October 2011 - kernelnewbies@kernelnewbies.org

Changes in block layer for IOScheduler Insertion selection
by mani Feb. 6, 2012

Feb. 6, 2012

Dear All, I am working on linux kernel 2.6.32.9 tegra NVIDIA board. I am getting ~8MBps speed of the Nand disk if i use hdparm hdparm -t /dev/mtdblock3 i made changes in block layer of kernel as below:- block/blk-core.c static inline void add_request(struct request_queue *q, struct request *req) { drive_stat_acct(req, 1); /* * elevator indicated where it wants this request to be * inserted at elevator_merge time */ __elv_add_request(q, req, ELEVATOR_INSERT_BACK, 0); //__elv_add_request(q, req, ELEVATOR_INSERT_SORT, 0); } changed ELEVATOR_INSERT_SORT to ELEVATOR_INSERT_BACK it improves my NAND speed to 10MBps. I am using "noop" I/O scheduler. Will this change have any adverse effect in kernel ? or any other side effect as far as i am using only Nand no Hard disk. Thanks Mani

3 4

Is there a way to build a cross reference with kernel file
by Jimmy Pan Dec. 18, 2011

Dec. 18, 2011

It seems we can use ctags, while, I don't really know how to implement it. I don't want to use the online cross reference every time... Thanks.

6 8

a few questions regarding acpi and kernel configuration
by Littlefield, Tyler Nov. 22, 2011

Nov. 22, 2011

Hello all: I have a few questions, I was hoping someone could help me out here. First, I'll start off by saying that my netbook is a sort of PDA for me. I just boot it up and when I'm taking care of stuff for work running around getting info and that, I can use it to grab the info. I use Linux for this, but my battery dies kind of quickly. I'm also looking to optomize performance as much as possible. So, I sat down and tried to optomize a kernel. Both to save space (since this is a smaller drive), and to make things a bit quicker, even by a little bit. I put the config here: http://tds-solutions.net/config.tar.bz2 First, I'm curious what else I can yank out of there to make everything smaller. I'd like to cut down on the drivers and all that, and just include the driver for my specific harddrive if that's possible, as well as the wireless card. Can I turn off scsi/pata/ata since I have a sata drive? Is there other stuff I missed that I could take out? Finally, my last two questions. 1) Is there a way to access acpi so I can put the computer into standby quickly? I'm not sure how to pull that off from the cli, but doing it with code would work. 2) Is there a way for me to detect when the lid of my laptop has been closed and put it into standby, and/or set an idle timer to put it into standby? If there are clients out there that can interface with acpi and help with this, awesome. If not, and I need to code something up, that works too. Finally, I'm trying to find a way, since I do not use it to turn off the display. Is this possible? -- Take care, Ty Web: http://tds-solutions.net The Aspen project: a light-weight barebones mud engine http://code.google.com/p/aspenmud Sent from my toaster.

3 2

why does the kernel use mutex to protect the address_space structure
by Zheng Da Nov. 13, 2011

Nov. 13, 2011

Hello, I try to understand why the kernel uses mutex to protect the address_space structure. I checked the critical areas protected by i_mmap_mutex, and didn't see these areas use IO. I don't see why we can't use spin locks. Do I miss something? It seems most of critical areas provided by i_mmap_mutex aren't very large. Thanks, Da

2 2

Question on CDROM access
by sahlot arvind Nov. 4, 2011

Nov. 4, 2011

Hi, I am running Ubuntu 11.10 and my CDROM drive's LED keeps blinking. I want to know which process is accessing CDROM for no reasons. There is no media in the drive, I believe OS seems to access it for media change events, but I am more interested in which process or deamon is actually doing it. I tried the followings and nothing gave anything: 'lsof | grep -in 'cdrom' 'lsof | grep -in 'sr0' 'lsof | grep -in 'sg0' What tools could I use to track down the process, which is accessing my CD drive? Thanks in advance. Asahlot

2 1

seagate drive and i/o errors
by Littlefield, Tyler Nov. 1, 2011

Nov. 1, 2011

Hello all, I have a quick question. When I do a ls on this drive and it's spun down, i get an i/o error. I'm curious if there's a way to write a module or do something to make the drive wake up when I do a r/w request on it? I've had an external and it never had this problem. I also know i could modify the scsi params to make it not sleep, but that doesn't work and that is not the best of ideas anyway. I like it sleeping, I don't even mind waiting for an LS. It's just hard to manage when i have to umount -l the directory, then figure out what sd* dev it's on. -- Take care, Ty Web: http://tds-solutions.net The Aspen project: a light-weight barebones mud engine http://code.google.com/p/aspenmud Sent from my toaster.

