October 2016 - kernelnewbies@kernelnewbies.org

where is __percpu_seg initialized?
by 落痕 March 28, 2018

March 28, 2018

Hi all: I knew the kernel implement the per-cpu variable via __percpu_seg. For example on x86 architecture, it's defined as: #ifdef CONFIG_X86_64 #define __percpu_seg gs #define __percpu_mov_op movq #else #define __percpu_seg fs #define __percpu_mov_op movl #endif But I'm curious where __percpu_seg is initialized?? I fail to find the relative code. Thanks.

2 2

How to handle Hotplug with UIO userspace driver
by Mandeep Sandhu Sept. 28, 2017

Sept. 28, 2017

Hi All, We're working on a custom PCIe based hardware, which can be hotplugged into the system. We're using the UIO approach of having a minimal kernel driver which basically handles the interrupt and provides a mmap of the device memory to userspace and userspace driver which mmap's the provided memory and manages the device. We're using systemd and udev rules to spawn off a user process whenever the hardware is plugged into the system. Now my question is, how can we gracefully inform the userspace process if someone pulls out the hardware (sitting on a PCI slot). Even with full hotplug support, there can always be a case of "surprise removal", where someone yanks the board without going through the whole "attention button" routine. While the kernel driver has no problem with hot-removal, I see that if I "unregister" our UIO driver in response to the removal, then the mapping for the hardware's physical memory would become "invalid" (done by UIO on unregsiter). This could possibly cause the user-space process to crash due to illegal memory access. Is there a "standard" way to handle this scenario, i.e for hotpluggable hardware using UIO? Basically, we'd want to "unregister" the UIO device only after the userspace process is informed of the removal so that it could unmap/close the device and exit gracefully. We could try doing it using some systemd trick, where we put a dependency on one of the UIO created files (/dev/uioX), so that when the file is destroyed, systemd could kill/shutdown the process. Although, there's no guarantee that the userspace process won't try to access the mmaped space during that interval, so this isn't a reliable solution either. Any hints, thoughts? Thanks., -mandeep

3 7

What are need to makes humidifier?
by 윤영석 April 1, 2017

April 1, 2017

Hi, I want to make own humidifier. What sensor need when making humidifier?? humidity indicator (detecting humidity), ultrasonic waves maker?? Please shares your wisdom to me. Thanks. Best regards.

2 1

Append Localversion to Kernel String
by Marco Gross Nov. 20, 2016

Nov. 20, 2016

Hello folks, I'm trying to automatically append version information to my kernel's version string. Therefore, I have in my .config following option: CONFIG_LOCALVERSION_AUTO=y Additionally, following is set: CONFIG_LOCALVERSION="++" The output that is expected is said to be the first eight characters of the output of 'git rev-parse --verify HEAD'. The output in my kernel directory is: $ git rev-parse --verify HEAD c8d2bc9bc39ebea8437fd974fdbc21847bb897a3 After installing the kernel the name of the kernel is as follows: vmlinuz-4.8.0++ I expected: vmlinuz-4.8.0++-gc8d2bc9b My question: Where is the version string that I expected? Marco Mit freundlichen Gr??en / Best regards Marco Gross Diplomand [cid:LogoFritzStephan72dpi120x55Pixel_cfdd7a92-cbea-435d-a419-8d088e6ed7a8.png] Fritz Stephan GmbH - Medizintechnik - Kirchstra?e 19, D-56412 Gackenbach Tel.: +49 6439 9125-249 Fax: +49 6439 9125-111 E-Mail: mgross(a)stephan-gmbh.com<mailto:mgross@stephan-gmbh.com> Web: www.stephan-gmbh.com Handelsregister Amtsgericht Montabaur HR-Nr. 4241 Gesch?ftsf?hrer: Tanja Stephan, Georg Mainusch [cid:Eve_Signatur_145525f8-5839-48a6-8419-4e43941739b5.jpg] [cid:YouTube-logo-full_color_362f623d-4265-4133-9721-da6b905f4c25.png] <https://www.youtube.com/channel/UCtILKk6D8klJo5FwlycY82w>

