- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

How to accout max_rss precisely
by 杨贺然 June 3, 2024

June 3, 2024

Hi, I'm working on a utility to report the maximum memory usage of a certain program. I intend to show the max_rss of this process and its children, but it's hard to use getrusage to get the RSS precisely, so I'm trying to leverage cgroups. However on the document of cgroups v2, `memory.peak` says: The max memory usage recorded for the cgroup and its descendants since the creation of the cgroup. It didn't explain the composition of memory usage. So I want to ask that is the memory usage in `memory.peak` the maximum of RSS? I made a quick test on an empty program like: ```c // a.c int main() {} ``` It shows that `memory.peak` of this program is ~500KiB, which does not make sense to me. I also wonder how to implement the same function under OS without cgroups, like BSD and macOS. Thanks

2 1

Memory reclaim mechanisms freezing system
by Yeongjin Kwon June 3, 2024

June 3, 2024

Hello, I have been using GNU/Linux as my daily driver for 3 years. During that whole time, I have experienced this problem where my system would often freeze up and become unresponsive for a few seconds here and there when it is under load, which has been a big pain. After experimenting and trying to find the cause of the problem, I now suspect the root cause is that the kernel memory reclaim mechanisms (kswapd, direct reclaim, etc.) use too much CPU time, not allowing my desktop environment processes to run. My computer is a Surface Pro (1st Gen). Here is a link to its specifications: https://support.microsoft.com/en-us/surface/surface-pro-1st-gen-specificati… Notably, my computer has 4GB of ram and 8 GB of swap. I am currently using the sway tiling window manager on arch linux, but have also experienced the freezing problem on KDE plasma, gnome, and other linux distros. There is some configuration I changed from the defaults to improve performance. I turned zswap off, and set the vm.page-cluster sysctl to 3 (which might have been the default anyways). There are several reasons I did this, but I will not expand on them here. My system generally freezes when it has low free memory, is using swap, and I switch to an application that has not been recently active. The problem gets worse when more swap is being used. When my system is using around 3-4 GB or more of swap, it even stalls for tens of seconds. And so I believe the problem occurs when my system is swapping a lot. I thought the freezing was due to my desktop processes getting swapped out, so I allocated memory, cpu, and io bandwidth protections using cgroups and systemd to my desktop/system processes so that they would not get swapped out and that they would have enough resources even under load. This significantly helped solve the freezing, but the problem still remained. As I understand it, there are two procedures involved in swapping when free memory is low, which are reading data into memory from swap when page faults occur, and swapping old data out of memory onto disk (reclaiming memory) to make room for the data being swapped in. Furthermore, I have observed the system freezing when it exhausts available ram and starts swapping memory out onto disk, even when there is no swap usage at the time. This leads me to believe the freezing is caused by the latter of the two procedures, memory reclamation. In addition, I did not observe the problem occurring when the system was swapping data in from disk while there was plenty of free memory. I also recently tried increasing the vm.watermark_scale_factor sysctl parameter. When I did, the freezing problem became worse. According to the documentation, this sysctl parameter is used to tune the aggressiveness of kswapd. Assuming kswapd is more aggressive and active the higher the parameter is, the amount of activity of kswapd would be positively correlated with the severity of the problem. I also tried using the linux zen kernel, but that did not seem to alleviate the problem, and I believe it actually made performance worse. And so now I think it's possible that the kernel memory reclaim mechanisms are using up too much CPU and ignoring the cpu protections (cpu.weight parameter) that I set for my desktop environment processes. It explains why my system freezes even though I allocated protections to my desktop environment. The kernel can just bypass them all. I know all the details I mentioned above may be vague and unreliable as evidence. I suppose it might be faster and more accurate to do some profiling, but I unfortunately do not know how to go about profiling my system. Is my suspicion correct? If so, is there any way to restrict the CPU time these reclaim mechanisms get? Will doing that be likely to improve performance? Perhaps this occurs because the reclaim mechanisms do not yield the CPU at all until the respective watermark is reached. If this is the case, is anyone working to make these mechanisms more preemptible and less CPU hogging? I would love to try implementing these changes myself, but at the moment, I unfortunately have no experience in kernel hacking. I thought it might be a good idea to post a question and see if anyone else has some more insight into the issue, before attempting further experimentation. Thank you for your time. Yeongjin Kwon

