- kernelnewbies@kernelnewbies.org - Kernelnewbies Mailing Lists

Linux kernel mailing list filtering issues
by hal＠halemmerich.com July 20, 2018

July 20, 2018

Hi all, I've been trying to post a bug report to the linux-usb mailing list but keep getting trapped in the filters despite ensuring I'm using plaintext. Heres what I get in the bounce back: The mail system <linux-usb(a)vger.kernel.org>: host vger.kernel.org[209.132.180.67] said: 550 5.7.1 Content-Policy accept-into-freezer-1 msg: Bayes Statistical Bogofilter considers this message SPAM. BF:<S 1> In case you disagree, send the ENTIRE message plus this error message to <postmaster(a)vger.kernel.org> ; I've sent the message to the postmaster but haven't heard anything. My message got through to the maintainer, who has requested I fix my message and get it on the mailing list proper but nothing I've tried is working. Heres part of the header from the source of my most recent message: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit Any suggestions? Thanks

2 1

Re: mmap with huge page
by David Frank July 18, 2018

July 18, 2018

Thanks Mike. I read the doc, which is not explicit on the non used file taking up huge page count On Tuesday, July 17, 2018, 4:57:04 PM PDT, Mike Kravetz <mike.kravetz(a)oracle.com> wrote: On 07/17/2018 12:05 PM, David Frank wrote: > Hi, > According to the instruction, I have to mount a huge directory to hugetlbfs and create file in the huge directory to use the mmap huge page feature. But the issue is that, the files in the huge directory takes up the huge pages configured through > vm.nr_hugepages = > > even the files are not used. > > When the total size of the files in the huge directory = vm.nr_hugepages * huge page size, then mmap would fail with 'can not allocate memory' if the file to be mapped is in the huge dir or the call has HUGEPAGETLB flag. > > Basically, I have to move the files off of the huge directory to free up huge pages. > > Am I missing anything here? > No, that is working as designed. hugetlbfs filesystems are generally pre-allocated with nr_hugepages huge pages. That is the upper limit of huge pages available. You can use overcommit/surplus pages to try and exceed the limit, but that comes with a whole set of potential issues. If you have not done so already, please see Documentation/vm/hugetlbpage.txt in the kernel source tree. -- Mike Kravetz

2 1

mmap with huge page
by David Frank July 17, 2018

July 17, 2018

Hi,According to the instruction, I have to mount a huge directory to hugetlbfs and create file in the huge directory to use the mmap huge page feature. But the issue is that, the files in the huge directory takes up the huge pages configured throughvm.nr_hugepages = even the files are not used. When the total size of the files in the huge directory = vm.nr_hugepages * huge page size, then mmap would fail with 'can not allocate memory' if the file to be mapped is in the huge dir or the call has HUGEPAGETLB flag. Basically, I have to move the files off of the huge directory to free up huge pages. Am I missing anything here? Thanks, David

2 1

newly add MAP_SYNC flag to mmap
by David Frank July 17, 2018

July 17, 2018

Hi, I installed LTS 18.04 with headers. I'm trying to use the new flag MAP_SYNC with mmap call, but it complained MAP_SYNC undefined.What #define do I need to enable this? Thanks, David

2 4

Question about memcpy
by bing zhu July 17, 2018

July 17, 2018

Dear Sir/Ma'am Thank you for your time ,i'm a student new to linux kernel. I have a question about memcpy,i noticed that memcpy is faster in kernel than in user space for example : in a module helloworld , i use memcpy to copy a 4096B to a block of memory for like 10000 times and in user space i do the same thing,I noticed that kernel is faster than user , is it possible that in kernel when i insmod hello it can not be scheduled but in user space it will so kernel is faster? is there a possible way that a user task can run a block of code that uninterruptable? No switch ,no schedule ? Thank you !