3 2

Review on spi driver
by Arthur LAMBERT Nov. 19, 2016

Nov. 19, 2016

Hi, Not sure that it is the good place to do that. I am quite new in kernel dev and it will be great to have some review/advise on a implementation in kernel. My need : Read periodically data from a sensor with spi. I know that data is available thanks to a GPIO. Data are available every 2 ms. First step for me was to : Declare a irq handler on the ready signal GPIO thanks to gpio_to_irq and request_irq function. First I try to read from spi in my irq handler.I finally understand that I cannot make call which implies sleep in irq handler like spi_sync for example. I finally use wake_up_interruptible call in my irq handler to make the spi work in another piece of code. static irqreturn_t irq_handler(int irq, void *id) { flag = 1; wake_up_interruptible(&wq); return IRQ_HANDLED; } On my first try I just use a kernel thread which read data from spi and then which send my data to userspace thanks to a netlink socket. while(!kthread_should_stop()) { wait_event_interruptible(wq, flag != 0 || kthread_should_stop()); if (kthread_should_stop()) do_exit (0); flag = 0; // read data from spi (...) // send data in netlink socket skb = nlmsg_new(NLMSG_ALIGN(msg_size + 1), GFP_ATOMIC); nlh = nlmsg_put(skb, 0, 1, NLMSG_DONE, msg_size + 1, 0); memcpy(nlmsg_data(nlh), spi->rx_buf, msg_size); ret = nlmsg_multicast(nl_sk, skb, 0, MYGRP, GFP_ATOMIC); schedule(); // not sure that I need this line of code. } Problem I was not sure that netlink socket was the best solution to my problem. Also I was not able to use the same socket during my data acquisition. I must call nlmsg_new on each new data sample. Other strange stuff, If I call nsmsg_free at the end of the process I have big stability issue during my test. Moreover, if I am to slow to send data from kernel space to user space, I will miss event from my irq handler and miss sample from my sensor. My second try was to use a character device to send data from kernel space to userpsace. I was already using the character device to drive my kernel driver. Open/Release to init/clean sensor. And ioctl to send command to the sensor like start/stop acquisition. So my plan was just to add a read callback on the character device. static const struct file_operations fops = { .owner = THIS_MODULE, .unlocked_ioctl = ads_ioctl, .open = ads_open, .release = ads_release, .read = ads_read, .unlocked_ioctl = ads_ioctl, }; static int ads_read(struct file *filp, char __user *data, size_t len, loff_t *ppos) { struct ads_dev *dev = filp->private_data; wait_event_interruptible(wq, flag != 0); flag = 0; // read data from spi (...) copy_to_user(data, spi->rx_buf, DATA_SIZE); return DATA_SIZE; }; This code seems to be easy compare to the netlink solution. Even in the user space code side. Just need to use a read instead of using socket and libnl. But the problem is always there. If the read operation over the spi and the copy_to_user take more than 2 ms, I will lost sample from my sensor. My last plan was to use mixed the two previous solutions and use a kthread + the character device. The irq wake up the thread. I read data over spi. I put the data in a queue. The read callback from character device will gather data from the queue. I have always a synchrone call which can take more than 2 ms. But At least I can send the data to user space in a synchrone way. But I was not able to find a proper way to use a queue with semaphore. So it was quite basic. A static queue + offset + wake_up_interruptible/wait_event_interruptible to synchronize kernel thread and my read callback. static uint8_t queue[50 * DATA_SIZE]; static int read_offset = 0; static int write_offset = 0; I guess that my problems are quite common in kernel driver which want to gather some data over spi. Is there a good kernel driver that I can take as good practice/example to implement my solution ? Feel free to say that my code is stupid ! Thanks, Arthur.

