December 2023 - kernelnewbies@kernelnewbies.org

Source code organization
by Billie Alsup (balsup) April 23, 2024

April 23, 2024

I have a set of drivers that I would like to upstream. These are primarily MFD style drivers supporting Cisco-8000 FPGAs. The devices can be instantiated through multiple top level drivers, which provide the access mechanism for the MFD driver. For example, an FPGA can be accessed via PCI, or via i2c, or via SLPC, or via P2PM (a point-to-point interface). We currently build these drivers out of tree, under a single directory drivers/cisco. I note that existing drivers are spread out across the kernel tree, sometimes by bus (pci, i2c), sometimes by function (gpio, net/mdio, spi), and sometimes under the generic mfd. I would like to start the upstream review process for our drivers, but first want to get recommendations on the source code layout. Is it permissible to keep a top level directory such as drivers/cisco to organize our code? It is not only the source code that is affected, but also provides a central place for Kconfig entries. Our FPGAs have multiple logical blocks, each of which is handled by a different MFD driver, e.g. i2c controllers, gpio controllers, spi controllers, mdio controllers. There can be multiple instances of each block as well (so multiple MFD devices are instantiated for each driver). And of course, there can be multiple FPGAs in a system, each with different combinations of logical blocks. The drivers themselves are pretty specific to our FPGAs, thus it makes sense to use Kconfig to select a hardware platform to automatically select the set of MFD drivers (and top level bus drivers) that would apply. Would a source layout putting all the code under drivers/cisco be acceptable in this case, or do I need to move things around and spread out the Kconfig entries across directories? I note that a single drivers/cisco would simplify any related modifications to MAINTAINERS as well.

2 5

Boot stucks at "Loading initial ramdisk"
by Dileep Sankhla Jan. 5, 2024

Jan. 5, 2024

Hello, I have cloned and built the latest rc kernel using the default config but on boot, it stucks at "Loading initial ramdisk". I have tried booting after removing the "quiet" option from the kernel parameters just before selecting the kernel from the grub menu. I have also tried adding "dis_ucode_ldr" parameter to the initial parameters but all in vain. How to look for the underlying issue and solve it? Regards, Dileep

5 12

mtd0 device is missing in /sys/class/mtd
by Manojkiran Eda Dec. 29, 2023

Dec. 29, 2023

Hi All, I was as beginner in linux device driver development, and recently I was working on a device driver for espi protocol and was stuck with a problem. I have an mtd device driver through which one could access the NOR flash connected to the other end of espi bus. The driver seems to have probed successfully , I see an entry for my mtd device in /proc/mtd but , I dont see an entry for my mtd device in the /sys/class/mtd. Here are few logs of interest : root@mybox:~# cat /proc/mtd dev: size erasesize name mtd0: 04000000 00001000 "aspeed-espi-flash” <<<<<—— This is my mtd device mtd1: 04000000 00010000 "bmc" mtd2: 000e0000 00010000 "u-boot" mtd3: 00020000 00010000 "u-boot-env" mtd4: 00900000 00010000 "kernel" mtd5: 02000000 00010000 "rofs" mtd6: 01600000 00010000 "rwfs" mtd7: 04000000 00010000 "alt-bmc" mtd8: 000e0000 00010000 "alt-u-boot" mtd9: 00020000 00010000 "alt-u-boot-env" mtd10: 00900000 00010000 "alt-kernel" mtd11: 02000000 00010000 "alt-rofs" mtd12: 01600000 00010000 "alt-rwfs” root@mybox:/sys/class/mtd# ls mtd1 mtd10ro mtd11ro mtd12ro mtd2 mtd3 mtd4 mtd5 mtd6 mtd7 mtd8 mtd9 mtd10 mtd11 mtd12 mtd1ro mtd2ro mtd3ro mtd4ro mtd5ro mtd6ro mtd7ro mtd8ro mtd9ro I was expecting mtd0 to be present in the /sys/class/mtd, but its not there. Looking at dmesg: root@mybox:/sys/class/mtd# dmesg | grep -ie mtd -A5 -ie espi [ 0.685006] Entered the aspeed_espi_ctrl_probe function [ 0.690990] device matched : 1 [ 0.699225] espi alloc started [ 0.704726] espi read from device tree, flash size : 67108864 [ 0.712416] espi flash enabled [ 0.716572] espi flash channel is ready [ 0.720940] espi flash allocation completed [ 0.725259] aspeed-espi-ctrl 1e6ee000.espi-ctrl: module loaded -- [ 0.962764] 5 fixed-partitions partitions found on MTD device bmc [ 0.969675] Creating 5 MTD partitions on "bmc": [ 0.974776] 0x000000000000-0x0000000e0000 : "u-boot" [ 0.981982] 0x0000000e0000-0x000000100000 : "u-boot-env" [ 0.989557] 0x000000100000-0x000000a00000 : "kernel" [ 0.996772] 0x000000a00000-0x000002a00000 : "rofs" [ 1.003759] 0x000002a00000-0x000004000000 : "rwfs" -- [ 1.048304] 5 fixed-partitions partitions found on MTD device alt-bmc [ 1.055563] Creating 5 MTD partitions on "alt-bmc": [ 1.061096] 0x000000000000-0x0000000e0000 : "alt-u-boot" [ 1.068672] 0x0000000e0000-0x000000100000 : "alt-u-boot-env" [ 1.076607] 0x000000100000-0x000000a00000 : "alt-kernel" [ 1.084211] 0x000000a00000-0x000002a00000 : "alt-rofs" [ 1.091594] 0x000002a00000-0x000004000000 : "alt-rwfs” and the mtd devices listed in /sys/class/mtd from mtd1 through mtd12 are basically from the partitions list above. root@mybox:/sys/class/mtd# for dir in *; do cat $dir/name; done bmc alt-kernel alt-rofs alt-rwfs u-boot u-boot-env kernel rofs rwfs alt-bmc alt-u-boot alt-u-boot-env Does anyone know what could be the possible reason for the problem ? Any hints/help would be appreciated. Thanks, Manoj

