rtlwifi: Fix EFUSE_ANA8M map value

The map entry EFUSE_ANA8M is supposed to be a bit mask of the SYS_CLK register (see efuse.c) It doesn't make sense to assign the enumeration value EFUSE_ANA8M. Assign the bitmask ANA8M instead. rtl8192se does not have ANA8M, so use 0 as bitmask.

---

This is RFC, because I don't really know the device.
The rtl8192se part of the patch is just a guess, because this driver's reg.h doesn't have an ANA8M define.

Index: wireless-drivers-next/drivers/net/wireless/rtlwifi/rtl8192ce/sw.c
===================================================================
@@ -282,7 +282,7 @@ static struct rtl_hal_cfg rtl92ce_hal_cf
 	.maps[EFUSE_PWC_EV12V] = PWC_EV12V,
 	.maps[EFUSE_FEN_ELDR] = FEN_ELDR,
 	.maps[EFUSE_LOADER_CLK_EN] = LOADER_CLK_EN,
-	.maps[EFUSE_ANA8M] = EFUSE_ANA8M,
+	.maps[EFUSE_ANA8M] = ANA8M,
 	.maps[EFUSE_HWSET_MAX_SIZE] = HWSET_MAX_SIZE,
 	.maps[EFUSE_MAX_SECTION_MAP] = EFUSE_MAX_SECTION,
 	.maps[EFUSE_REAL_CONTENT_SIZE] = EFUSE_REAL_CONTENT_LEN,

Index: wireless-drivers-next/drivers/net/wireless/rtlwifi/rtl8192cu/sw.c
===================================================================
@@ -206,7 +206,7 @@ static struct rtl_hal_cfg rtl92cu_hal_cf
 	.maps[EFUSE_PWC_EV12V] = PWC_EV12V,
 	.maps[EFUSE_FEN_ELDR] = FEN_ELDR,
 	.maps[EFUSE_LOADER_CLK_EN] = LOADER_CLK_EN,
-	.maps[EFUSE_ANA8M] = EFUSE_ANA8M,
+	.maps[EFUSE_ANA8M] = ANA8M,
 	.maps[EFUSE_HWSET_MAX_SIZE] = HWSET_MAX_SIZE,
 	.maps[EFUSE_MAX_SECTION_MAP] = EFUSE_MAX_SECTION,
 	.maps[EFUSE_REAL_CONTENT_SIZE] = EFUSE_REAL_CONTENT_LEN,

Index: wireless-drivers-next/drivers/net/wireless/rtlwifi/rtl8192se/sw.c
===================================================================
@@ -330,7 +330,7 @@ static struct rtl_hal_cfg rtl92se_hal_cf
 	.maps[EFUSE_PWC_EV12V] = 0, /* nouse for 8192se */
 	.maps[EFUSE_FEN_ELDR] = 0, /* nouse for 8192se */
 	.maps[EFUSE_LOADER_CLK_EN] = 0,/* nouse for 8192se */
-	.maps[EFUSE_ANA8M] = EFUSE_ANA8M,
+	.maps[EFUSE_ANA8M] = 0, /* nouse for 8192se */
 	.maps[EFUSE_HWSET_MAX_SIZE] = HWSET_MAX_SIZE_92S,
 	.maps[EFUSE_MAX_SECTION_MAP] = EFUSE_MAX_SECTION,
 	.maps[EFUSE_REAL_CONTENT_SIZE] = EFUSE_REAL_CONTENT_LEN,

I think the patch is correct. Certainly loading an array entry with the offset of the array makes little sense. This change makes the older drivers (rtl8192c{e,u} and rtl8192se match the newer ones.

Once I test the patch, I will send it on to Kalle with my S-o-b.

Thanks,

Larry

Thanks. I did this, because I have an rtl8192cu device that does not work (even with this patch). It doesn't return any scan results. I'll have to debug this further. Any suggestions?

Michael

What does lsusb say about your device?

The latest Realtek driver that I have can be gotten from http://github.com/lwfinger/rtl8192cu.git. That one works pretty well with device 7392:7811, which is an Edimax EW-7811Un, but does not work with device 2357:0100, which is a TP-Link WN8200ND.

The GitHub repo needs a bit of work as it tends to spam the logs a lot, but it might work for your device.