2 1

ip rule suppress_prefixlength
by Matthias Peter Walther Nov. 1, 2016

Nov. 1, 2016

Dear Linux-friends, I asked this question before on the netdev-Mailinglist, but never got an answer (Link: https://www.mail-archive.com/netdev@vger.kernel.org/msg133574.html) Maybe someone here can help me. I have a configuration like this: root@des1 ~ # ip rule 0: from all lookup local 32765: from all iif lo lookup ffnet suppress_prefixlength 0 32766: from all lookup main 32767: from all lookup default (ffnet is table 42) root@des1 ~ # ip r s default via 5.9.86.151 dev eth0 5.9.86.151 dev eth0 proto kernel scope link src 5.9.86.144 root@des1 ~ # ip r s t 42 blackhole default I have the default routing table, and a routing table number 42. I could use an ip rule filtering by destination ip, but I wanted to try suppress_prefixlength. Let's say I want to ping 8.8.8.8. What I expect is, that the package is put into routing table 42 by the ip rule 32765. As there is no more specific route for 8.8.8.8 than the default route in table 42, I expect the suppress_prefixlength 0 option to put it back to the default routing table and then to be send out through eth0. Instead this configuration takes the whole machine offline: root@des1 ~ # ping 8.8.8.8 connect: Invalid argument When I delete the ip rule 32765 containing the suppress_prefixlength, the machine is back online. Do I not understand the suppress_prefixlength-feature correctly or is this a bug? I tested with Kernel 4.7 and 4.6, both show the same behaviour as described above. Thanks for any replies in advance. Regards, Matthias

2 2

Re: Kernelnewbies Digest, Vol 71, Issue 26
by Laurence Rochfort Oct. 27, 2016

Oct. 27, 2016

The BBC Micro which was 6502 based had an add on that let you connect a Z80 straight onto the same bus. Different regards "operating system" I know, but it has been tried before. On Fri, 21 Oct 2016, 17:00 , <kernelnewbies-request(a)kernelnewbies.org> wrote: > Send Kernelnewbies mailing list submissions to > kernelnewbies(a)kernelnewbies.org > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies > or, via email, send a message with subject or body 'help' to > kernelnewbies-request(a)kernelnewbies.org > > You can reach the person managing the list at > kernelnewbies-owner(a)kernelnewbies.org > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Kernelnewbies digest..." > > > Today's Topics: > > 1. Is that possible to implement a single machine with > heterogeneous architecture. (Douglas Su) > 2. Re: Is that possible to implement a single machine with > heterogeneous architecture. (Felix Bytow) > 3. Re: Is that possible to implement a single machine with > heterogeneous architecture. (Valdis.Kletnieks(a)vt.edu) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Fri, 21 Oct 2016 03:10:28 +0000 > From: Douglas Su <d0u9.su(a)outlook.com> > Subject: Is that possible to implement a single machine with > heterogeneous architecture. > To: kernelnewbies <kernelnewbies(a)kernelnewbies.org> > Message-ID: > < > SN1PR04MB2096827D49CA8F766DA1392EECD40(a)SN1PR04MB2096.namprd04.prod.outlook.com > > > > Content-Type: text/plain; charset="iso-8859-1" > > Hello everyone, > > > Is that possible to install multiple CPUs which have different > architecture on a single machine? For example, on a single machine with two > different cpu sockets for X86 and MIPS cpu respectively, and these two cpus > are inter-connected with some sorts of bus (PCI or other advanced buses). > > > If it is possible, Is this machine still SMP? What will lscpu (or cat > /proc/cpuinfo) dump? > > > Thx! >

