November 2011 - kernelnewbies@kernelnewbies.org

Re: Memory mapping in X86
by Sathishkumar Duraisamy Nov. 19, 2011

Nov. 19, 2011

Hi Santhosh, I got this thing in dmesg: [10256.068664] driver_test:7114 map pfn RAM range req uncached-minus for 2f866000-2f867000, got write-back. I am writing some random numbers in mmapped memory. But I am not able to get it from the user space. Note: ( I have allocated memory via adc_dev.adcMemoryArea = (adcdata_t *) __get_free_pages(GFP_KERNEL, 1); ) On Sat, Nov 19, 2011 at 12:16 PM, santhosh kumars <karuna.linux(a)gmail.com> wrote: > hi, > what is the exact error you are getting. > thanks, > santosh. > > On Mon, Nov 14, 2011 at 10:26 AM, Sathishkumar Duraisamy > <bewithsathish(a)gmail.com> wrote: >> Hi, >> >> I trying to implement mmap method (And by the way I writing driver to >> model adc). I had implemented as shown below. I am not getting the >> result as I expected. But got some error: map pfn RAM range req >> uncached-minus for 36a0e000-36a0f000, got write-back. >> >> In user space, I call mmap as follows: map_ptr = (signed short *) >> mmap(0,ADC_USER_MEM_SIZE,PROT_READ, MAP_SHARED,fd,0); >> >> Am I doing anything wrong here? >> >> >> static int simdrv_mmap(struct file *filp,struct vm_area_struct *vma) >> { >> unsigned long off = vma->vm_pgoff << PAGE_SHIFT; >> unsigned long physical = (__pa((void *)adc_dev.adcMemoryArea) >> >> PAGE_SHIFT) + off; // adcMemory I got via >> __get_free_pages(GFP_KERNEL, 1); >> unsigned long vsize = vma->vm_end - vma->vm_start; >> unsigned long psize = __pa((void *) adc_dev.adcMemoryArea) - off; >> >> // Module spanning too high >> if (vsize > psize) >> { >> printk(KERN_ALERT "Module spanning too high"); >> return -EINVAL; >> } >> >> vma->vm_flags |= VM_IO; >> vma->vm_flags |= VM_RESERVED; >> vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot); >> >> if (remap_pfn_range(vma,vma->vm_start, physical,vsize,vma->vm_page_prot)) >> { >> return -EAGAIN; >> } >> printk(KERN_ALERT "Sucessfully memory mapped"); >> return 0; >> } >> -- Regards, Sathishkumar D http://flowersopenlab.weebly.com/

1 0

Is an IRQ line disabled on local CPU or globally?
by Parmenides Nov. 19, 2011

Nov. 19, 2011

Hi, It is said that an IRQ line is disabled when the same IRQ line is in process. But, is the IRQ line disabled on the local CPU or all CPUs? Another relevant question is that the disable_irq() funciton will disable an IRQ line on local CPU or all CPUs.

2 1

Re: Kernelnewbies Digest, Vol 12, Issue 25
by Sazzad Nov. 18, 2011

Nov. 18, 2011

> > <surajkhurana27(a)gmail.com <mailto:surajkhurana27@gmail.com>> wrote: > > Hi all, > I am a newbie in linux world and want to learn linux device drivers. > I'm also new to Linux device driver world and am interested to learn. Besides learning and fun, I want to write a driver for Green Packet's<http://www.wimaxian.com/2009/12/02/green-packet-dx-tower-and-uh-shuttle-sty…> UH-235 wimax modem. Probably you could also join? :) If you are interested, we could find out, as a first step, How the Kernel responses, when *any* hardware is attached to a system that > runs Linux? Probably, dmesg <http://en.wikipedia.org/wiki/Dmesg> command can help us a bit. Our next step would be to find out, How to hack/know about a device attached with a Linux system, provided that > there is no driver currently available for the device? I've also collected the source file provided by Green Packet. It was actually a deb file and I extracted it. There is really no source except shared object (.so/.ko) files. This deb file, reportedly, only works in Ubuntu 9.04, which uses Linux Kernel Version *2.6.28*. Any other installation on newer Ubuntu failed. Possibly be a Kernel issue? Let's find out! If you want the deb/source file, let me know. I'll mail you. Green Packet's UH-235 wimax modem uses Mediateks m7118 SoC. And reportedly, Mediatek's chips are not developer friendly. Like they don't provide useful information to the developers outside. :P Happy hacking! -- Sincerely, Sazzad Bin Kamal

1 0

User Mode Linux
by Vishal Thanki Nov. 18, 2011

Nov. 18, 2011

Hi All, I am trying to create a loopback block device under a UML (running kernel 3.1.0). I used following commands: # dd if=/dev/zero of=/tmp/store1 bs=1024 seek=2047 count=1 # losetup /dev/loop1 /tmp/store1 But the 2nd command (i.e. losetup) returns with the error complaining "/dev/loop1" is not present, but I can see /dev/loop[0..7] files present. My question, is it possible to do this exercise under UML (User mode linux)? Thanks, Vishal