Larry

It is exactly this device.
Does yours also fail with vanilla kernel?

Michael

No, it is able to scan and connect with the vanilla kernel, but it is not terribly reliable with about 25% ping loss. I just got a report of a regression. Kernel 4.0 fails whereas 3.19 works. I have been seeing a lot of patches from various sources. I thought they were OK, but I will need to bisect.

Some days I am sorry that I ever heard of Linux. :)

Larry

I tried with 3.16 and 4.0 based Debian kernels. Neither worked.
I use the firmware provided by Debian's firmware-realtek package.
I have two of these devices, so I think the hardware should be ok, unless they are broken both.

When I first tried to use it, I forgot to install firmware. It tried to load it, then some 'alternative firmware' and then froze the system partially. I guess there's some unreleased lock or such in an error path somewhere. I didn't check this, because it didn't seem too terribly important for now.

I will also try your driver from your github repo soon and see if that makes it work.

Retired, but still out of time. :)

Michael

Oh well. To put it mildly, this driver is awful.
It compiles with frightening array overrun warnings in rtl8192c_rf6052.c, on load spews a ton of messages and then replies to a scan request with:

root@milhouse:~# ip link set up dev wlan4
root@milhouse:~# iwlist wlan4 scan
wlan4     Interface doesn't support scanning.

root@milhouse:~# iw dev wlan4 scan
command failed: No such device (-19)

Stand clear, here come the messages:

[ 3973.868312] rtl8192cu driver version=v4.0.2_9000.20130911
[ 3973.868453] usbcore: registered new interface driver rtl8192cu
[ 4028.492113] usb 1-2: new high-speed USB device number 6 using ehci-pci
[ 4028.625897] usb 1-2: New USB device found, idVendor=7392, idProduct=7811
[ 4028.625912] usb 1-2: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 4028.625921] usb 1-2: Product: 802.11n WLAN Adapter
[ 4028.625930] usb 1-2: Manufacturer: Realtek
[ 4028.625938] usb 1-2: SerialNumber: xxx
[ 4028.626959] 
               usb_endpoint_descriptor(0):
[ 4028.626976] bLength=7
[ 4028.626983] bDescriptorType=5
[ 4028.626991] bEndpointAddress=81
[ 4028.626998] wMaxPacketSize=200
[ 4028.627004] bInterval=0
[ 4028.627011] RT_usb_endpoint_is_bulk_in = 1
[ 4028.627017] 
               usb_endpoint_descriptor(1):
[ 4028.627026] bLength=7
[ 4028.627032] bDescriptorType=5
[ 4028.627038] bEndpointAddress=2
[ 4028.627044] wMaxPacketSize=200
[ 4028.627050] bInterval=0
[ 4028.627056] RT_usb_endpoint_is_bulk_out = 2
[ 4028.627063] 
               usb_endpoint_descriptor(2):
[ 4028.627073] bLength=7
[ 4028.627080] bDescriptorType=5
[ 4028.627087] bEndpointAddress=3
[ 4028.627095] wMaxPacketSize=200
[ 4028.627102] bInterval=0
[ 4028.627108] RT_usb_endpoint_is_bulk_out = 3
[ 4028.627115] 
               usb_endpoint_descriptor(3):
[ 4028.627126] bLength=7
[ 4028.627133] bDescriptorType=5
[ 4028.627141] bEndpointAddress=84
[ 4028.627147] wMaxPacketSize=40
[ 4028.627154] bInterval=1
[ 4028.627162] RT_usb_endpoint_is_int_in = 4, Interval = 1
[ 4028.627171] nr_endpoint=4, in_num=2, out_num=2