5 20

Re: newly add MAP_SYNC flag to mmap
by Alex Arvelaez July 17, 2018

July 17, 2018

16.07.2018 10:15 PM David Frank <david_frank95(a)yahoo.com> napisał(a): > > Thanks Valdis. > > Yes, I got the kernel and header files installed. > > But there seems to be a lot of mman.h. I found one in /usr/include/asm-generic that defined MAP_SYNC, and I included it also, now I got passed compile and linking. I feel like your toolchain should be able to pick up the right file when you do #include <sys/mman.h> Regards, Alex

1 0

Trying to export per process network bandwidth stats under /proc/[pid]/
by Amos Bird July 14, 2018

July 14, 2018

Hello! I'm trying to add a proc entry under /proc/[pid] in order to get per process network bandwidth stats. I suppose it's similar to /proc/[pid]/io . However I couldn't find a modular way to add entries under /proc/[pid]. This thread https://lists.kernelnewbies.org/pipermail/kernelnewbies/2011-January/000475… suggests modifying the kernel which isn't practical to me. Here is what I've found by going through a bit of the related kernel code. 1. I need to find a way to override or hooking `sock_sendmsg`/`sock_recvmsg` functions with accounting variables like `task_io_accounting` 2. a way to add proc entries under /proc/[pid] with readops printing `current` netio accounting variables. Am I on the right path of implementing this? Thanks! -- Amos Bird amosbird(a)gmail.com

2 1

mmap huge page and MAP_SYNC flag support
by David Frank July 12, 2018

July 12, 2018

Hi,How do I enable the MAP_SYNC flag in mmap for linux 4.15, my code does not compile with that flag on linux 4.15.0-041500rc7-generic #201801072330 Also, how do I make mmap support map size of 2GB or larger, currently the call to mmap fails with mapsize of 2GB. thanks, David

2 7

What does the "-DPACKAGE" means?
by Dengke Du July 12, 2018

July 12, 2018

Hi all What does the "-DPACKAGE" means? ``````````````````````````````````````````````````````````` $(OUTPUT)test-libbfd.bin: $(BUILD) -DPACKAGE='"perf"' -lbfd -lz -liberty -ldl ``````````````````````````````````````````````````````````` In this link: https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf.git/tree/tools/buil…

2 2

Use of enums, why?
by John Whitmore July 11, 2018

July 11, 2018

I only learning the ropes, and might have missed the memo on the use of enums so forgive me. I have looked at https://www.kernel.org/doc/html/v4.10/process/coding-style.html#macros-enum… but that didn't answer my question. I'm doing a clean up of staging:rtl8192u: clearing out checkpatch errors and triming out data structures which aren't used in the code. So I came across a typedef of an enumerated type in the file: drivers/staging/rtl8192u/r8192U.h typedef enum rf_optype { RF_OP_By_SW_3wire = 0, RF_OP_By_FW, RF_OP_MAX } rf_op_type; A quick grep for this type in drivers/staging/rtl8192u directory shows that the type is never used outside that header definition. So I can remove it? Well no you can't because the values defined in the enum are used. In drivers/staging/rtl8192u/r8192U_core.c for example: priv->Rf_Mode = RF_OP_By_SW_3wire; that element priv->Rf_Mode is defined in the structure typedef struct r8192_priv { ... u8 Rf_Mode; /* For Firmware RF -R/W switch */ ... } So now to the question, as I understand it the compiler will use an int type for the enumerated type. The data structure r8192_priv doesn't use this int type because the programmer knows that a u8 will do to hold the 3 possible values defined by the enumerated type. So you're saving a few bytes in a data structure, which I'm happy about, but the point of enumerated types is, as I understand it, so that the compiler can do some checking to ensure a value is not assigned in error. Since the enum isn't being used, there is no compiler type checking, so why use an enumerated type? Might as well have used three #define values. The obvious thing to do is leave well enough alone. But I had to ask what is the correct implementation? Use the enum in the data structure, instead of u8? Or just #define a few constants? #define RF_OP_By_SW_3wire 1 #define RF_OP_By_FW 2 #define RF_OP_MAX 3 Given the space saving of u8 over int I'd probably go with the #define. Guess it depends on how many of those data structures are being declared. thanks for any thoughts or clarification John

2 2