- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

Re: Finding physical and logical core number in a kernel module ?
by Alexandre Courbot Sept. 6, 2011

Sept. 6, 2011

> I agree that the kernel module is not restricted to any core but is it > possible that whenever it is executed > it gives the core number whichever is executing it ? Have a look at the end of include/linux/smp.h: smp_processor_id() gives you the number of the current executing CPU. get_cpu() will do the same and will also disable preemption so that you will stay on that CPU until put_cpu() is called. >From user-space, you can use sched_getcpu() to obtain the same information. Hope this helps, Alex.

1 0

CRC calculation in Module.symvers
by Shakthi Kannan Sept. 6, 2011

Sept. 6, 2011

Greetings! I would like to know as to: 1. How the CRC is computed for an exported symbol in the Linux kernel, and 2. Given a kernel .c file, how can I obtain the CRC checksums for a symbol. For linux-3.0.2 (example), I compiled with CONFIG_MODVERSIONS=y, and running make built vmlinux, and generated the Module.symvers file. I believe genksyms is used to generate the CRC information. I had referred the "Kernel Symbols and CONFIG_MODVERSIONS" documentation [1] written by Mark McLoughlin. I tried the following in the top-level kernel sources: $ gcc -E -D__GENKSYMS__ -DCONFIG_MODVERSIONS -DEXPORT_SYMTAB crypto/api.c -I/tmp/linux-3.0.2/include -I/tmp/linux-3.0.2/arch/x86/include | ./scripts/genksyms/genksyms and I get the following error output: === Output === /tmp/linux-3.0.2/include/linux/jiffies.h:43:42: error: division by zero in #if In file included from /tmp/linux-3.0.2/include/linux/tracepoint.h:19:0, from /tmp/linux-3.0.2/include/linux/module.h:18, from crypto/api.c:22: /tmp/linux-3.0.2/include/linux/rcupdate.h:160:2: error: #error "Unknown RCU implementation specified to kernel configuration" In file included from /tmp/linux-3.0.2/include/linux/module.h:21:0, from crypto/api.c:22: /tmp/linux-3.0.2/arch/x86/include/asm/module.h:59:2: error: #error unknown processor family === End === The documentation for genksyms is old, and now there is no -k option? Appreciate any inputs in this regard, Thanks! SK [1] Kernel Symbols and CONFIG_MODVERSIONS. http://www.skynet.ie/~mark/pub/symbols.txt -- Shakthi Kannan http://www.shakthimaan.com

1 0

Finding physical and logical core number in a kernel module ?
by Vaibhav Jain Sept. 6, 2011

Sept. 6, 2011

Hi, Are there kernel functions that would return the number of the physical core and logical core (in case of Hyperthreading) on which a kernel module is running ? Thanks Vaibhav Jain

2 1

Uniquely Identify Machine
by Gaurav Mahajan Sept. 6, 2011

Sept. 6, 2011

Hi, I wan to recognize a particular machine uniquely. I would like to know which attribute of any machine can uniquely identify a particular machine. The machine may be a Virtual machine. Are BIOS information for machine are unique???? Is processor id is unique???? or we can form unique identification number of machine with the combination of the bios attributes.???? Regards, Gaurav Mahajan.

3 3

Parallel IO port conflict (driver problem)
by Shenli Zhu Sept. 6, 2011

Sept. 6, 2011

Hi all, I am reading "Linux device driver 3rd". I try to insmod the "short" example(a example use parallel port). But the port 0x378 is allocated to parport. And parport is used by ppdev, parport_pc and lp. Shall I rmmod all these 4 module to insmod "short"? Thanks, David

2 1

Can we put the pages allocated by alloc_pages in the page cache?
by Da Zheng Sept. 5, 2011