1 0

why high io requests will slow down kernel scheduling?
by kipade Dec. 16, 2023

Dec. 16, 2023

Hello, Im puzzled about the influence about iowait. Almost all the document say, the iowait "Show the percentage of time that the CPU or CPUs were idle during which the system had an outstanding disk I/O request." And, if there was a programs writing disk frequently, the system's iowait would be very high and the cpu was very free(do nothing). If so, the kernel's process scheduling should work fine, but the process scheduling lantency was very high(some programs were rescheduled after a long time). Why? Thanks

1 0

Insight into Load, Runnable, and Utilization Averages in Linux Scheduler Code
by iso m Dec. 15, 2023

Dec. 15, 2023

Hello there, Below snippet is taken from file: include/linux/sched.h >/* > * The load/runnable/util_avg accumulates an infinite geometric series > * (see __update_load_avg_cfs_rq() in kernel/sched/pelt.c). > * > * [load_avg definition] > * > * load_avg = runnable% * scale_load_down(load) > * > * [runnable_avg definition] > * > * runnable_avg = runnable% * SCHED_CAPACITY_SCALE > * > * [util_avg definition] > * > * util_avg = running% * SCHED_CAPACITY_SCALE > * > * where runnable% is the time ratio that a sched_entity is runnable and > * running% the time ratio that a sched_entity is running. what's the denominator that's being considered here? Is it a period? or time between two scheduler ticks? > * > * For cfs_rq, they are the aggregated values of all runnable and blocked > * sched_entities. > * > * The load/runnable/util_avg doesn't directly factor frequency scaling and CPU > * capacity scaling. The scaling is done through the rq_clock_pelt that is used > * for computing those signals (see update_rq_clock_pelt()) > * > * N.B., the above ratios (runnable% and running%) themselves are in the > * range of [0, 1]. To do fixed point arithmetics, we therefore scale them > * to as large a range as necessary. This is for example reflected by > * util_avg's SCHED_CAPACITY_SCALE. Thanks Vishal

1 0

when/how is the schedule() function actually called?
by Dawei Li Dec. 14, 2023

Dec. 14, 2023

Greetings! I am reading several resources regarding the linux kernel scheduling (in kernel 2.6.34). There seems to be the periodic scheduler (tick_periodic()), which invokes the scheduler_tick() function, then the entity_tick() function, and then the resched_task() function. However, eventually, the resched_task() function doesn't invoke the schedule() function; it only invokes the set_tsk_need_resched() function. So, it is only setting the need_resched flag. My question is, when/how is the schedule() function actually called? Also, I notice that at many locations in the code, when there is the need to do a scheduling, the code is just setting the need_resched flag. Same question: when/how is the schedule() function actually called? (Although I am using the 2.6.34 version, I believe the question is generally applicable to any kernel version.) Not sure if this list is still active. Appreciate any response. Thanks!

4 6

Question regarding FlexSPI controller available on NXP boards
by Patryk Dec. 13, 2023

