Re: [PATCH] Add stlc45xx, wi-fi driver for stlc4550/4560
On Sat, 2008-12-06 at 19:32 +0200, Kalle Valo wrote:
+static const u8 default_cal_channels[] = {
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6c, 0x09,
+ 0x00, 0x00, 0xc9, 0xff, 0xd8, 0xff, 0x00, 0x00, 0x00, 0x01, 0x10,
+ 0x01, 0x10, 0x01, 0x10, 0x01, 0x10, 0x01, 0xe0, 0x00, 0xe0, 0x00,
+ 0xe0, 0x00, 0xe0, 0x00, 0xd0, 0x00, 0xd0, 0x00, 0xd0, 0x00, 0xd0,
+ 0x00, 0x54, 0x01, 0xab, 0xf6, 0xc0, 0x42, 0xc0, 0x42, 0xc0, 0x42,
+ 0xc0, 0x42, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00,
+ 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x22, 0x01, 0x37, 0xa9,
+ 0xc0, 0x33, 0xc0, 0x33, 0xc0, 0x33, 0xc0, 0x33, 0x00, 0xbc, 0x00,
+ 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc,
+ 0x00, 0xbc, 0xfb, 0x00, 0xca, 0x79, 0xc0, 0x2b, 0xc0, 0x2b, 0xc0,
+ 0x2b, 0xc0, 0x2b, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4,
+ 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0xd0, 0x00, 0x5d,
+ 0x54, 0xc0, 0x21, 0xc0, 0x21, 0xc0, 0x21, 0xc0, 0x21, 0x00, 0xaa,
+ 0x00, 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0x00,
+ 0xaa, 0x00, 0xaa, 0xa7, 0x00, 0xa9, 0x3d, 0xc0, 0x17, 0xc0, 0x17,
+ 0xc0, 0x17, 0xc0, 0x17, 0x00, 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x00,
+ 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x7a, 0x00,
+ 0x06, 0x2c, 0xc0, 0x0d, 0xc0, 0x0d, 0xc0, 0x0d, 0xc0, 0x0d, 0x00,
+ 0x96, 0x00, 0x96, 0x00, 0x96, 0x00, 0x96, 0x00, 0x96, 0x00, 0x96,
+ 0x00, 0x96, 0x00, 0x96, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x06, 0x80, 0x80, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x71, 0x09, 0x00, 0x00, 0xc9, 0xff, 0xd8,
+ 0xff, 0x00, 0x00, 0x00, 0x01, 0x10, 0x01, 0x10, 0x01, 0x10, 0x01,
+ 0x10, 0x01, 0xf0, 0x00, 0xf0, 0x00, 0xf0, 0x00, 0xf0, 0x00, 0xd0,
+ 0x00, 0xd0, 0x00, 0xd0, 0x00, 0xd0, 0x00, 0x54, 0x01, 0xab, 0xf6,
+ 0xc0, 0x42, 0xc0, 0x42, 0xc0, 0x42, 0xc0, 0x42, 0x00, 0xcb, 0x00,
+ 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb,
+ 0x00, 0xcb, 0x22, 0x01, 0x37, 0xa9, 0xc0, 0x33, 0xc0, 0x33, 0xc0,
+ 0x33, 0xc0, 0x33, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc,
+ 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0xfb, 0x00, 0xca,
+ 0x79, 0xc0, 0x2b, 0xc0, 0x2b, 0xc0, 0x2b, 0xc0, 0x2b, 0x00, 0xb4,
+ 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00,
+ 0xb4, 0x00, 0xb4, 0xd0, 0x00, 0x5d, 0x54, 0xc0, 0x21, 0xc0, 0x21,
+ 0xc0, 0x21, 0xc0, 0x21, 0x00, 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0x00,
+ 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0x00, 0xaa, 0xa7, 0x00,
+ 0xa9, 0x3d, 0xc0, 0x17, 0xc0, 0x17, 0xc0, 0x17, 0xc0, 0x17, 0x00,
+ 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x00, 0xa0, 0x00, 0xa0,
+ 0x00, 0xa0, 0x00, 0xa0, 0x7a, 0x00, 0x06, 0x2c, 0xc0, 0x0d, 0xc0,
+ 0x0d, 0xc0, 0x0d, 0xc0, 0x0d, 0x00, 0x96, 0x00, 0x96, 0x00, 0x96,
+ 0x00, 0x96, 0x00, 0x96, 0x00, 0x96, 0x00, 0x96, 0x00, 0x96, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x06, 0x80, 0x80,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x76,