1 0

notifying clock consumers of frequency change
by Sam Winchenbach May 30, 2024

May 30, 2024

Hello, I am writing a driver for a frequency synthesizer that does not currently have any support in the kernel. This synthesizer is clocked by a fixed external oscillator. In my driver, I: 1. Register the clock hardware with devm_clk_hw_register 2. Describe the clocks to the DTS nodes with devm_of_clk_add_hw_provider Various external stimuli and commands (through IIO) can change the frequency of this synthesizer. How do I notify downstream devices that the frequency has changed? Ultimately it looks like __clk_notify() is what I want to call, but that is a non-exported function. Tracking it through, the function stack that calls this looks like it is used to request a new frequency, not to notify consumers that the frequency has changed. Does anyone have any tips, or perhaps know of an in-tree driver I could study to understand this better? Thanks. -Sam

1 0

Which driver to use for patches?
by Philipp Hortmann May 18, 2024

May 18, 2024

Dear knewbies and developers, maybe you ask yourself: Which driver to use for patches? Here the list of drivers I can support you with tests (from the staging subsystem): rtl8723bs rtl8192e rtl8712 vt6655 vt6656 Please feel free to contact me for questions. Thanks for your support. Bye Philipp Hint: You like to get your own hardware? Here some hints where I got mine from: rtl8723bs This hardware can be bought directly. https://www.ebay.de/sch/i.html?_from=R40&_nkw=rtl8723bs&_sacat=0 But since it has a SDIO Interface it is not so easy to connect it. Luckily I found out that the Laptop ODYS TRENDBOOK NEXT 14 (not so easy to install linux on) Laptop ODYS TRENDBOOK NEXT 14 Pro do have it build in. I just got one so I am able to test. rtl8192e This hardware can be bought directly. https://www.ebay.de/sch/i.html?_from=R40&_nkw=rtl8192e&_sacat=0&_odkw=rtl87… But you need a mini PCIe Card slot to connect it to your computer. rtl8712 I found this hardware some time ago. It is a USB Stick brand packard bell Part Number: AW-NU120 vt6655 I found this hardware long time ago. It is a mini PCI card. It seems that it is build in a One A400 Notebook. But I do not have this Notebook. https://forum.ubuntuusers.de/topic/via-vt-8739/ vt6656 I found this hardware long time ago. It is a USB Dongle. It was build into video cameras. That is the reason why I found it and was able to buy it.

1 0

Seeking Assistance with Spin Lock Usage and Resolving Hard LOCKUP Error
by Muni Sekhar May 17, 2024

May 17, 2024

Dear Linux Kernel Community, I am reaching out to seek assistance regarding the usage of spin locks in the Linux kernel and to address a recurring issue related to hard LOCKUP errors that I have encountered during testing. Recently, I developed a small kernel module that involves ISR handling and utilizes the spinlock_t primitive. In my module, I have employed spin locks both in process context using spin_lock() and spin_unlock() APIs, as well as in ISR context using spin_lock_irqsave() and spin_unlock_irqrestore() APIs. Here is a brief overview of how I have implemented spin locks in my module: spinlock_t my_spinlock; // Declare a spin lock // In ISR context (interrupt handler): spin_lock_irqsave(&my_spinlock, flags); // ... Critical section ... spin_unlock_irqrestore(&my_spinlock, flags); // In process context: (struct file_operations.read) spin_lock(&my_spinlock); // ... Critical section ... spin_unlock(&my_spinlock); However, during testing, I have encountered a scenario where a hard LOCKUP (NMI watchdog: Watchdog detected hard LOCKUP on cpu 2) error occurs, specifically when a process context code execution triggers the spin_lock() function and is preempted by an interrupt that enters the ISR context and encounters the spin_lock_irqsave() function. This situation leads to the CPU being stuck indefinitely. My primary concern is to understand the appropriate usage of spin locks in both process and ISR contexts to avoid such hard LOCKUP errors. I am seeking clarification on the following points: Is it safe to use spin_lock_irqsave() and spin_unlock_irqrestore() APIs in ISR context and spin_lock() and spin_unlock() APIs in process context simultaneously? In scenarios where a process context code execution is preempted by an interrupt and enters ISR context, how should spin locks be used to prevent hard LOCKUP errors? Are there any specific guidelines or best practices for using spin locks in scenarios involving both process and ISR contexts? I would greatly appreciate any insights, guidance, or suggestions from the experienced members of the Linux kernel community to help address this issue and ensure the correct and efficient usage of spin locks in my kernel module. Thank you very much for your time and assistance. -- Thanks, Sekhar

