My task at the moment is to port a driver for some 16550-compatible chip from QNX to Linux. The chip provides several UARTs, each one seen as a standard 16550 serial port, albeit with some extensions.
Now, in QNX, the whole device driver is packed into a standalone executable, that acts both as a driver and as an initial configurator for the provided UARTs (baud rates, loopback modes etc.) That is just natural in QNX because there device drivers run in user space and are little more than standard executables.
On Linux, OTOH, the driver is now implemented as a kernel module, loadable at will. More, that module is provided by the producer, so I would not want to modify or patch it too much.
For me, the remaining task is to provide some mechanism for setting up those UARTs' parameters. They are seen as /dev/ttyPREFIXX devices. I intend to do that through a standard C-programmed executable calling standard termios (ie tcsetattr() or ioctls) on the serial ports of interest.
Which leads me to the question: is my approach right? And, if yes, then how to achieve a persistent configuration? As I perceive the fact (from this example: http://www.easysw.com/~mike/serial/serial.html), the termios functions act on OPEN devices. In short: they open a device, they set up the parameters, they read or write, then close the port. After closing the port, is the configuration (baud rate etc.) lost? I hope it is not, because it is stored, already, into the hardware.
Can somebodey confirm to me that the configuration is persistent? And, if not, how to achieve that persistence, for the future applications that would open again that port and will expect it with some pre-established parameters? If not, should I modify the module kernel to accept some parameters and, then, do the configuration at the load time?
The approach that I intend for now is to write that C executable that opens the ports, sets up their configuration, then close the ports. I hope the latter applications will see the ports with the desired configuration.
Thank you.
You might want to have a look at stty and setserial. The venerable Serial-HOWTO (wow, when was the last time I actually recommended a HOWTO to anyone?) is probably also a good starting point.
Well, I found the answer here: http://www.gnu.org/software/libc/manual/html_node/Mode-Functions.html#Mode-Functions
Quote: "Although tcgetattr and tcsetattr specify the terminal device with a file descriptor, the attributes are those of the terminal device itself and not of the file descriptor. This means that the effects of changing terminal attributes are persistent; if another process opens the terminal file later on, it will see the changed attributes even though it doesn't have anything to do with the open file descriptor you originally specified in changing the attributes."
This clears the issue.
Related
I am comparing a mainline Linux kernel source with a modified copy of the same source that has many drivers added. A little background: That modified source is an Android kernel source, it contains many drivers added by the vendor, SoC manufacturer, Google etc.
I am trying to identify all drivers added in the modified source that are reachable from userspace via any syscalls. I'm looking for some systematic or ideally automatic way to find all these to avoid the manual work.
For example, char device drivers are of interest, since I could perform some openat, read, write, ioctl and close syscalls on them if there is a corresponding device file. To find new character device drivers, I could first find all new files in the source tree and then grep them for struct file_operations. But besides char drivers, what else is there that I need to look for?
I know that the syscalls mentioned above do some kind of "forwarding" to the respective device driver associated with the file. But are there other syscalls that do this kind of forwarding? I think I would have to focus on all these syscalls, right?
Is there something I can grep for in source files that indicates that syscalls can lead there? How should I go about this to find all these drivers?
Update (narrowing down):
I am targeting specific devices (e.g. Huawei P20 Lite), so I know relevant architecture and hardware. But for the sake of this question, we can just assume that hardware for whatever driver is present. It doesn't really matter in my case if I invoked a driver and it reported back that no corresponding hardware is present, as long as I can invoke the driver.
I only look for the drivers directly reachable via syscalls. By directly reachable I mean drivers designed to have some syscall interface with userspace. Yes, syscalls not aimed at a certain driver may still indirectly trigger code in that driver, but these indirect effects can be neglected.
Maybe some background on my objective clarifies: I want to fuzz-test the found drivers using Syzkaller. For this, I would create descriptions of the syscalls usable to fuzz each driver that Syzkaller parses.
I'm pretty sure there is no way to do this programmatically. Any attempt to do so would hit up against a couple of problems:
The drivers that are called in a given case depend on the hardware. For example, on my laptop, the iwlwifi driver will be reachable via network syscalls, but on a server that driver won't be used.