+ 0x09, 0x00, 0x00, 0xc9, 0xff, 0xd8, 0xff, 0x00, 0x00, 0x00, 0x01,
+ 0x10, 0x01, 0x10, 0x01, 0x10, 0x01, 0x10, 0x01, 0xf0, 0x00, 0xf0,
+ 0x00, 0xf0, 0x00, 0xf0, 0x00, 0xd0, 0x00, 0xd0, 0x00, 0xd0, 0x00,
+ 0xd0, 0x00, 0x54, 0x01, 0xab, 0xf6, 0xc0, 0x42, 0xc0, 0x42, 0xc0,
+ 0x42, 0xc0, 0x42, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb,
+ 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x00, 0xcb, 0x22, 0x01, 0x37,
+ 0xa9, 0xc0, 0x33, 0xc0, 0x33, 0xc0, 0x33, 0xc0, 0x33, 0x00, 0xbc,
+ 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00, 0xbc, 0x00,
+ 0xbc, 0x00, 0xbc, 0xfb, 0x00, 0xca, 0x79, 0xc0, 0x2b, 0xc0, 0x2b,
+ 0xc0, 0x2b, 0xc0, 0x2b, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00,
+ 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0x00, 0xb4, 0xd0, 0x00,
+ 0x5d, 0x54, 0xc0, 0x21, 0xc0, 0x21, 0xc0, 0x21, 0xc0, 0x21, 0x00,
+
Re: [PATCH] Add stlc45xx, wi-fi driver for stlc4550/4560
Johannes Berg [EMAIL PROTECTED] writes:
On Sat, 2008-12-06 at 19:32 +0200, Kalle Valo wrote:
+static const u8 default_cal_channels[] = {
+0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
+0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x6c, 0x09,
[...]
+0x00 };
+
+static const u8 default_cal_rssi[] = {
+0x0a, 0x01, 0x72, 0xfe, 0x1a, 0x00, 0x00, 0x00, 0x0a, 0x01, 0x72,
[...]
+0x00, 0x00, 0x00, 0x00, 0x00 };
Is this data actually usable?
Yes, it is. I tested it yesterday and it works ok.
The static data in cx3110x didn't even parse correctly iirc.
I don't even remember cx3110x having any static calibration data,
wlan-cal provided it always. But then again I haven't looked at
cx3110x for a long time, I might remember wrong.
+static void stlc45xx_tx_edcf(struct stlc45xx *stlc);
+static void stlc45xx_tx_setup(struct stlc45xx *stlc);
+static void stlc45xx_tx_scan(struct stlc45xx *stlc);
+static void stlc45xx_tx_psm(struct stlc45xx *stlc, bool enable);
+static int stlc45xx_tx_nullfunc(struct stlc45xx *stlc, bool powersave);
+static int stlc45xx_tx_pspoll(struct stlc45xx *stlc, bool powersave);
Can you reorder the file?
Sure.
And I think you must be basing this on top of your PS patches, so
could you not put Vivek's in the middle too?
Actually I didn't use my PS patches when I tested this patch. But
true, I should try Vivek's patch also with stlc45xx.
+static ssize_t stlc45xx_sysfs_store_cal_rssi(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf, size_t count)
I still think it would be better to use the firmware framework with a
custom helper for this. At least it should provide a uevent like crda
does so userspace knows automatically when to upload the files after you
reload the module for example, I think?
Yeah, you talked about this already and I liked it. I haven't
implemented it just because a lack of time. I'll add it to the todo
list.
+static ssize_t stlc45xx_sysfs_show_tx_buf(struct device *dev,
+ struct device_attribute *attr,
+ char *buf)
+{
That should be in debugfs, I think?
I think I should even remove it.