Sept. 5, 2011

Hello, I'm modifying a filesystem. I need to allocate multiple contiguous pages to store data, and I use alloc_pages to do so. In order to reduce memory copy, I should put these pages directly to the page cache. I don't know if it's proper. Can I allocate multiple pages together and free them and return them to the buddy system separately? VFS uses page_cache_alloc to allocate a page each time, so the pages can be freed independently. My other concern is the reference counters of the pages. When I allocate multiple contiguous pages with alloc_pages, only the first page's counter is 1. The reference counters of the remaining pages are 0. It confuses me. Should I set the reference counters of all pages to 1 first before I add them in the page cache? Understanding the Linux Kernel says a page is unused when its reference counter is -1, instead of 0. I wonder who is using the page when its counter is 0. Thanks, Da

1 0

kernel 3.0.4 installation failed
by Vishnupradeep Sept. 4, 2011

Sept. 4, 2011

Downloaded and compiled the Kernel 3.0.4 (linux-3.0.4.tar.bz2) Full compile log. http://pastebin.com/qbQ4tREb Then installed using the following command [vishnu@localhost linux-3.0.4]$ sudo make modules_install install [sudo] password for vishnu: INSTALL arch/x86/ia32/ia32_aout.ko INSTALL arch/x86/kernel/test_nx.ko INSTALL drivers/scsi/scsi_wait_scan.ko INSTALL fs/ext4/ext4.ko INSTALL fs/jbd2/jbd2.ko INSTALL lib/crc16.ko INSTALL net/netfilter/xt_mark.ko DEPMOD 3.0.4-vish sh /media/Portable/linux-kernel/tar_packs/linux-3.0.4/arch/x86/boot/install.sh 3.0.4-vish arch/x86/boot/bzImage \ System.map "/boot" [vishnu@localhost linux-3.0.4]$ Just need these modules ?? The system is not booting up with just these files installed. -- -------------- *When you do something right, people wont notice you've done anything at all.*

3 3

How make modprobe find my kernel module?
by Parmenides Sept. 3, 2011

Sept. 3, 2011

Hi, I have write a 'hello, world!' module which is in a directory rather than the kernel source tree. I compiled it by: make -C /usr/src/linux SUBDIRS=$PWD modules and installed it by: make -C /usr/src/linux SUBDIRS=$PWD modules_install I find it was installed at /lib/modules/2.6.34/extra. Then, I invoked modprobe hello.ko to load this module, but get a message: FATAL: Module hello.ko not found. It seems that my module is not in the modprobe's search path. Is that true? If so, how can I configure its search path?

3 2

Compile a Single Kernel Module
by Nestor Waldyd Sept. 1, 2011

Sept. 1, 2011

Hello, I am a Fedora 15 user and i would like to compile only certain kernel drivers (rt2x000) that are modules for the linux kernel. How can this be done? Any documents recommendations? Beforehand Thanks. Best Regards, -- -------------------------------------------------------------- Nestor Waldyd Alvarez Villa Ingeniero Electrónico. MSc. C. Telecomunicaciones. Universidad Pontificia Bolivariana mailto:nestor.alvarez@upb.edu.co Medellín - Antioquía. COLOMBIA -------------------------------------------------------------- Using Opera's revolutionary email client: http://www.opera.com/mail/

2 1

Kernel Lock Debugging
by Venkatram Tummala Aug. 31, 2011

Aug. 31, 2011

Hi, i wrote a linux kernel module which uses mmap_sem in lots of places. I grab mmap_sem for both reads & writes. And, i also use page_table_lock spinlock under them. The problem i have is that there is a hard to reproduce a race condition bug in which a thread doesn't release the semaphore at all. Hence, ps hangs as ps needs to grab mmap_sem for the processes. The question is how do i find the thread that is holding the semaphore forever. It will help me figure out the part of the code the race is in. So, the question is : I have some number of threads grabbing the semaphore for read/write. One of them grabs it & doesn't release at all. How do i find the thread that is holding it. This is not a soft lockup & i dont get any hung_task_timeout messages. Any help is appreciated. Thanks. Venkatram

3 3