2 1

Handling I2C eeprom devices
by Patryk May 17, 2024

May 17, 2024

Hi, I've got a question regarding how to handle I2C eeprom devices. I see that in the kernel/drivers/misc/eeprom there are not that many drivers available. Does it mean that they support various chips that do not necessarily come from one vendor? Looking at example at24.c, in the compatible properties I see that all of them refers to atmel, however when I compared the following devices: AT24C04C and BRCB008GWZ, at the first glance they look kinda similar: - they are organized internally in the same way - 1024 x 8 (8K) - read/write commands looks the same I'm not asking you to tell me whether or not it's possible this driver at24 with the device I need (BRCB008GWZ), rather I would like to get a suggestion on how to approach this and what factors I should take into account to make the right decision. Best regards Patryk

3 2

[RFC PATCH] kvm: replace spaces with a tab
by Mohamed Karaoui May 16, 2024

May 16, 2024

Hi list, This is an attempt to get a feeling of sending a first patch. Please feel free to comment. I have few questions: - I have build this patch against the torvalds/linux.git. Does it matter which tree I use as long as it works on the maintainer tree? - In this particular case I checked the maintainer. It seems that its tree is something of 11 years old. - Should I just go ahead and send this patch to the emails in ./scripts/get_maintainer.pl anyway? (replacing this message of course) Signed-off-by: Mohamed Karaoui <moharaka(a)gmail.com> --- virt/kvm/kvm_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index ff0a20565f90..cd4702e069ca 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -6368,7 +6368,7 @@ EXPORT_SYMBOL_GPL(kvm_get_running_vcpu); */ struct kvm_vcpu * __percpu *kvm_get_running_vcpus(void) { - return &kvm_running_vcpu; + return &kvm_running_vcpu; } #ifdef CONFIG_GUEST_PERF_EVENTS -- 2.34.1

2 1

Looking for advise
by Jules Irenge May 15, 2024

May 15, 2024

I am looking for advice on what to do after reading the Linux device driver book. Someone suggests I get a beaglebone and write some i2c driver to get hands-on experience, I will consider that. Any other suggestion will be appreciated Thanks, Jules *..* Open source contributor: https://bit.ly/40GblaQ Blog: https:// <https://www.strava.com/athletes/72390635> www.julesirenge.blog twitter: @jules_irenge strava: https://www.strava.com/athletes/72390635 linkedIn: *https://www.linkedin.com/in/jules-irenge-87282238/ <https://www.linkedin.com/in/jules-irenge-87282238/>*

2 1

Fwd: Qemu hangs at Starting KernelMemorySanitizer
by Camila Alvarez May 14, 2024

May 14, 2024

Hello all, I'm trying to debug a KMSAN related issue. To do so I created a QEMU instance to replicate the bug, however, it gets stuck at startup. It gets stuck after saying: Starting KernelMemorySanitizer ATTENTION: KMSAN is a debugging tool! Do not use on production machines! Has anyone experienced something similar? Do you have any idea of which configuration options I should look into? To give some context: I'm working on an ARM64 machine, and I'm cross compiling the kernel for x86_64 using clang. Thanks!

1 0

Seeking Assistance: Implementing USB Device Suspend/Resume in User Space Driver
by Muni Sekhar May 13, 2024

May 13, 2024

Dear Linux USB Subsystem Community, I hope this email finds you well. We are currently working on a USB device driver in the Linux kernel space, which supports the suspend/resume functionality for a USB device using the struct usb_driver's .suspend and .resume callbacks. In our driver code, we perform a check to determine if the device has remote wakeup capability. If the device supports remote wakeup, we enable the device's autosuspend mode by setting the struct usb_device->dev.power.autosuspend_delay parameter and then calling usb_enable_autosuspend(). Similarly, to resume the suspended device, we expose an API to user space via IOCTL. In this process, we set struct usb_device->dev.power.autosuspend_delay to -1 and then call usb_disable_autosuspend(). While we have successfully converted most of the USB device driver functionality from kernel space to user space using libusb, we are now seeking guidance on how to implement USB device suspend/resume in user space. Specifically, we would like to know if libusb provides support for achieving the suspend/resume functionality mentioned above. Additionally, any insights or suggestions on how to accomplish this task would be greatly appreciated. It's worth mentioning that our USB device driver also supports Bulk\Coontrol I/O transfers and other functionalities. Thank you in advance for any assistance you can provide. I look forward to your valuable inputs and recommendations. -- Thanks, Sekhar

2 12