Dec. 13, 2023

Hi, I've got a question regarding FlexSPI controller available on e.g imx8mm or ls1028/2160 boards. To not waste your time here is tldr; Does spi-nxp-fspi controller (and Linux driver) support simultaneous read/write transactions? More details here: I have two boards, the first one is LS1028a, and the second one is Hilsher netX90 and I need to connect these two over SPI. The only available SPI controller I have on my board is FlexSPI. From what I've found in the documentation by using LUT sequences I should be able to adapt it to the protocol that is needed by netX90. There is just one problem: netX immediately sends some status information (1 status byte followed by 3 undefined bytes) once the master starts clocking, which results in simultaneous transfer on both MOSI and MISO lines. Eg when I want to issue a read operation on netx, I send the following byte stream:(1 byte cmd, 3 bytes addr, 1 byte length) MOSI: 0x80 | 0x12 0x34 0x56 | 0x4 MISO: 0x11 | 0xXX 0xXX 0xXX | data[0]... data[3] As you can see the status byte - 0x11 (followed by some undefined 3 bytes) is immediately sent over the MISO line. Then, data[0]...data[3] is actual data that was requested to fetch. I tried to use spi-mem framework that is supported by spi-nxp-fspi controller driver, but it does not support handling simultaneous transfer thus at the end of the SPI transaction I ended up with a rx buffer consisting of data[0]...data[3] only. Kindly asking for any advice. Unfortunately, there is no way to use a traditional SPI controller in this case, as all of them are used for other purposes. BR Patrick

1 0

from CoLinux to MoreLinux :D
by Mario Marietto Dec. 10, 2023

Dec. 10, 2023

Hello. maybe someone of you know the old project called "coLinux" : Cooperative Linux is the first working free and open source method for optimally running Linux on Microsoft Windows natively. More generally, Cooperative Linux (short-named coLinux) is a port of the Linux kernel that allows it to run cooperatively alongside another operating system on a single machine. For instance, it allows one to freely run Linux on Windows 2000/XP/Vista/7, without using a commercial PC virtualization software such as VMware, in a way which is much more optimal than using any general purpose PC virtualization software. In its current condition, it allows us to run the KNOPPIX Japanese Edition on Windows. CoLinux is very old and not maintained for a lot of time and I'm not interested in resurrecting it (and I don't have the competences to do it),BUT I'm interested to gather some information about a similar project that I have in mind. What about if,instead of having a Linux kernel which can run Windows cooperatively,we have a Linux kernel that can run more Linux distributions (maybe only 2 as a starting point,as CoLinux already does) at the same time,without using virtualization software ? Is the technology behind Colinux the same that's under the lxc or docker containers or the WSL2 subsystem ? What are the differences ? I don't use WSL2,I don't use Windows so much. I like Linux and FreeBSD. So,an even nicer idea is to create a coLinux variant that allows the Linux kernel to cooperate with FreeBSD. This is even nicer than making a cooperation between 2 Linuxes. -- Mario.

3 14

Re: Your confirmation is required to leave the Kernelnewbies mailing list
by shrikanth malawad Dec. 8, 2023

Dec. 8, 2023

Please remove from mailing list On Fri, Dec 8, 2023 at 7:10 PM shrikanth malawad < shrikanth.malawad(a)gmail.com> wrote: > Please remove from mailing list > > On Fri, Dec 8, 2023 at 7:09 PM < > kernelnewbies-confirm%2B9286dc1e48a84934f6dad5be0d38d7fa330fcef9(a)kernelnewbies.org> > wrote: > >> Mailing list removal confirmation notice for mailing list >> Kernelnewbies >> >> We have received a request for the removal of your email address, >> "shrikanth.malawad(a)gmail.com" from the kernelnewbies(a)kernelnewbies.org >> mailing list. To confirm that you want to be removed from this >> mailing list, simply reply to this message, keeping the Subject: >> header intact. Or visit this web page: >> >> >> https://lists.kernelnewbies.org/mailman/confirm/kernelnewbies/9286dc1e48a84… >> >> >> Or include the following line -- and only the following line -- in a >> message to kernelnewbies-request(a)kernelnewbies.org: >> >> confirm 9286dc1e48a84934f6dad5be0d38d7fa330fcef9 >> >> Note that simply sending a `reply' to this message should work from >> most mail readers, since that usually leaves the Subject: line in the >> right form (additional "Re:" text in the Subject: is okay). >> >> If you do not wish to be removed from this list, please simply >> disregard this message. If you think you are being maliciously >> removed from the list, or have any other questions, send them to >> kernelnewbies-owner(a)kernelnewbies.org. >> >

2 1

Linux 4 block device driver infrastructure
by Alexandru Goia Dec. 8, 2023

Dec. 8, 2023

Greetings ! I am a Unix/Linux hobbyist from Romania, interested in kernel stuff. I need some clarifications, related to block device subsystem in Linux 4, so I will ask them here, if you can answer me, please... 1) Why (in Linux 4) in struct block_device_operations, the (*open)(struct block_device *, fmode_t) has a different signature than the (*release)(struct gendisk *, fmode_t) ? Why open() uses block_device, while release() uses gendisk ? They are both in a struct block_device_operations. Why they not refer to the same thing ? I understand that gendisk refer to a real disk, and block_device (s) to logical disk (s) and partitions. But why the kernel developers have chosen to use different signatures ? 2) Release() is also synonim to close() ? 3) Why is not explicitely present a close() function ? 4) Why struct gendisk does not have inside it pointers to struct block_device ? Thank you very much, Alexander, Computer hobbyist, Romania

2 2