Virtually any code loaded into the kernel is reachable from some syscall if the hardware is present. Drivers interact with hardware, which in turn either interacts with users, external devices, or networks, and all of these operations are reachable by syscalls. People don't write drivers that don't do anything.
Even drivers that aren't directly reachable by a system call can affect execution. For example, a driver for a true RNG would be able to affect execution by changing the behavior of the system PRNG, even if it weren't accessible by /dev/hwrng.
So for a generic kernel that can run on any hardware of a given architecture, it's going to be pretty hard to exclude any driver from consideration. If your hope is to trace the execution of the code by some programmatic means without actually executing it, then you're going to need to solve the halting problem.
Sorry for the bad news.
The goal is to "lock" access to a serial device or other Linux device, to ensure exclusive access to the device while it's in-use. This prevents, for example, two programs both opening the same serial device and "competing" to read bytes from the device.
The advice has been to use SYSV-style UUCP device lock files such as /var/lock/LCK..ttyS1. That is what is recommended by Linux Serial HOTWO: Locking Out Others. It is also documented in the Filesystem Heirarchy Standard. It is implemented by serial terminal programs such as gtkterm, picocom. There are libraries such as liblockdev and liblockfile to support this (although the implementation details differ between these two libraries).
However, I have found Debian bug #734086, which says on Linux, SYSV-style UUCP device locks are deprecated, and flock() advisory locks should be used instead.
However, I can't find a reliable document source to describe deprecation of these SYSV-style UUCP device locks, and recommendation of flock(), other than that Debian bug itself.
I've also found ioctl(fd, TIOCEXCL) which is used by the screen utility to lock a terminal.
Which is the modern "best practice" for locking serial ports and other devices in Linux? Where can we find up-to-date documentation describing this?
As far as I can tell, using flock(fd, LOCK_EX | LOCK_NB) locking of serial ports or other devices is probably the best the way to go in Linux, following Debian's lead in Debian bug #734086. Note that the original advocate of this Debian change, Roger Leigh, has moved away from Debian and Linux and onto FreeBSD in 2014/2015 (see his comments here). But Debian seems to be sticking with the flock() method, so that's worth something.
However given how poorly this change has been communicated to the broader Linux community at this point, it could be good to support the older SYSV-style UUCP device lock files (/var/lock/LCK..ttyS1) as a compile-time option, for use in systems still using the older lock method.
E.g. picocom has now changed to using the flock() method, with a compile-time optional selection of SYSV-style UUCP device lock files.
Another answer on StackOverflow describes two methods:
ioctl(fd, TIOCEXCL)
the flock(fd, LOCK_EX | LOCK_NB) method
It says "An application can choose to do one, or both", further explaining:
The reason to use both, is to detect if another process has already opened the device without putting it into exclusive mode, but has hopefully set the advisory lock. In that case, the open() and ioctl() both succeed, but the flock() fails.
So, it is worth implementing ioctl(fd, TIOCEXCL) additionally.
I need to write a program in linux (debian, to be exact) that disables a USB device if it doesn't pass specific filters. For example, the program might be set to disallow webcams, usb sticks and keyboards, enable mice and printer (through usb). The filters may change in runtime. For example, the program might receive a message to enable usb sticks and it should respond without rebooting the system. The program is written in python but embedding c code (or others) is also acceptable.
What I have tried
I have tried many methods but some of them aren't about programming.
First, I tried to mess with udev. I can monitor the activities when a device is plugged and write filters. There used to be an option "ignore_device" that ignored the filtered devices. For example, to ignore all the devices that are a member of usb subsystem, I would write this as a udev rule:
ACTION=="add", SUBSYSTEM=="usb", OPTIONS+="ignore_device"
But this option is released in this release of udev. What I get so far is that udev can currently be used mainly for monitoring. Sure, I can write additional rules for the rule above that runs a script, but I have to do the disabling elsewhere.
