September 2017 - 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

Re: How to verify linux-next
by Pintu Kumar Oct. 1, 2017

Oct. 1, 2017

On Fri, Sep 29, 2017 at 6:44 PM, Damian Tometzki <damian.tometzki(a)icloud.com> wrote: > Hello, > > Ubuntu 16.04 with current linux-next Kernel workson virtualbox 5.1.28 > > Host: Windows 10 > Guest: Ubuntu 16.04 > > Best regards > Damian > > > Am Freitag, den 29.09.2017, 08:41 -0400 schrieb > valdis.kletnieks(a)vt.edu: >> On Fri, 29 Sep 2017 16:08:07 +0530, Pintu Kumar said: >> >> > >> > I have a general question. >> > How do we normally verify linux-next tree? >> The same exact way you "verify" any other Linux kernel, for whatever >> definition of "verify" you plan to use. >> >> > >> > 1) For Oracle virtual box 5.1.26 with ubuntu-32 bit, the linux-next >> > kernel is not booting. >> Does an Ubuntu kernel boot correctly under VirtualBox? If not, fix >> that issue first. Also, "is not booting" isn't detailed enough for >> anybody >> to make even a guess as to what's wrong. >> >> Also, note that 5.1.28 is out. Ok, I just updated to 5.1.28. And my Ubuntu version is already 16.04. Let me try again if it works. Thanks all for your reply. BTW, I am more interested in my another query about QEMU arm. This will be much quicker and easy for me. But the problem is I wanted to use multiple ssh shell on qemu. Also I needed a pre-built rootfs image for qemu-arm, cortex-a9 versatilepb machine. It should have networking and ssh built-in so that I can connect to it using PUTTY client. Note that I could able to build my own minimal busybox and boot qemu using linux-next (non graphical mode). But I could able to get only one shell. I need to test some client/server problem, so I need multiple shell. I am not able to configure net/ssh on this qemu system. So, I have 2 things to ask: 1) If you have pointers on how to setup ssh/net connection on QEMU with busybox, do let me know. 2) Let, please point me to a pre-built qemu-arm busy box image with full features. Thanks, Pintu

7 12

Input device driver
by Bruno E. O. Meneguele Sept. 30, 2017

Sept. 30, 2017

Hi folks, I'm trying to write an input device driver basicaly to play with interrupt handlers. The idea at first was to catch all keys pressed on my keyboard, nothing else, but some doubts rised after I wrote the driver and it didn't work out. The code so far is as follows: https://gist.github.com/bmeneguele/621d4bbbfa28fca200df6ba289cfb3fc Now the questions: 1) from /proc/interrupts I saw the i8042 keyboard+mouse controller on IRQ line 1 and 12: CPU0 CPU1 CPU2 CPU3 0: 55 0 0 0 IR-IO-APIC 2-edge timer 1: 101 5 7 2 IR-IO-APIC 1-edge i8042 8: 0 0 0 1 IR-IO-APIC 8-edge rtc0 9: 7216 3565 16517 3485 IR-IO-APIC 9-fasteoi acpi 12: 212 29 323 51 IR-IO-APIC 12-edge i8042 16: 0 0 0 0 IR-IO-APIC 16-fasteoi i801_smbus 19: 15 0 6 3 IR-IO-APIC 19-fasteoi [...] then I thought I could share the line along with it, once in drivers/input/serio/i8042.c the request_irq() is being called with IRQF_SHARED. Is there a problem here? 2) I'm using a USB keyboard as the testing device, and TBH I got confused if I could actually use the input subsystem for that or I _should_ use HID instead (considering the keyboard is HID compliant). 3) the output I'm receiving is just one message regardless the number of keys I press: [... (make && insmod) ...] $ dmesg | tail [75013.625311] input: Bmeneg's Keyboard as /devices/virtual/input/input25 [75056.682676] [my_kbd] kbd_irq_handler:22:: irq 12 occured! [... (rmmod && make && insmod) ...] [75132.203324] input: Bmeneg's Keyboard as /devices/virtual/input/input26 [75149.669112] [my_kbd] kbd_irq_handler:22:: irq 1 occured! [... (rmmod && make && insmod) ...] [75164.213562] input: Bmeneg's Keyboard as /devices/virtual/input/input27 [75168.720791] [my_kbd] kbd_irq_handler:22:: irq 1 occured! [... (rmmod && make && insmod) ...] [75390.584746] input: Bmeneg's Keyboard as /devices/virtual/input/input28 [75397.149047] [my_kbd] kbd_exit:73:: irq reference counter: 1 [... (rmmod && make && insmod) ...] [79146.583034] input: Bmeneg's Keyboard as /devices/virtual/input/input29 [79159.647871] [my_kbd] kbd_exit:73:: irq reference counter: 1 as you can see I tried with prints inside IRQ handler, that I changed to a ref counter considering a previous knowledge from embedded systems that prints inside IRQs are bad :D (is it still true for linux kernel?), and the output from prints (inside IRQs) just appears after I `rmmod` the driver. Of course something is wrong in my code, but what? (handle kbd as input device? IRQ shared with i8042? wrong way to handle IRQ?) Well, this is my first time trying to accomplish something using real hardware in kernel space, not using just tutorial copy/paste or software-only things like timers. Because of that any info would be really great. What I'm trying to get here is just dive into bottom-helves execution (work queues) to understand how they actually works and of course more advanced things later! Thanks in advance! Note: using kernel 4.12.12 here. -- bmeneg PGP Key: http://bmeneg.com/pubkey.txt

