gmane.linux.kernel.mentors

ISO 8601 Timestamp <-> UNIX Timestamp

Dan Luedtke — 2012-05-20T11:07:29

Hi everyone,

are there any functions to convert from and to ISO 8601 formatted timestamps?

I found
  void time_to_tm(time_t totalsecs, int offset, struct tm *result)
in include/linux/time.h around line 200, which comes close, but the
other way round would be more interesting.

regards
  Dan

Some questions in v6_copy_user_highpage_aliasing

Gavin Guo — 2012-05-08T03:40:44

Hi all,

I have traced v6_copy_user_highpage_aliasing in arch/arm/mm/copypage-v6.c
and found some questions.

 72 /*
 73  * Copy the page, taking account of the cache colour.
 74  */
 75 static void v6_copy_user_highpage_aliasing(struct page *to,
 76         struct page *from, unsigned long vaddr)
 77 {
 78         unsigned int offset = CACHE_COLOUR(vaddr);
 79         unsigned long kfrom, kto;
 80

I think the following statement is to solve aliasing problem that
kernel & user address map to the same physical frame. When
the kernel mapping is dirty, writing back is needed before the copy_page()
is used. Because the color bit of user space address & kernel one
may be difference which cause inconsistency. Now, the problem is
what is the scenario of dirty cache. When does the PG_dcache_dirty
set ? In normal copy-on-write case, I think the if clause will be false.

 81         if (test_and_clear_bit(PG_dcache_dirty, &from->flags))
 82                 __flush_dcache_page(page_mapping(from), from);
 83

The probl

Timer drivers using clocksource and clockevents api

Bahadir Balban — 2012-04-09T21:01:16

Hello,

I wrote a virtual timer for a virtual platform. I do not actually have a
timer hardware device, but I have a periodic irq that I can use and I
know the period which every irq occurs. 

The result I get is a working clock source, but it is slow. The driver
is not %100 right as I have little knowledge of clocksource and
clockevents API. Can you help in spotting issues in the below driver?

All the numbers are made up :) I would appreciate guidance for what to
use on clkrate, cycle value, and any other api parameters.

Thanks in advance,
Bahadir

Invitation to connect on LinkedIn

Carter Cheng via LinkedIn — 2012-02-28T18:05:27

LinkedIn
------------




    Carter Cheng requested to add you as a connection on LinkedIn:
  

------------------------------------------

Greg,

I'd like to add you to my professional network on LinkedIn.

- Carter

Accept invitation from Carter Cheng
http://www.linkedin.com/e/-13f7ym-gz78yq20-a/gUfpE6MtuEhIxgEzB5uqY5tHupjIxRpxsUQoK7Qq/blk/I179014284_75/1BpC5vrmRLoRZcjkkZt5YCpnlOt3RApnhMpmdzgmhxrSNBszYRdRYQe38Qcj0VdP59bSV5hCVTt3xIbPsRc3gNd38McjoLrCBxbOYWrSlI/EML_comm_afe/?hs=false&tok=0rSLxfV1vZml81

View invitation from Carter Cheng
http://www.linkedin.com/e/-13f7ym-gz78yq20-a/gUfpE6MtuEhIxgEzB5uqY5tHupjIxRpxsUQoK7Qq/blk/I179014284_75/3kTnPgUczgNc3ATckALqnpPbOYWrSlI/svi/?hs=false&tok=120Df1qBfZml81

------------------------------------------

Why might connecting with Carter Cheng be a good idea?

Carter Cheng's connections could be useful to you:

After accepting Carter Cheng's invitation, check Carter Cheng's connections to see who else you may know and who you might want an introduction to. Building t

Invitation to connect on LinkedIn

Frankie onuonga via LinkedIn — 2012-01-21T18:57:07

LinkedIn
------------




    Frankie onuonga requested to add you as a connection on LinkedIn:
  

------------------------------------------

Greg,

I'd like to add you to my professional network on LinkedIn.

