Ever since we started our pc sales and services division we have been trying to get a system with 100% linux supported hardware. One of the pieces which had been the stumbling block till now had been the internal modem. We had tried all the internal modems available in the Trivandrum local market and we had failed, ... well not all. We have found the missing link!
We recently tried out the WeP (Wipro e-Peripherals) Spider 56K V.92 PCI Modem with the Conexant HSFi CX11252-412 0610y1DB chipset (PCI ID 14F1:2F30 - from running lspci from Linux). This modem is supported in Linux and it costs only around Rs. 375. Up till now we were hesitant in offering to install linux on PCs with internal modems because of the dreaded unsupported winmodem issue. No longer so. New PCs are seldom bought with an internal modem but there are quite a lot of people with PCs who still depend on dialup as their primary means of accessing the internet. So this has opened up a whole new market for us and Linux. The Conexant HSFi CX11252 driver installation is pretty straigtforward and is given below
Драйвер Модем Linux Скачать Hsf
It wouldn't surprise me to learn that sometime in the past decade that the driver had been upstreamed into the generic linux kernel, but since I have no concrete information to that effect, I'll lay out the process that I took back then and hope that it remains workable.
See my attached XFree86 configuration (see _linux/XF86Config-4). You see, that I included support for any USB mouse that might be attached to the laptop. Since I use the /dev/input/mice device for the touchpad and the USB mouse, any USB mouse can even be attached and will recognized after X is started (thanks to kernel 2.6.0)!
The script is available at _linux/h2. See h2 -h for usage. This script needs x11vnc and a VNC viewer. Take care to use x11vnc 0.62 or above, because older versions do not convert the mouse pointer position correctly when scaling the display.
Your specific error is because the Raspberry Pi kernel header package doesn't follow the standard naming scheme. The displaylink-debian script needs to be edited to change all instances of linux-headers-$(uname -r) to raspberrypi-kernel-headers so that the Pi's kernel headers are installed properly.
Additional informations MPEG: 1 & 2 (not tested on Linux) Connectors left side (photo) : Battery, 2xPCCard, IEEE1394(FireWire), Headphones, Mic, TV-OUT back (photo) : Power, USB (#1), Ethernet (RJ45), VGA, parallel, serial, USB (#2), modem (RJ11) right side (photo) : CD-ROM, FDD (can be replaced with second battery) bottom : port replicator (photo) Other related resources Linux-Sony Home Page Linux-Sony mailing list archive Linux on Laptops home (and Sony there) NEW Linux Kernel for laptops -- laptopkernel is a patchset for the linux kernel containing several useful patches for laptop-users. It contains acpi, software suspend, supermount and some hardware compatibility patches. NEW Battery FAQ Official specifications sheet (.pdf 196KB) NEW Intel(R) 810 and 815 Framebuffer Driver project page (I didn't tested) NEW! Very good detailed description of FX220 by David Levitan Linux on a Sony Vaio FX209K by Peter Koellner Sony VAIO PCG-FX200 by Ben Crowder Other models Some other Vaios has similar problems as FX240. Some patches/solutions can be sucessfully used on them. This is list of informations I got from other people about their laptops. Model What can be fixed Who wrote me about it(and maybe can help you) Comments, etc. FX101 USB Matt Willsher Model also known as PCG-955C FX140 USB Kevin FX200 USB Satyajit Behari FX220 USB Chris Woodruff Dennis McGuire FX250 USB Daniel de Young FX270 USB Andrew Zavjalov FX290 USB Victor Khimenko Read Victor's BIOS trail! FX340 USB, ethernet Leonardo Andersen Lopes FX370 USB Mauricio Di Fulvio Richard G. Froese Carl Rehbein FX390 USB F. Oliver Gathmann FX403 USB Volker Christian Behr FX403 has 2 USB hubs. Here is patch that will address flaw in both. Thanks to Volker. FX950 USB Gwyn Morfey FX503 USB Tony Geraghty R505JE USB Dylan Oliver R505JS USB Kirk VanOpdorp R505JL USB Girish Pahlya There were no error messages on USB in log files -- see Girish's mail R600MX USB Phil Anderson Acer Travelmate723txv USB Gino Pelliccia This patch works for Gino on his Acer, after little modifications. Ask him for details Fujitsu-SiemensC1010 USB Robert Klein This patch works for Robert on his F-S, after modifications. Ask him for details. (USB stopped working after upgrading BIOS) Note: this is information about models/hardware that somebody had tested. It doesn't mean that these solution cannot work with other models or with other hardware with these models. If you have other useful information about your model, send it to me . I'll add it here. Thanks! Thanks for all people for sending their expiriences, comments & kind words!!! UPDATED I removed information about working modems above. I think the official drivers should make work all Conexants. CPU SpeedStep NEW Way of accessing SpeedStep has been once again in recent kernels (tested on 2.4.21-rc1-ac3 and later). Basing on Dominik Brodowski's mail there are 3 ways of accessing SpeedStep. Two of them are supported by current kernels (2.4 with ACPI or -ac patch or 2.5) ACPI "P-States". At /proc/acpi/processor/CPU0 