3 4

keyboard behavior change
by John de la Garza Sept. 30, 2017

Sept. 30, 2017

I have changed drivers/tty/vt/keyboard.c to allow me to use my return key (or any other key) as a control key if pressed with another key or simply as the return key when pressed and released. This only works in virtual consoles. I would like to figure out where to move this code to make it work all the way up, so it will work in virtual consoles, X11 apps, and for all keyboards. I want to make it appear that 'keyboards' just work this way. Anyone have any feedback to where to do something like this? I have considered doing it with X, but I also would like it to work in virtual consoles as well as X.

1 0

Re: How to verify linux-next
by Pintu Kumar Sept. 29, 2017

Sept. 29, 2017

Hi, I have a general question. How do we normally verify linux-next tree? I wanted to work on linux-next but I am facing some issues. I could able to build linux-next for both x86 and arm, but I could not verify it on any machine. Currently I don't have a real Linux PC to boot with linux-next kernel. So I am trying to find alternative ways, like using the virtual box or qemu-arm. 1) For Oracle virtual box 5.1.26 with ubuntu-32 bit, the linux-next kernel is not booting. 2) For qemu-arm (versatilepb), I am able to build the kernel, but I could not figure out which rootfs to use with it. I tried creating minimal rootfs using busybox, but it does not contain enough interface. I am not able to open multiple terminal and also could not setup ssh to access it using PUTTY. So, if you know of any better rootfs to use with qemu-arm please let me know. Or, if you know of any better option to use linux-next please tell me. It will be really helpful. Thank You! Regards, Pintu