2 1

generic I/O
by Kai Meyer Oct. 31, 2011

Oct. 31, 2011

Are there existing generic block device I/O operations available already? I am familiar with constructing and submitting 'struct bio's, but what I'd like to do would be greatly simplified if there was an existing I/O interface similar to the posix 'read' and 'write' functions. If they don't exist, I would probably end up writing functions like: int blk_read(struct block_device *bdev, void *buffer, off_t length); int blk_write(struct block_device *bdev, void *buffer, off_t length); Pros and cons to this sort of approach? -Kai Meyer

1 0

ext4 extents
by Mike Gibson Oct. 30, 2011

Oct. 30, 2011

I'm updating a user space program that finds free blocks on a ext2/3 filesystem so that it works with ext4. I'm concerned about how extents might affect my program. My program currently looks only at the block allocation bitmaps. Are the block allocation bitmaps still updated when an extent is created? >From what I can tell, extent information is only stored in the inode, and is only a way to improve the efficiency of the process of creating the structure needed in the inode. But I haven't been able to find out anything about whether the block allocation bitmap is updated at the same time. Do extents provide a new way of expressing which blocks are allocated on disk? Or are the block allocation bitmaps still the canonical way to determine if a given block is in use? Mike Gibson

2 2

[ANNOUNCE] Xvisor: eXtensible Versatile hypervISOR
by Chauhan, Himanshu Oct. 28, 2011

Oct. 28, 2011

We are pleased to announcing Xvisor v0.1.0, a new open source bare-metal hypervisor, which aims towards providing virtualization solution, which is light-weight, portable, and flexible with small memory foot print and less virtualization overhead. It is distributed under GNU Public License (GPLv2). Xvisor has most of the features expected from a modern full fledged hypervisor: - Tree based configuration - Tickless and high resolution timekeeping - Threading framework - Device driver framework - CPU virtualization - Address Space virtualization - Device emulation framework - Serial port virtualization - Managment terminal Xvisor is currently being ported for two architectures: ARM and MIPS. The development & testing is being done using QEMU (0.14.xx or higher). Xvisor ARM is able to boot multiple unmodified Linux-2.6.30.10 or Linux-3.0.4 guest with a fairly interactive and smooth Busybox 0.19.2 console. Currently supported host for Xvisor ARM is Realview-PB-A8 Board emulated by QEMU. It is also being ported to real hardware, specifically Beagle Board. Please try out our Xivsor ARM demo on QEMU (0.14.xx or higher). To download the demo tarball visit the link below: https://docs.google.com/a/brainfault.org/leaf?id=0B0ABS_s60oP_OGE2MDA0MWYtN… Xvisor MIPS is still a work in progress and will be announced very soon. Xvisor is currently hosted on Github, to clone or download source code please visit following links: Mainline: https://github.com/xvisor/xvisor ARM Development: https://github.com/avpatel/xvisor-arm MIPS Development: https://github.com/hschauhan/xvisor-mips Our developer mailing list is xvisor-devel[at]googlegroups[dot]com, please feel free to post your queries or join our development mailing list.

1 0

Kernel Module from C program
by Luis Morales Oct. 27, 2011

Oct. 27, 2011

Greetings, I'm new to the world of kernel modules programming. I've read the Linux Kernel Module Programming Guide by Peter Jay Salzman, among other documents, and I have the following question: I'd like to know if it's possible to take a program developed in C (in this case OpenSIPs, which is a SIP Proxy Server) and turn it into a kernel module (I know "to turn it" is probably not the right expression, what I mean is if I could modify it somehow so that it can be loaded as a kernel module) ? As far as I've understand, kernel modules are developed for drivers or to do some modifications to system calls, but I like to know if it's possible to do what I stated earlier. Thanks, Luis Morales.

2 1