there is performance file, and using this it should be possible to select SpeedStep mode. Unfortunately, currently, my FX240 with 2.4.21-rc6-ac1 kernel doesn't support this method. /proc/acpi/processor/CPU0/performance contains . CPUfreq -- legacy speedstep driver. After enabling the driver in kernel (Processor type and features -> CPU Frequency scaling -> ...) it's possible to switch frequency using /proc/cpufreq interface. Description of this method can be found in cpufreq info in kernel's Documentation/cpufreq. Note: you have to add speedstep_coppermine=1 to kernel boot parameters (or module parameters if cpufreq support compiled as module). Marc's speedstep driver. Dominik: ...undocumented, uncertain, in my opinion highly dangerous interface... So I don't continue. ACPI throttling. (my thoughts) There is also /proc/acpi/processor/CPU0/throttling interface. It works on FX240. Throttling is lowering the cpu clock cycles per second and, if I understand well, the voltage level. I don't know if there is direct connection between throttling and SpeedStep. Probably... See my (2.4.21-rc6-ac1 kernel): /proc/acpi/processor/CPU0/performance /proc/acpi/processor/CPU0/throttling /proc/cpufreq Special SpeedStep daemon can be used to automatically switch to high (performace) or low (powersave) level, as cpu load changes. Some projects below. This is old (outdated) info for earlier kernels... I just upgraded the kernel to linux-2.4.19-pre9-ac3 and SpeedStep works! Only thing you have to do is edit file linux/arch/i386/kernel/speedstep.c and delete line check = 1; Now you can: check current speed: $ cat /proc/sys/cpu/0/speed switch to hi speed (800MHz in case of FX240): $ cat /proc/sys/cpu/0/max-speed >/proc/sys/cpu/0/speed switch to lo speed (650MHz in case of FX240): $ cat /proc/sys/cpu/0/min-speed >/proc/sys/cpu/0/speed That's it! Some additional Linux/SpeedStep links: Linux CPUFreq project (by Dominik Brodowski, seems to be empty (???)) NEW autospeedstep daemon (AFAIK needs ACPI "P-States" mode) speedstep daemon (doesn't work for me...) cpufreq list archives VGA By default 80x25 textmode window is "small". It just doesn't fill area of the screen. It look like this: Black is 14" screen area, and white is area used to display 80x25 text. You can try to change textmode (with VGATextMode), but I didn't get any satisfactory results. There is no option in CMOS for this Similar problem is with graphics mode. Only supported resolution is 1024x768. If you pick up lower, you got the same resolution, but... in smaller window. You may want to see Intel(R) 810 and 815 Framebuffer Driver project page. I didn't tested yet. Antonino Daplas/Ken Mano utility i810switch to switch between LCD & VGA. Intel's technical specification for i810 chipset. (PDF, 2.9MB) Modem Modem works! You have to download HSF modem drivers. The official page of Linux drivers for Conexant is What to do? -- Install the hsflinmodem package, and run hsfconfig. Sound I'm not quite sure if I was doing everything alright, but i810 audio driver supported only 48kHz sampling rate. When I have switched to ALSA driver - everything is ok. (I use 0.9.0beta9 version) One problem - voice from internal speakers has very poor quality. But what we can expect from speakers that doesn't have more than 1.5 cm diameter. CD-ROM cdrom works I didn't test the DVD player, but Gordon Runkle has reported that it works in his FX220. This is pretty the same machine. He is using Ogle player to watch DVD. You can change region for DVD only 5 times (probably from Windows only?). To make cdrecord work - you must use ide-scsi driver To make cdrdao (v1.1.5) work - use ide-scsi and add following line to: /usr/local/share/cdrdao/drivers WMATSHITAUJDA710generic-mmcOPT_MMC_USE_PQOPT_MMC_PQ_BCDOPT_MMC_READ_ISRC Ethernet This device works great at 100Mbps (100Base-T), but... not at 10Mbps (10Base-T). After period of time (at 10Mbps), driver says: eepro100: wait_for_cmd_done timeout and no longer transmit anything. You can unload module eepro100.o and load it again, and then it works again for some time. Update I just found, that EEPROM of this card has set a sleep mode enabled. And this is potential source of problems on heavy load. I have switched this now, and I'll write what are the results, when I have chance to test it on 10Mbps network. If you want to check this condition - download eepro100-diag.c program and run: jajo:# ./eepro100-diag -e -e -f If result contains this text, maybe it is yours problem also... Sleep mode is enabled. This is not recommended. Under high load the card may not respond to PCI requests, and thus cause a master abort. You can change the value of this setting by the command: jajo:# ./eepro100-diag -G -w -w -w -f Warning: Do everything on your own risk. Solution: There is other module called e100.o It's part of 2.4 kernel. More info on Intel's site. It can be downloaded directly from Intel. USB IRQ routing tables for USB devices seems to be broken. Wrong IRQ is assigned to USB (#10 instead of #9), and in effect USB is dead. You can use my patch (at your risk). It works good for me. This is the receipt how to apply it: untar kernel source jajo:/usr/src/linux$ cd arch/i386/kernel jajo:/usr/src/linux/arch/i386/kernel$ patch pci-irq.c
2ff7e9595c
Comments