3 3

Re: How to verify linux-next
by Pintu Kumar Sept. 29, 2017

Sept. 29, 2017

** Re sending ** Hi, I have a general question. How do we normally verify linux-next tree? I wanted to work on linux-next but I am facing some issues. I could able to build linux-next for both x86 and arm, but I could not verify it on any machine. Currently I don't have a real Linux PC to boot with linux-next kernel. So I am trying to find alternative ways, like using the virtual box or qemu-arm. 1) For Oracle virtual box 5.1.26 with ubuntu-32 bit, the linux-next kernel is not booting. 2) For qemu-arm (versatilepb), I am able to build the kernel, but I could not figure out which rootfs to use with it. I tried creating minimal rootfs using busybox, but it does not contain enough interface. I am not able to open multiple terminal and also could not setup ssh to access it using PUTTY. So, if you know of any better rootfs to use with qemu-arm please let me know. Or, if you know of any better option to use linux-next please tell me. It will be really helpful. Thank You! Regards, Pintu On Fri, Sep 29, 2017 at 3:12 PM, Pintu Kumar <pintu.ping(a)gmail.com> wrote: > Hi, > > I have a general question. > How do we normally verify linux-next tree? > > I wanted to work on linux-next but I am facing some issues. > I could able to build linux-next for both x86 and arm, but I could not > verify it on any machine. > Currently I don't have a real Linux PC to boot with linux-next kernel. > So I am trying to find alternative ways, like using the virtual box or qemu-arm. > 1) For Oracle virtual box 5.1.26 with ubuntu-32 bit, the linux-next > kernel is not booting. > 2) For qemu-arm (versatilepb), I am able to build the kernel, but I > could not figure out which rootfs to use with it. > I tried creating minimal rootfs using busybox, but it does not contain > enough interface. I am not able to open multiple terminal and also > could not setup ssh to access it using PUTTY. > > So, if you know of any better rootfs to use with qemu-arm please let me know. > Or, if you know of any better option to use linux-next please tell me. > It will be really helpful. > > > Thank You! > Regards, > Pintu

1 0

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

Sept. 28, 2017

On Thu, Sep 28, 2017 at 11:51 AM, Divakar <divakar.chitturi(a)hpe.com> wrote: > Hi Mandeep, > Recently i encountered similar issue with UIO framework in hotplug > scenario. I reviewed the earlier patch from 2010 and your 4 patch > series. 4 patch series seems to work. I checked the latest kernel > source and i dont think your patches have been merged. Hopefully > i am not looking at the wrong place. If they are not merged is there > a plan to merge them ? I do think these patches are required. > > Thats correct. The patches have not been merged. I no longer have the hardware available and have moved away from the project that needed it, therefore I was unable to carry with the patches (they were still under review). It'd be great if you can resubmit those. Make sure they apply cleanly to the latest staging branch (?). I think Greg had asked some question regarding during the review, it would be good if you can pick it up from there. GKH (who's on this list too) can weigh in more on this. Good luck. -mandeep > I also found a bug in your patch set which is discussed here > https://lkml.org/lkml/2017/6/14/30. I was able to workaround/fix it. > I can upload a patch if needed. > > Please advice if you plan to get the patchset merged. > > Thanks > Divakar >

1 0

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

iovector changes in kernel versions from 3.19
by Sreejith M M Sept. 27, 2017

Sept. 27, 2017

I have sendmsg function which sends messages using iovec array. I am not understanding the new way of getting the length and getting the data from msghdr from the sendmsg api. Has the sendmsg api signature changed recently in kernel? Where do I find more iovec apis? -- Regards, Sreejith

1 0

Accessing registers from another driver
by Luca Ellero Sept. 27, 2017

Sept. 27, 2017

Hi, in a watchdog driver on an ARM platform, I have the need to write a hw register managed by another driver (system reset controller). I'm wondering what is the correct/clean way of doing that. What I've done until now is adding a second couple in the "reg" property of the device-tree &wdog1 { pinctrl-names = "default"; pinctrl-0 = <&pinctrl_wdog>; reg = <0x30280000 0x10000>, <0x30390000 0x10000>; fsl,wdog_b; }; and then in the wdog driver adding the following code: static void __iomem *src_base; ... res = platform_get_resource(pdev, IORESOURCE_MEM, 1); src_base = devm_ioremap(&pdev->dev, res->start, res->end - res->start + 1); if (IS_ERR(src_base)) return PTR_ERR(src_base); ... writel(0x3, src_base + 4); In this way it works as expected, but I'm not sure if this is correct? Thanks Best regards Luca

1 0