June 2015 - kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

blank mmodules.dep
by Shérab June 16, 2015

June 16, 2015

Dear all, I'm trying to compile a kernel from Linus' tree, commit 181e505952c5e4b4beb287d0022ea3cd20f23e74 The .cnfig I use is here: http://paste.debian.net/230990/ As far as I can tell, make works fine and so does make modules_install install But after all this, the modules.dep file I get in /boot is empty. I couldn't find any clue so far. Can anybody help,please? Thanks a lot! Shérab.

2 3

refrence to communicate with hardware
by john byron June 15, 2015

June 15, 2015

Anyone know where i need to do/read to know all the flag (or something else in low level, or bit level) to communicate with hardware ? thank you

2 1

Re: Kernel contributions from organisations and individual privacy
by Valdis.Kletnieks＠vt.edu June 15, 2015

June 15, 2015

On Sun, 14 Jun 2015 19:36:23 -0400, nick said: > a embedded version of Linux called NX OS I believe and its all open to my > knowledge. Again this is stupidity in my option or companies being lazy by > hiding there terribly written drivers. Remember Nick - it's not a technical decision, it's a business decision. If a company can improve the bottom line by $1M by doing something technically ugly, something technically ugly *will* be done. And it's usually no use standing up and fighting it if you work at the company, because quite frankly, you're expendable. Unless you're a superstar who will cost more than $1M in recruiting costs and signing bonuses to replace, the company will can your ass, recruit a replacement, do the ugly thing, and *still* come up with more money. And let's face it - if you're a superstar, why are you working at a company whose product is so messed up that technically ugly work-arounds are called for? :)

1 0

Re: Kernel contributions from organisations and individual privacy
by Valdis.Kletnieks＠vt.edu June 14, 2015

June 14, 2015

On Sun, 14 Jun 2015 17:25:40 -0400, Nicholas Krause said: > Why are we even trying to prove this, it seems any intelligent company who has > a interest in staying in business will comply The problem is companies that have mixed in their own "secret sauce", and are afraid of said sause becoming non-secret as part of their required GPL compliance. If you have a competent legal department, you'll file lots of motions and so on to delay that final court order as long as you can. See VMWare for the details - they've got enough legal staff that it took like 7 years for Hellwig to team up with funded legal support. And it will almost certainly be an eventual loss for VMWare - but they can delay that for months or years. And VMWare certainly doesn't want to release their code - mostly because it's almost certainly a creeping horror worthy of mockery. If you don't believe me, go look at the GPL kernel drivers for VMWare Workstation or Player, which are probably similar to the ESXi code that VMWare would have to release. How bad are the WS/Player drivers? Let's just say that Greg KH probably wouldn't let them into drivers/staging. :)

1 0

Building which kernel?
by John Whitmore June 14, 2015

June 14, 2015

I've a question regarding the building of the linux kernel, more specifically which kernel to build. I've previously built the kernel for an 86 arch but this time I want to build for the RaspberryPi. I can easily get the current version of the installed kernel with the "uname -a" command but that doesn't tell me the git repo where the kernel originated. I'd imagine that it's from the RPi git repo, but that adds another level of confusion as that git repo doesn't have any tags at all. So even given the version number from uname how to check out a suitable version of the kernel. I could take the latest from the RPi repo and use "make oldconfig" but I wonder how far, in kernel versions, I can jump and not affect the installed libs and utilities which have been built against another version of the kernel? I've read various documentation on building the linux kernel but that's never really mentioned, mainly I guess as you're building for the x86 Arch. But even there can you jump major versions of kernel without effecting other installed SW in the system? John

3 3

Global variables
by roni June 14, 2015

June 14, 2015

I know i should avoid using global variables in kernel programming. What problem it create if i use more global variables?

3 3

opening a BUG in kernel bugzilla (https://bugzilla.kernel.org/)
by Kevin Wilson June 12, 2015

June 12, 2015

Hi all, Can we open a BUG in kernel bugzilla (https://bugzilla.kernel.org/) against a kernel image which was found on an image which was built from most recent build of Linus tree (https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git) but is not reproducible (yet) against the official kernel images on https://www.kernel.org/ ? Or could it be event worse ? I mean, could we open BUGS against **each** of the images under kernel.org (for example, 4.1-rc7, which is the latest) ? Or only for those labeled as "longterm" or "stable" ? Regards, Kevin

2 1

Userpace serial port
by Ricardo Ribalda Delgado June 11, 2015

June 11, 2015

Hello I have a serial port accesible via Ehtercat. http://www.beckhoff.dk/english.asp?ethercat/el6001.htm Ethercat is a fancy way of accessing i/o devices via ethernet. You can read/write registers with close to real-time restrictions. Nowadays there is a kernel land implementation of ethercat master, but it is not merged (and it is definately not an easy job). And a software implementation http://fh-soft.de/src/ethercat-userspace.html My hope is that there is something like uinput, but for serial devices,i.e. I make a userland program that looks like a normal serial device to the other applications. Does this exist? Regards! -- Ricardo Ribalda

3 5

what's the mean of num in address_sapce
by feng247＠yeah.net June 11, 2015

June 11, 2015

Hi, in kernel 2.6.39, in /include/linux/compiler.h, I found the followed code: # define __percpu __attribute__((noderef, address_space(3))) #ifdef CONFIG_SPARSE_RCU_POINTER # define __rcu __attribute__((noderef, address_space(4))) #else # define __rcu address_space(num), if num is 0 ,it means kernel space, 1 , user space; 2, io space, so 3 and 4 means what? I found nothing about this by google. Thanks, feng feng247(a)yeah.net

1 0

miscdevice and default attributes
by David Bakin June 8, 2015

June 8, 2015

Where do I set the default attributes for a device created with misc_register()? In kernel 3.19 struct miscdevice looks like this: struct miscdevice { int minor; const char *name; const struct file_operations *fops; struct list_head list; struct device *parent; struct device *this_device; const char *nodename; umode_t mode; }; No attributes there that I can see. I know that after calling misc_register() the this_device field is filled in and I can use device_create_file() to make attributes with it. But it is too late to use this_device's own default attributes, right? The device has already been created. The reason this is puzzling me is because Greg KH, here <http://kroah.com/log/blog/2013/06/26/how-to-create-a-sysfs-file-correctly/> (in 2013), explains that the right time to set up attributes is when the device is created, using the device's, the driver's, the class's, or the bus's default attributes field - if you wait until later you run the risk of a losing a race against udev. But maybe I'm just confused (I'm rather new here, after all), because here <http://lists.kernelnewbies.org/pipermail/kernelnewbies/2015-April/013894.ht…> Greg explains that all udev is doing is setting up permissions and symlinks, so maybe I don't care about the potential race (if I don't need different permissions). I guess another way to ask this question is to ask: What are the circumstances when I should be concerned about this race between the driver and udev? In the case of a new misc device for my own private purposes ... it seems unlikely I'd want to use udev to rename it in any particular way or modify permissions separately from the umode I give the kernel, so ... nothing to worry about, right? Thanks! -- David

2 1