- Frankie

Accept invitation from Frankie onuonga
http://www.linkedin.com/e/-13f7ym-gxp02siu-35/gUfpE6MtuEhIxgEzB5uqY5tHupjIxRpxsUQoK7Qq/blk/I158095177_75/1BpC5vrmRLoRZcjkkZt5YCpnlOt3RApnhMpmdzgmhxrSNBszYRdRYTdP4Rej0Udj59bPtSqQJCmClVbPAVd3cUcz4QcPkLrCBxbOYWrSlI/EML_comm_afe/?hs=false&tok=1BWonvg2BNx541

View invitation from Frankie onuonga
http://www.linkedin.com/e/-13f7ym-gxp02siu-35/gUfpE6MtuEhIxgEzB5uqY5tHupjIxRpxsUQoK7Qq/blk/I158095177_75/3kTnPsTcjkVc3wRckALqnpPbOYWrSlI/svi/?hs=false&tok=2qEhUzdhNNx541

------------------------------------------

Why might connecting with Frankie onuonga be a good idea?

Frankie onuonga's connections could be useful to you:

After accepting Frankie onuonga's invitation, check Frankie onuonga's connections to see who else you may know and who you might want an int

tcp/ip

Johan Nestaas — 2012-01-15T10:35:43

Hi,

Sorry if this is too newb of a question for the mailing list, but I tried
the kernelnewbies irc chat and no one was responding. Maybe someone could
direct me to a proper mailing list if this isn't the right place.

I've been trying to follow the latest kernel source of the tcp/ip stack but
I feel overwhelmed and I don't know the best file to start reading from.
Right now I'm reading linux/net/ipv4/af_inet.c but I have a feeling this is
not the best place to start. Does anyone have any suggestions? This would
basically be my start to analyzing the kernel source and operating system
data structures, outside of an Operating Systems class I took a semester
ago.

Thanks,
Johan
_______________________________________________
Kernel-mentors mailing list
Kernel-mentors< at >selenic.com
http://selenic.com/mailman/listinfo/kernel-mentors

One question about the kernel timer

Wizard — 2012-01-14T15:49:29

All

I am reading this code in kernel/timer.
my thought is for most of the timer in tv->vec, after cascade they
will stay in the same level.
Am I right?

1014static int cascade(struct tvec_base *base, struct tvec *tv, int index)
1015{
1016        /* cascade all the timers from tv up one level */
1017        struct timer_list *timer, *tmp;
1018        struct list_head tv_list;
1019
1020        list_replace_init(tv->vec + index, &tv_list);
1021
1022        /*
1023         * We are removing _all_ timers from the list, so we
1024         * don't have to detach them individually.
1025         */
1026        list_for_each_entry_safe(timer, tmp, &tv_list, entry) {
1027                BUG_ON(tbase_get_base(timer->base) != base);
1028                internal_add_timer(base, timer);
1029        }
1030
1031        return index;
1032}

I think this function is trying to move some of the timer to another level.
Since the internal_add_timer() will calculate the hash number based on
the timer->expire,
so I guess for the ti

Should I generate inode numbers?

Dan Luedtke — 2011-12-13T17:43:29

Hello everyone,

I am thinking about a concept for a filesystem (read: kernel-module),
and when studying the sources I found in in include/linux/fs.h

 751 struct inode {
 ...
 771         /* Stat data, not accessed from path walking */
 772         unsigned long           i_ino;

Source: http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=blob;f=include/linux/fs.h;h=057434af2b900e18745b375b109e52ef03a9c6c0;hb=HEAD#l772

I want to use the blocknumber as inode number to
-keep the code simple
-avoid collisions and
-always know which block of a device an inode can be found in.

When I would my code like intended, and it would be compiled on a 32
bit machine, then unsigned long would be 32 bit. That would cause some
trouble when mounting partitions with more than 2^32 blocks, wouldn't
it?

So, as long as i_ino is not 64 bit (which was already discussed years
ago by some ext4 developers I think), what should I do? Should I
generate inode numbers (e.g. by incrementing a "global" variable) and
check for

advice on next step

Frankie Onuong'a — 2011-12-02T10:15:59


Hi guys,
I would like to kindly ask for an extension on the time I gave.I came across a book :
Understanding the Linux Kernel by Daniel P Bovet and Marco Cesati.
I believe to further help i need to go from page on to the end of it.I have already started but i believe with another two weeks all should be on my finger tips.
Thanks again to : vijaitm< at >gmail.com and Greg for the direction.
keep well.

frankie
La presente comunicación tiene carácter confidencial y es para el exclusivo uso del destinatario indicado en la misma. Si Ud. no es el destinatario indicado, le informamos que cualquier forma de distribución, reproducción o uso de esta comunicación y/o de la información contenida en la misma están estrictamente prohibidos por la ley. Si Ud. ha recibido esta comunicación por error, por favor, notifíquelo inmediatamente al remitente contestando a este mensaje y proceda a continuación a destruirlo. Gracias por su colaboración.This communication contains confidential information. It is for the exclu

What should I do when I cant get any response to my patch

Denis Kuzmenko — 2011-11-26T14:02:43

Hi Linux Mentors.