4 4

NAT with unique egress port
by Rui Santos Oct. 27, 2016

Oct. 27, 2016

Hi all, I'm currently using NAT to provide Basic address translation from private to public IP's. However, linux kernel uses both destination IP and Port as part of it's NAT mapping process. This way (client1 and client2 are on the same internal network): - if client1 connects to server1 using source port X, the NAT will be mapped: client1IP:SourcePortX -> server1IP:SourcePortX; - if client2 then connects to server2 using source port X, the NAT will be mapped: client2IP:SourcePortX -> server2IP:SourcePortX. Basically, SourcePortX is used on both mappings for client1 and client2. But, if client2 tries to connect do server1 instead, using the same source port X, the NAT will be mapped: client2IP:SourcePortX -> server2IP:SourcePortY SourcePortY will be an available (randomly generated?) ephemeral port. My goal is to force this behavior on all outgoing connections. This way I would get a unique egress port mapping to an internal IP:Port in a specific point in time: - if client1 connects to server1 using source port X, the NAT will be mapped: client1IP:SourcePortX -> server1IP:SourcePortX; - if client2 then connects to server2 using source port X, the NAT will be mapped: client2IP:SourcePortX -> server2IP:SourcePortY. SourcePortY will be an available (randomly generated?) ephemeral port. I am aware that this will imply a concurrent NAT connections limit, equal to the ephemeral port range, per egress IP. Is there any way I can accomplish this kind of behaviour? Thanks for all your help, -- Rui Santos Veni, Vidi, Linux

2 2

Latest kernel documentation tree
by Jai Krishna Oct. 26, 2016

Oct. 26, 2016

Hello, Im trying to follow the documentation work of converting to Sphinx etc that Jon, Mauro and others are doing ( inferring from the patches coming into kernel-docs@ ). Id like to be able to pull and build the Sphinx based docs, learn and possibly even contribute (wish, wish...). However, Im stuck on a very mundane thing. Im not quite sure which tree has the latest sources. I looked at kernel/kernel-docs.git on git.kernel.org but that seems stale. Im not quite sure if its kernel/gregkh/staging.git. Can someone kindly point me to the correct tree that has the latest sources of Sphinx docs? Thanks Jai

3 4

Stale pages: FAQ/VariousKernelTrees OutreachyfirstpatchSetup
by Duncan Oct. 26, 2016

Oct. 26, 2016

[I'm not subscribed. Please CC me on followups.] Two points: 1) The FAQ/VariousKernelTrees page is embarrassingly outdated with a last update in 2007 and still referring to 2.6.* as current, while both it and 3.x are arguably now historical status. 2) I actually followed a link to kernelnewbies from kernel.org, while looking for current recommendations on /secure/ git remote URLs -- it occurred to me that I really should replace my current git:// fetches with something more secure, and I wondered what was current best-practice. While rather less technical than most of the material covered here, it seems neither kernel.org nor kernelnewbies.org has an immediately obvious (including in the FAQ, at least that's immediately obvious) link to a discussion of something so basic. kernel.org does list https://git.kernel.org , but there's no real discussion or recommendation of https vs. less secure protocols, pointing out that git:// (apparently) isn't secure or indication of whether the obvious gits:// will or won't work, and no indication that the just as prominently listed rsync:// url (apparently) isn't secure either, let alone any distinction in terms of what's available, releases vs live git tree, between the three listings there (http/git/rsync). kernelnewbies.org does appear to have some basic instructions on the OutreachyfirstpatchSetup page, but there's three problems with that: a) Outreachy isn't immediately obvious as where one should look for this. I thought the front page outreachy link was to some organizational blurb (for all I know it is as I followed a more convoluted path to the above page), not a howto, and in terms of the patches part, I already know how to apply patches but am more personal systems admin and git kernel builder/tester/bug-reporter than coder so don't really generate patches for more than private use, and simply wanted information on updating my git pull URLs to something more secure. b) The staging tree might not be the best example for true kernel git sources newbies, mainline is arguably a better one. c) The example again appears dated and insecure, using the git:// protocol that the git-fetch manpage's GIT URLS discussion specifically warns should be used with caution on unsecured networks like the general internet, because it does no authentication and is /not/ secured. FWIW, I'm going with the https://git.kernel.org link for now, but thought it worth the trouble to at least post this as someone who can do something about it definitely needs to update (or delete if not considered with the trouble to update) those two pages at least, and possibly provide a more obvious basic mainline kernel git setup link on the front page, as well. Because right now they're just adding to the huge pile of outdated and now actively security-dangerous advice about Linux out there on the net. -- Duncan - No HTML messages please; they are filtered as spam. "Every nonfree program has a lord, a master -- and if you use the program, he is your master." Richard Stallman

2 1