+static int stlc45xx_request_firmware(struct stlc45xx *stlc)
+{
+const struct firmware *fw;
+int ret;
+
+/* FIXME: should driver use it's own struct device? */
+ret = request_firmware(fw, 3826.arm, stlc-spi-dev);
own struct device? I don't see why it should? It doesn't have one
anyway, does it?
There is stlc45xx_device:
ret = platform_device_register(stlc45xx_device);
All this device class stuff is not yet fully clear to me, that's why I
added that comment. I guess I should study this more on a boring train
ride.
+static int stlc45xx_upload_firmware(struct stlc45xx *stlc)
+fw_addr = FIRMWARE_ADDRESS;
+fw_len = stlc-fw_len;
+stlc45xx_spi_write(stlc, SPI_ADRS_DMA_DATA, stlc-fw, _fw_len);
+
+/* FIXME: I think this doesn't work if firmware is large,
+ * this loop goes to second round. fw-data is not
+ * increased at all! */
+}
Indeed, the last _spi_write above needs to have something like
stlc-fw + fw_addr - FIRMWARE_ADDRESS
Thanks, I'll fix it at some point.
+/*
+ * FIXME: this gives warning from __ieee80211_rx()
+ *
+ * status.rate_idx = data-rate;
+ */
That works on p54, afaik?
Yes, I haven't looked at this yet.
+edcf-queues[0].aifs = 2;
+edcf-queues[0].pad0 = 1;
+edcf-queues[0].cwmin = 3;
+edcf-queues[0].cwmax = 7;
+edcf-queues[0].txop = 47;
+edcf-queues[1].aifs = 2;
+edcf-queues[1].pad0 = 0;
+edcf-queues[1].cwmin = 7;
+edcf-queues[1].cwmax = 15;
+edcf-queues[1].txop = 94;
+edcf-queues[2].aifs = 3;
+edcf-queues[2].pad0 = 0;
+edcf-queues[2].cwmin = 15;
+edcf-queues[2].cwmax = 1023;
+edcf-queues[2].txop = 0;
+edcf-queues[3].aifs = 7;
+edcf-queues[3].pad0 = 0;
+edcf-queues[3].cwmin = 15;
+edcf-queues[3].cwmax = 1023;
+edcf-queues[3].txop = 0;
Shouldn't that be taking values from mac80211?
Definitely, just has been pending on my todo list.
+setup-bratemask = 0x;
That should also take values from the BSS config.
Will fix.
+scan-flags = LM_SCAN_EXIT;
+scan-bratemask = 0x15f;
or that? not sure
Me neither :)
+static int stlc45xx_op_tx(struct ieee80211_hw *hw, struct sk_buff *skb)
+{
+entry = stlc45xx_txbuffer_alloc(stlc, skb-len);
+if (!entry) {
+/* the queue should be stopped before the firmware buffer
+ * is full, so firmware buffer should always have enough
+ * space */
+if (net_ratelimit())
+stlc45xx_warning(firmware buffer full);
+
Re: [PATCH] Add stlc45xx, wi-fi driver for stlc4550/4560
Kalle Valo [EMAIL PROTECTED] writes: This patch adds a new driver called stlc45xx, which supports wi-fi chipsets stlc4550 and stlc4560 from ST-NXP Wireless. The chipset can be found, for example, from Nokia N800 and N810 products. The driver is implemented based on the firmware interface information published by ST-NXP Wireless here: http://wireless.kernel.org/en/developers/Documentation/specs#STMicroelectronicshardware Currently only SPI interface is supported. Crap, just now I realised to try to compile this with pristine wireless-testing tree: In file included from drivers/net/wireless/stlc45xx.c:24: drivers/net/wireless/stlc45xx.h:27:24: error: mach/board.h: No such file or directory drivers/net/wireless/stlc45xx.c: In function 'stlc45xx_power_off': drivers/net/wireless/stlc45xx.c:958: error: dereferencing pointer to incomplete type drivers/net/wireless/stlc45xx.c:959: error: dereferencing pointer to incomplete type I usually merge linux-omap to wireless-testing and that's why I didn't notice this. I will implement the platform data solution in v2 which should fix this compile error. But I would still love to hear any comments people have about v1. -- Kalle Valo -- To unsubscribe from this list: send the line unsubscribe linux-omap in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