Second, I tried ioctl to send a DISCONNECT signal to device handle. I'm testing this with a USB mouse. This is the python code for that: (I have also tried this in C, nothing changed)
import fcntl
import sys
USBDEVFS_RESET = ord('U') << (4*2) | 20
USBDEVFS_DISCONNECT = ord('U') << (4*2) | 22
raw_name = "/dev/bus/usb/{:03d}/{:03d}"
filename = raw_name.format(1,2)
fd = open(filename, "wb")
fcntl.ioctl(fd, USBDEVFS_DISCONNECT, 0)
Here, if I would send USBDEVFS_RESET, it works, the mouse input is ignored for a second or two. But disconnect signal raises an error:
IOError: [Errno 25] Inappropriate ioctl for device
What I get from here is, I cannot send disconnect signal to a mouse. Maybe a usb stick or printer or some other devices would work, I haven't tried. I want to develop the program as generic as possible so as to prevent writing additional device-specific code, so this approach seems useless for me. And another point here is that when I manually disconnect/connect my mouse, I see events in udev monitor. But when I send reset signal, no event is sent.
The udev monitor says that the mouse was mounted to this path: /sys/bus/usb/devices/1-3 (which is a symlink for /sys/devices/pci0000:00/0000:00:14.0/usb1/1-3). Some documents told me that this folder contains the device's settings and setting /power/level to "off" or "suspended" would turn the device power off. But I cannot manipulate any files in /power. Come to think about it, it might not be a good idea after all.
The Question
So, the question is, what is the best way to achieve such task? I have an idea but I'm not sure whether it will work and even if it does, it might be overkill. My idea is writing a "wrapper driver" that identifies itself to linux kernel as driver for all usb devices. The "wrapper driver" reads the device information and if the device is good to go, it acts as a wrapper for real drivers in the kernel, calling their functions. If not, the "wrapper driver" just ignores the device.
I'm not sure it can even be done, I'm not experienced in kernel or driver programming.
Another way is, -somehow- getting the handle of the device programatically and telling it to power off (or making it busy forever, whatever works). I have also done some little research but couldn't find a proper-easy way. They say all devices are considered as "files" but I cannot reach those files at all.
Note:
The question is mostly about linux internals but it also involves kernel programming. I read a lot about usb manipulation/monitoring programs, I read manual pages of udevadm. But these approaches do not help me at all. I think I need to alter either kernel or device internals programatically.
I have also tried manipulating authorized file that resides in /sys/devices/pci0000:00/0000:00:14.0/usb1/1-4 (for a keyboard). It's default value is 1. Changing it to 0 successfully disabled the bus (NOT device, but the physical usb port. The same device can still be used when plugged in another port). But making it 0 also stopped udev events from this usb port. So, I can disable the port if the user plugs in a forbidden device but I cannot decide when to enable it since I cannot listen to remove events in udev. Would it make sense if I delve deeper to lower levels of code (possibly kernel) and listen to usb events in some other way?
I think the simplest way to solve your problem is balcklisting all usb device drivers excepting mouse, keyboard etc.
The cleanest way is whitlisting mouse etc. with udev using usb device id's
Writing a wrapper driver is an absolute overkill never chose that way.
One possible solution could be tying usb access to user permissions. here is a related link you may find handy
http://robots.mobilerobots.com/wiki/Linux_udev_USB_Device_Permissions_Configuration
I am running embedded linux 3.2.6 on an ARM processor. I am using a modified version of atmel's serial driver to control the 4 USART ports on my device. When I use the driver compiled with the kernel, all works fine. But I want to run the driver as a kernel module instead. I make all of the necessary changes and disable the internal driver and everything seems fine. The 4 tty devices are registered successfully and I can see that the all of my probe and initialization functions work correctly.
So here's the problem:
When I try to write to any of the devices, my "start transmit" function gets called but then waits for an interrupt from the usart which never occurs. So the write just hangs, and using a logic analyzer I can see that RTS gets asserted but no bytes show up on the tx line. I know that my call to request_irq succeeds and yet i never see any of the irq entries in /proc/interrupts. In the driver, I have also tried using request_irq to register a separate interrupt handler for a gpio line, and this works fine.
I know that this is a problem that is probably hard to diagnose, but I am looking for ANY possible suggestions that could lead me in the right direction to finding a solution. Let me know if you need any clarifications. Thank you
The symptoms reads like a peripheral clock that has not been enabled (or turned off): the device can be initialized w/o errors and an I/O operation can be setup, but the device doesn't do anything; it plays dead. Since no I/O ever starts, you're never going to get an interrupt indicating completion!