3 4

Registering for only Specific Packets from Net-Filter!!!!
by Kesava Srinivas Nov. 18, 2011

Nov. 18, 2011

HI Guys, I'm a Newbie to Linux Kernel Development. Need some Clarification in understanding Net-Filter Hooks. Can I register only for the Necessary Packets [Ex: IPSEC etc...] from the Net-Filters?? AFAIK ; Net-Filter Module send the Packets to the Corresponding Module based on pre-defined HOOKS. Let us say; there are Many Packets reaching that Hooks; and if I am Interested only in some Packets from Net-Filter module; can I do that ??? Any Help is Really Appreciated !! -Thnks, VKS

4 5

Can I get the information of bounce limit from sys?
by loody Nov. 18, 2011

Nov. 18, 2011

hi all: I saw sys can show some information such as max_sectors_kb , physical_block_size, hw_sector_size and max_segments. Where i can get bounce limit information, except digging through source code. -- Regards,

1 0

Getting Information from Multiple nodes
by Praveen kumar Nov. 18, 2011

Nov. 18, 2011

Hi All, I have a I2C chip driver,and there are two identical chips on two different i2c bus. I have registered the driver and initialized it and created two nodes /dev/chip_0 and /dev/chip_1 If I open the node from user how will I make driver identify which device to access .ie in the ioctl. Am I clear ??? Praveen

2 2

Re: User Mode Linux
by Vishal Thanki Nov. 17, 2011

Nov. 17, 2011

On Thu, Nov 17, 2011 at 7:04 PM, Anuz Pratap Singh Tomar <chambilkethakur(a)gmail.com> wrote: > > > On Thu, Nov 17, 2011 at 12:04 PM, Vishal Thanki <vishalthanki(a)gmail.com> > wrote: >> >> On Wed, Nov 16, 2011 at 9:14 AM, Vishal Thanki <vishalthanki(a)gmail.com> >> wrote: >> > >> > On Tue, Nov 15, 2011 at 8:09 PM, Lakshmipathi.G >> > <lakshmipathi.g(a)gmail.com> wrote: >> > > But the 2nd command (i.e. losetup) returns with the error complaining >> > > "/dev/loop1" is not present, but I can see /dev/loop[0..7] files >> > > present. >> > > >> > > Try loading the module as - >> > > >> > > #modprobe loop >> > >> >> I have one more question in the same line (i.e. related to UML). I >> tried loading an external module (build for ARCH=um type). I am able >> to load that module only if i've booted UML with following command: >> >> ./linux ubda=<raw_file_system_path> >> >> If i pass the mem parameter, i.e. >> >> ./linux ubda=<raw_file_system_path> mem=512M >> >> I am not able to load the module. The error that I am seeing is: >> >> overflow in relocation type 11 val a121c3c0 >> `tun' likely not compiled with -mcmodel=kernel >> >> I tried to load the tun.ko. This holds true for any manually built >> simple hello_world kind of module too (where i added -mcmodel=kernel >> parameter in Makefile). >> >> What can be wrong with "mem=512M" parameter? I change the size of mem >> to 128K to 1G, but nothing works. It just doesn't like "mem" parameter >> as bootargs. >> >> http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > > try compiling it statically I gave tun.ko as an example. I want to load my own module which is not part of linux kernel.

1 0

How can I know what is tainting my kernel
by Daniel Hilst Selli Nov. 16, 2011

Nov. 16, 2011

Two questions about taint system.. 1. I have a installation of archlinux with kernel 3.0, just after boot I cat /proc/sys/kenrel/tainted and receives 1024, but cat /var/log/messages | grep -i taint gives me no output. How can I know what is tainting my kernel? 2. I was studying char devices, and have writed some code. I have put the MODULE_LICENSE("GPL"); on the module, build it, put it on a initrd image and boot with qemu using -kernel and -initrd options. When I loadit with modprobe I receives a message on console saying that kernel was tainted and /proc/sys/kernel/tainted gives me 4024. There is any other thing on module, besides license that can taint the kernel? Here is the code -> http://sprunge.us/RHNa?c []'s

3 5

Generic I/O
by Kai Meyer Nov. 15, 2011

Nov. 15, 2011

I'm finding it's really simple to write generic I/O functions for block devices (via a "struct block_device") to mimic the posix read() and write() functions (I have to supply the position, since I don't have a fd to keep a position for me, but that's perfectly ok). I've got a little hack that allows me to run synchronously or asynchronously, relying on submit_bio() to create the threads for me. My caller function has an atomic_t value that I set equal to the number of bios I want to submit. Then I pass a pointer to that atomic_t around to each of the bios which decrement it in the endio function for that bio. Then the caller does this: while(atomic_read(numbios) > 0) msleep(1); I'm finding the msleep(1) is a really really really long time, relatively. It seems to work ok if I just have an empty loop, but it also seems to me like I'm re-inventing a wheel here. Are there mechanisms that are better suited for waiting for tasks to complete? Or even for generic block I/O functions? -Kai Meyer

2 3