I've made a patch to add support of LCD W35i for Mini2440 board. Looks
like this is a new version of LCD which is supplied with all new boards.

First time I've posted that patch:

https://lkml.org/lkml/2011/6/13/345
https://lkml.org/lkml/2011/6/14/331

got comments and fixed all things required to be fixed.

I've tried to resend that patch in five months and now I can't get any
replies

https://lkml.org/lkml/2011/11/17/376

I've formed recipients list from the output of get_maintainer.pl script.
This looks like no one is interested in this feature being in mainline?

There is no way I can prove that this feature is needed by many users
but I've received a letter from someone googled W35i support with a
question on how to use that display with Linux.

So suggest me please what should I do next?

Should I email this path to Andrew Morton for inclusion in -mm three?
Or should i just periodically repost that waiting or someone got interested?

introduction

Frankie Onuong'a — 2011-11-24T08:30:55

Hi guys,My name is Frankie Onuonga .I am a student at university taking Computer science and math.
I am passionate about open source technology.I just waned to introduce myself and ask where does start.
thanks
Frankie Onuonga


La presente comunicación tiene carácter confidencial y es para el exclusivo uso del destinatario indicado en la misma. Si Ud. no es el destinatario indicado, le informamos que cualquier forma de distribución, reproducción o uso de esta comunicación y/o de la información contenida en la misma están estrictamente prohibidos por la ley. Si Ud. ha recibido esta comunicación por error, por favor, notifíquelo inmediatamente al remitente contestando a este mensaje y proceda a continuación a destruirlo. Gracias por su colaboración.This communication contains confidential information. It is for the exclusive use of the intended addressee. If you are not the intended addressee, please note that any form of distribution, copying or use of this communication or the information in it i

automount codeing in linux

joshi dhaval — 2011-11-15T16:46:30

Hi Guys,

can anyone suggest me about codeing for automount in linux ? how can i debug issues related to automount and how can i analyze dumps for it 

sorry i am bit new to kernel so not sure how to start with troubleshooting kernel issues.

Regards,
Dhaval
_______________________________________________
Kernel-mentors mailing list
Kernel-mentors< at >selenic.com
http://selenic.com/mailman/listinfo/kernel-mentors

Where to ask questions related to kernel development?

Denis Kuzmenko — 2011-11-13T23:33:17

Hi All,

I'm getting problems with learning the linux kernel.
Particularly, working with kernel timers from interrupt handler i got oops indicating that I'm trying to timer_add() on currently pending timer.

Can you suggest me some mailing list or board or anything where I can post some piece of code, describe it's behavior and get some help?
Suppose, linux-kernel mailing list is bad place for questions like that :)

Thank you, Denis Kuzmenko.

Which source three to use as base for development

Denis Kuzmenko — 2011-11-12T22:00:44

Hi All,

I want to learn kernel development. I;ve downloaded latest stable
tarball and started with it. During development of some "hello
world"-like module the bug was founded and it looks like i have a fix.
So I need to post pathes as well. I would be good to have a convenient
way to update my source three each time new stable release comes up
(at this time I don't want to use anything except stable, because I
don't want to get so deep in bufixing). Looks like I need to use some
git-repo to update my three easily and have a convinient way to
produce pathes (with `git diff`), but I don't know which repo to pick.

Is this command set to get needed three is best for me:

git clone git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
cd linux-stable
git checkout v3.0.9

Any suggestions?

Thank you, Denis Kuzmenko.

May be a bug in scatterlist.c

Michael Wang — 2011-11-07T04:01:21

Hi, Experts

I found there are may be a bug in

file: lib/scatterlist.c
func: __sg_alloc_table
line: 248

left -= sg_size;

if the max_ents is 1 and nents is 3 and sg chain enabled,
the sg_size will be 0, and it will always while because the
left will not go less.

I think here should use

left -= alloc_size;

I will be very happy if anyone can help me to see whether
this is really a bug, and tell me where should I report it,
and I will be very glad to provide a patch if it is a bug.

