-08 08:23:56.000000000 +0800
> +++ include/linux/usb.h 2012-07-13 17:45:59.000000000 +0800
> @@ -828,6 +828,27 @@ static inline int usb_make_path(struct u
> .bInterfaceClass = (cl), \
> .bInterfaceSubClass = (sc), \
> .bInterfaceProtocol = (pr)
> +/**
> + * * USB_VENDOR_AND_INTERFACE_INFO - describe a specific usb device with a class of usb interfaces, but independent of product ID
This chunk seems like a copy of the patch Gustavo Padovan just
submitted? Should really be listed as a precondition instead of
included here, should it not?
-----In your opinions, it is better to declare this defining in the option.c file, but not usb.h file, right?
> --- ../test/linux-3.5-rc6/drivers/usb/serial/option.c 2012-07-13 14:22:52.000000000 +0800
> +++ drivers/usb/serial/option.c 2012-07-13 17:38:38.000000000 +0800
> @@ -572,6 +572,115 @@ static const struct option_blacklist_inf
> };
>
> static const struct usb_device_id option_ids[] = {
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x01) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x02) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x03) },
I guess this means that the device specific entries matching this could
and should be removed, does it not? All these seem redundant with your
patch:
------The new matching rule is independent of the special product ID, so it can support a series products of Huawei Data Card.
-----The following matching rule is only for the specific product, and it is covered by the new matching rule, so I think that we can remove the following matching sentences.
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K4605, 0xff, 0x01, 0x31) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K4605, 0xff, 0x01, 0x32) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K5005, 0xff, 0x01, 0x31) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K5005, 0xff, 0x01, 0x32) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K5005, 0xff, 0x01, 0x33) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K3770, 0xff, 0x02, 0x31) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K3770, 0xff, 0x02, 0x32) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K3771, 0xff, 0x02, 0x31) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K3771, 0xff, 0x02, 0x32) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K4510, 0xff, 0x01, 0x31) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K4510, 0xff, 0x01, 0x32) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K4511, 0xff, 0x01, 0x31) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_K4511, 0xff, 0x01, 0x32) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x01, 0x01) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x01, 0x02) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x01, 0x03) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x01, 0x10) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x01, 0x12) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x01, 0x13) },
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x02, 0x01) }, /* E398 3G Modem */
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x02, 0x02) }, /* E398 3G PC UI Interface */
{ USB_DEVICE_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, HUAWEI_PRODUCT_E353, 0xff, 0x02, 0x03) }, /* E398 3G Application Interface */
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x04) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x05) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x06) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x0A) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x0B) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x0D) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x0E) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x0F) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x10) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x01, 0x12) },
I see that the holes I know of are there, like protocol 0x08, 0x09 and
0x11. You wouldn't happen to have a complete list of those as well? I
am interested in the ones which can be expected to work with the
qmi_wwan driver. Currently we match
0xff, 0x01, 0x09 (along with a 0xff, 0x01, 0x08 data interface)
0xff, 0x01, 0x11
0xff, 0x01, 0x39 (along with a 0xff, 0x01, 0x38 data interface)
but I suspect there are more variants which could be handled by that
driver.
------Sorry, protocol 0x08, 0x09 and 0x11 is defined as the ECM port, they are supported by other kernel driver, but not option.c driver.
------So, we don't declare them in option.c.
[..]
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x02, 0x79) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x02, 0x7A) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x02, 0x7B) },
> + { USB_VENDOR_AND_INTERFACE_INFO(HUAWEI_VENDOR_ID, 0xff, 0x02, 0x7C) },
> +
> { USB_DEVICE(OPTION_VENDOR_ID, OPTION_PRODUCT_COLT) },
> { USB_DEVICE(OPTION_VENDOR_ID, OPTION_PRODUCT_RICOLA) },
> { USB_DEVICE(OPTION_VENDOR_ID, OPTION_PRODUCT_RICOLA_LIGHT) },
And should the new entries be sorted with the other Huawei entries?
Putting them in front of the Option entries looks a bit strange
-------OK, we will improve this, and then update the patch.
Best Regards,
Franko Fang
Powered by blists - more mailing lists