[ 4028.627182] USB_SPEED_HIGH
[ 4028.627240] CHIP TYPE: RTL8188C_8192C
[ 4028.627274] register rtw_netdev_ops to netdev_ops
[ 4028.627415] Chip Version ID: VERSION_NORMAL_TSMC_CHIP_88C.
[ 4028.627428] RF_Type is 3!!
[ 4028.627768] EEPROM type is E-FUSE
[ 4028.627780] ====> ReadAdapterInfo8192C
[ 4028.627898] Boot from EFUSE, Autoload OK !
[ 4028.778757] EEPROMVID = 0x7392
[ 4028.778765] EEPROMPID = 0x7811
[ 4028.778769] EEPROMCustomerID : 0x00
[ 4028.778773] EEPROMSubCustomerID: 0x00
[ 4028.778777] RT_CustomerID: 0x00
[ 4028.778784] _ReadMACAddress MAC Address from EFUSE = xxx
[ 4028.778790] EEPROMRegulatory = 0x0
[ 4028.778794] _ReadBoardType(0)
[ 4028.778799] BT Coexistance = disable
[ 4028.778804] mlmepriv.ChannelPlan = 0x02
[ 4028.778809] _ReadPSSetting...bHWPwrPindetect(0)-bHWPowerdown(0) ,bSupportRemoteWakeup(0)
[ 4028.778813] ### PS params=>  power_mgnt(1),usbss_enable(0) ###
[ 4028.778817] ### AntDivCfg(0)
[ 4028.778821] readAdapterInfo_8192CU(): REPLACEMENT = 1
[ 4028.778826] <==== ReadAdapterInfo8192C in 152 ms
[ 4028.779451] rtw_macaddr_cfg MAC Address  = xxx
[ 4028.779461] bDriverStopped:1, bSurpriseRemoved:0, bup:0, hw_init_completed:0
[ 4028.780924] _rtw_drv_register_netdev, MAC Address (if1) = xxx
[ 4029.844223] rtl8192cu 1-2:1.0 wlan4: renamed from wlan1
[ 4029.872791] systemd-udevd[30617]: renamed network interface wlan1 to wlan4
[ 4032.495346] +871x_drv - drv_open, bup=0
[ 4032.499722]  ===> FirmwareDownload91C() fw:Rtl819XFwImageArray_TSMC
[ 4032.499737] FirmwareDownload92C accquire FW from embedded image
[ 4032.499746] fw_ver=v88, fw_subver=2, sig=0x88c0
[ 4032.532832] fw download ok!
[ 4032.532841] Set RF Chip ID to RF_6052 and RF type to 1T1R.
[ 4032.906818] IQK:Start!!!
[ 4032.918190] Path A IQK Success!!
[ 4032.926443] Path A IQK Success!!
[ 4032.931584] IQK: final_candidate is 0
[ 4032.931595] IQK: RegE94=106 RegE9C=13 RegEA4=fd RegEAC=3fd RegEB4=0 RegEBC=0 RegEC4=0 RegECC=0
                Path A IQ Calibration Success !
[ 4033.043931] pdmpriv->TxPowerTrackControl = 1
[ 4033.048684] rtl8192cu_hal_init in 556ms
[ 4033.061576] MAC Address = xxx
[ 4033.061789] -871x_drv - drv_open, bup=1
[ 4033.062150] IPv6: ADDRCONF(NETDEV_UP): wlan4: link is not ready
[ 4035.068534] ==>rtw_ps_processor .fw_state(0)
[ 4035.068546] ==>ips_enter cnts:1
[ 4035.068552] ===> rtw_ips_pwr_down...................
[ 4035.068771] ====> rtw_ips_dev_unload...
[ 4035.092200] usb_read_port_cancel
[ 4035.092508] usb_read_port_complete()-1284: RX Warning! bDriverStopped(0) OR bSurpriseRemoved(0) bReadPortCancel(1)
[ 4035.092761] usb_read_port_complete()-1284: RX Warning! bDriverStopped(0) OR bSurpriseRemoved(0) bReadPortCancel(1)
[ 4035.092998] usb_read_port_complete()-1284: RX Warning! bDriverStopped(0) OR bSurpriseRemoved(0) bReadPortCancel(1)
[ 4035.093247] usb_read_port_complete()-1284: RX Warning! bDriverStopped(0) OR bSurpriseRemoved(0) bReadPortCancel(1)
[ 4035.093263] usb_write_port_cancel 
[ 4035.093315] ==> rtl8192cu_hal_deinit 
[ 4035.093319] bkeepfwalive(0)
[ 4035.093323] card disble without HWSM...........
[ 4035.098401] <=== rtw_ips_pwr_down..................... in 28ms

If you want to have this device, I'll ship it to you for free.

Michael

I know about the log spamming. All the drivers for Realtek USB devices are released with full debugging mode enabled. I did not have time to fix that before I posted the driver on GitHub.

Sure, send me the device. That will give me a chance to try it as well as the other devices I have.

Larry