Best regards,
Michael Wang

(unknown)

王振 — 2011-11-02T03:38:42

_______________________________________________
Kernel-mentors mailing list
Kernel-mentors< at >selenic.com
http://selenic.com/mailman/listinfo/kernel-mentors

Drivers

Emil Goode — 2011-10-24T21:37:22

Hello,

I'm kernel newbie and I have some questions regarding how to write drivers.
I started to write some software so that a picoscope (a usb oscillioscope)
can be used on Linux. But I soon realized
that the proprietary Linux driver that picotech have provided with the
device is not good enough.
So my question is, what can I do? Since no hardware specification is
avaliable I guess the only way to make your
own driver is through reverse engineering? I'm not sure if its realistic as
I think it is probably very difficult?

When developers write their first driver do they always have the hardware
specification at hand?  Im guessing most hardware manufacturers don't make
that avaliable.

Here is the device:
http://www.picotech.com/entry-level-oscilloscopes.html

I would be very thankful for some advise from anyone that is experienced in
this kind of matter.

Best regards, Emil
_______________________________________________
Kernel-mentors mailing list
Kernel-mentors< at >selenic.com
http://selenic.com/mailman/listin

how to get the current on x86 platform?

Wizard — 2011-10-18T15:43:00

Experts

On LKD 3rd, it mentioned the current on x86 is calculated by masking
out the 13 least-significant bits
of the stack pointer to obtain the thread_info structure.

While I use cscope to looking for the current definition and find it
in arch/x86/include/asm/current.h


DECLARE_PER_CPU(struct task_struct *, current_task);

static __always_inline struct task_struct *get_current(void)
{
return percpu_read_stable(current_task);
}

#define current get_current()

So first each cpu has a task_struct pointer called current_task.

Then I have difficulty in understanding the percpu_read_stable().
I find the definition but...
Someone could give me some help? I don't see it has some
relationship with the thread_info structure.

#define percpu_read_stable(var)percpu_from_op("mov", var, "p" (&(var)))

#define percpu_from_op(op, var, constraint)\
({\
typeof(var) pfo_ret__;\
switch (sizeof(var)) {\
case 1:\
asm(op "b "__percpu_arg(1)",%0"\
    : "=q" (pfo_ret__)\
    : constrain

Several questions on the Highmem definition

Wizard — 2011-10-13T04:38:03

Hi, Experts

I am reading some articles on the Highmem. And I have several confused points.
Hope someone could share me some light.

1. this is the kernel policy for 1g/3g division.
    Some one have an idea on the reason of this division?
   And Highmem are the physical memory above 1G with the default kernel
   config?

2. the 128M reserved for kernel data structure is a must?
   on website, http://kerneltrap.org/node/2450, it says
   "This means that the kernel virtual address in this 128 MB
   is not mapped to physical memory."
   This is not in the physical memory? I think it should be in the
   physical memory.  Don't understand the meaning.

3. Then is there a limitation for the minimum physical memory requirement
   for linux system? more than 128M, since the kernel needs at least this
   amount to store some importance information.

4.  If the kernel want to access the mem above highmem, it should use kmap
   to create a PTE in the reserved 128M area?

    After that, some part of the kernel virtual

how the kernel call those __init calls?

Wizard — 2011-10-04T15:27:35

Hi, all

I am curious about how the kernel call these __init calls and keep an order?
For example, the function in kernel/resource.c.
static int __init ioresources_init(void).

I see the script in vmlinux.lds.
. = ALIGN((1 << 12)); .init.text : AT(ADDR(.init.text) - 0xC0000000) {
_sinittext = .; *(.init.text) *(.meminit.text) _einittext = .; }

I think every __init call will be allocated in the _sinittext section.
But I don't find where those functions are called.
And how the order of those init functions are guaranteed?

And also the meaning of the AT(ADDR(.init.text) - 0xC0000000).
I am not very clear.

Use of inode fields for storing encryption key of file

Parag Amritkar — 2011-09-27T15:49:51

I am working on inherent kernel functionality for file encryption
and decryption ; directory name with some *predefined prefix* will be
automatically be encrypted. Now I am stucked on storing encryption key for
file securely ,can I use any unused field of inode for that? Will it be
efficient or please suggest any other idea.
_______________________________________________
Kernel-mentors mailing list
Kernel-mentors< at >selenic.com
http://selenic.com/mailman/listinfo/kernel-mentors