The other thing to check are the conditional compilation directives for HW configuration structures in your arch/arm/mach-xxx/zzz_devices.c file.
Make sure that the serial port structures have something like:
#if defined(CONFIG_SERIAL_ATMEL) || defined(CONFIG_SERIAL_ATMEL_MODULE)
and not just
#if defined(CONFIG_SERIAL_ATMEL)
Addendum
I could be wrong but the clock shouldn't have any effect on the CTS pin causing an interrupt, right?
Not right.
These digital circuits are synchronous state machines: without a clock, a change-of-state by an input cannot be processed.
Also, SoCs and modern uControllers use the peripheral clocks as on/off switches for those integrated peripherals. There is often way more functionality, i.e. peripherals, on the silicon chip than can actually be used, mostly due to insufficient quantity of pins to the board. So disabling the clocks to unused devices is employed to reduce power consumption.
You are far too focused on interrupts.
You do not have a solvable interrupt problem; those are secondary failures.
The lack of output when attempting to transmit is far more significant and revealing.
The root cause is probably a flawed configuration of the USART devices, since transmitting bits is an automatic operation for a configured & operational USART.
If the difference between not-working versus working is loadable module versus static linking, then the root cause is going to be something fundamental (and trivial) like my two suggestions.
Also your lack of acknowledgement regarding the #if defined(), e.g. you didn't respond with "Oh yeah, we already knew that", raises a gigantic red flag that says "Fix me first!"
Addendum 2
I'm tempted to delete this answer after discovering that the Atmel serial driver cannot be configured/built as a loadable module using make menuconfig (which is the premise for half of the answer). (Of course the Kconfig file could be hacked to make the config variable tristate instead of boolean to overcome the module restriction.) I've left a comment for the OP. But I also wanted to preserve the comment to Mr. Stratton pointing out how symbols in the .config file are (not) used.
So I did finally fix my problem. Thank you for the responses, none of them directly solved my problem but they did prompt further examination of my code. After some trial and error I finally got it working. I had originally moved the platform_device structures for each usart from /mach-at91/xxx_devices.c to my loadable module. Well for some reason the structures weren't getting the correct data to map to the hardware, I suppose because it wasn't correctly linking the symbols from the kernel (never got an error message though) and so some of the registration functions weren't even getting called. I ended up moving the structures and platform_device_register calls back into the devices file. I also decided to keep the driver for the console built-in using the original atmel_serial.c driver. I had to change the platform_device name for the console in both the devices file and in the built-in atmel_serial.c file in order for it to not conflict with my usart ports driver. I found that changing the platform_device and platform_driver name for the usarts from anything but "atmel_usart" resulted in usart transmission failing. I really don't understand why, but i'm just leaving it as atmel_usart so it works.
Thanks again to everybody who responded to my problem.
I'm working an a system with embedded Linux (Kernel 2.6.31).
It is a AT91SAM9G20 chip inside, and some of the Pins are forwarded to the outside.
Now I want to use them as GPIO Inputs.
I read the gpio.txt documentation about using the GPIOs via filesystem, and that works very well 'til here. I connected some switches to the gpio-pins and I can see the result in /sys/class/gpio/gpioX/value. But now I'd like to react on a change without busy-waiting in a loop. (i.e echo "Switch1 was pressed").
I guess I need interrupts here, but I couldn't find out how to use them without writing my own kernel driver. I'm relatively new to Linux and C (I normally program in Java), so I'd like to handle the Interrupts via sysfs too. But my problem is, that there is no "edge"-file in my GPIO directory (I guess because this is only since Kernel version 2.6.33+). Is that right? Instead of "edge" I've got a uevent file in there, which is not described in gpio.txt.
In the gpio.txt documentation there was a Standard Kernel Driver mentioned: "gpio_keys". Is it possible to use this for my problem?
I guess it would be better to work with this driver than allowing a userspace program to manipulate kernel tasks.
I found a lot of codesnippets for writing my own driver, but I wasn't even able to find out which of the 600 gpio.h files to include, and how to refer to the library (cross compiler couldn't find the gpio.h file).
Sorry for newbie questions, I hope you could give me some advices.
Thanks in advance
See this for an example on how to do that. Basically, the thing you're missing is the usage of the select or poll system calls.