decode /proc
in [General Linux]
|
From: Robert Heller on 1 Feb 2010 15:00 At Mon, 1 Feb 2010 18:41:19 +0000 spike1(a)freenet.co.uk wrote: > > And verily, didst Tauno Voipio <tauno.voipio(a)notused.fi.invalid> hastily babble thusly: > > unruh wrote: > >> On 2010-02-01, peter <cmk128(a)gmail.com> wrote: > >>> I am here, I posted a post a few days ago. > >>> > >>> Thanks for the replies. I guess I have no way to read the /proc from > >>> outside. May be i need to write a agent that run inside the linux, > >>> then read the information from it from outside. > >>> thanks > >>> from Peter > >> > >> Maybe if you told us what it is that you really want to do someone could > >> give some useful advice. I do not know what "read the /proc from > >> outside" could mean. /proc only exists if the linux kernel is running. > >> If it is not running, it does not exist, and there is nothing to read. > >> What does "from the outside" mean? How would you imagine being able to > >> read kernel things "from the outside" and at the same time not being > >> able to read /proc as a filesystem? > > > > > > It seems that the OP is running a virtual machine or debugger, > > and he is interested to look at /proc entries without running > > the target system. > > > I think he wants to read them from the running system from OUTSIDE the > virtual machine... for some peculiar reason. Maybe he wants to avoid killing the perverbial cat? :-) I think the whole point of Quantum Machanics and the Uncertainity Principle is that you can't do that. (Sorry -- I just finished reading "A brief History Of Time" by Stephen Hawkings.) -- Robert Heller -- 978-544-6933 Deepwoods Software -- Download the Model Railroad System http://www.deepsoft.com/ -- Binaries for Linux and MS-Windows heller(a)deepsoft.com -- http://www.deepsoft.com/ModelRailroadSystem/
From: Tauno Voipio on 2 Feb 2010 10:36 Robert Heller wrote: > At Mon, 1 Feb 2010 18:41:19 +0000 spike1(a)freenet.co.uk wrote: > >> And verily, didst Tauno Voipio <tauno.voipio(a)notused.fi.invalid> hastily babble thusly: >>> unruh wrote: >>>> On 2010-02-01, peter <cmk128(a)gmail.com> wrote: >>>>> I am here, I posted a post a few days ago. >>>>> >>>>> Thanks for the replies. I guess I have no way to read the /proc from >>>>> outside. May be i need to write a agent that run inside the linux, >>>>> then read the information from it from outside. >>>>> thanks >>>>> from Peter >>>> Maybe if you told us what it is that you really want to do someone could >>>> give some useful advice. I do not know what "read the /proc from >>>> outside" could mean. /proc only exists if the linux kernel is running. >>>> If it is not running, it does not exist, and there is nothing to read. >>>> What does "from the outside" mean? How would you imagine being able to >>>> read kernel things "from the outside" and at the same time not being >>>> able to read /proc as a filesystem? >>> >>> It seems that the OP is running a virtual machine or debugger, >>> and he is interested to look at /proc entries without running >>> the target system. >>> >> I think he wants to read them from the running system from OUTSIDE the >> virtual machine... for some peculiar reason. > > Maybe he wants to avoid killing the perverbial cat? :-) I think the > whole point of Quantum Machanics and the Uncertainity Principle is that > you can't do that. (Sorry -- I just finished reading "A brief History > Of Time" by Stephen Hawkings.) In principle, he could, by following the data structures forming the /proc file system. The task is about the same as digging in with kgdb - a quite desperate measure, and it needs at least the kernel symbol table to work. - My lecturer of quantum mechanics (about 45 years ago) said that if the thing does not feel weird, the student is not following. -- Tauno Voipio
From: unruh on 2 Feb 2010 17:33 On 2010-02-02, Tauno Voipio <tauno.voipio(a)notused.fi.invalid> wrote: > Robert Heller wrote: >> At Mon, 1 Feb 2010 18:41:19 +0000 spike1(a)freenet.co.uk wrote: >> >>> And verily, didst Tauno Voipio <tauno.voipio(a)notused.fi.invalid> hastily babble thusly: >>>> unruh wrote: >>>>> On 2010-02-01, peter <cmk128(a)gmail.com> wrote: >>>>>> I am here, I posted a post a few days ago. >>>>>> >>>>>> Thanks for the replies. I guess I have no way to read the /proc from >>>>>> outside. May be i need to write a agent that run inside the linux, >>>>>> then read the information from it from outside. >>>>>> thanks >>>>>> from Peter >>>>> Maybe if you told us what it is that you really want to do someone could >>>>> give some useful advice. I do not know what "read the /proc from >>>>> outside" could mean. /proc only exists if the linux kernel is running. >>>>> If it is not running, it does not exist, and there is nothing to read. >>>>> What does "from the outside" mean? How would you imagine being able to >>>>> read kernel things "from the outside" and at the same time not being >>>>> able to read /proc as a filesystem? >>>> >>>> It seems that the OP is running a virtual machine or debugger, >>>> and he is interested to look at /proc entries without running >>>> the target system. >>>> >>> I think he wants to read them from the running system from OUTSIDE the >>> virtual machine... for some peculiar reason. >> >> Maybe he wants to avoid killing the perverbial cat? :-) I think the >> whole point of Quantum Machanics and the Uncertainity Principle is that >> you can't do that. (Sorry -- I just finished reading "A brief History >> Of Time" by Stephen Hawkings.) > > > In principle, he could, by following the data structures > forming the /proc file system. The task is about the same > as digging in with kgdb - a quite desperate measure, and > it needs at least the kernel symbol table to work. If he is not running Linux, there is no data structure to look at ( never mind the fact that there is no data structure to look at, because for many of the files in /proc, the contents are manufactured when a read request is made to that file). If a linux kernel is running, then the /proc filesystem is, or should be there and the data can be read from it. I am having immense difficulty in understanding what he wants. He is either running a Linux kernel, in which case /proc is there, or he is not, in which case the contents of /proc do not exist and make no sense. > > - > > My lecturer of quantum mechanics (about 45 years ago) said that > if the thing does not feel weird, the student is not following. >
From: peter on 3 Feb 2010 11:47 take a look http://peter-bochs.googlecode.com , It is a debugger, I made it for my operating system development. I want it to support / proc. In my debugger, I can read anywhere in memory, I can also read the CPU status. But I don't know where to find the /proc information in memory (perhaps it doesn't exist in memory). For operating system development, only reading the memory for debug is not enough. I need to read the data-structure(data structure for paging information, segmentation information, process information or etc... ) of kernel. But for linux, I don't know where is those structures in memory. If the /proc is not hold in memory or unable to decode it. I am going to write a small standard about "how to put and decode kernel debug information in memory/harddisk". If your operating system is follow this small-standard, then you can use my peter-bochs debugger to debug your operating system in high-level way. "high-level way" mean not just reading some bytes in memory, it means allows us to read the data structure of your kernel in real time (forgive my bad English). Also, I am applying a mPhil degree. If I use this as a research topic, is it too naive? Nice guys, thanks. Here is a good forum.
From: Robert Heller on 3 Feb 2010 12:18
At Wed, 3 Feb 2010 08:47:13 -0800 (PST) peter <cmk128(a)gmail.com> wrote: > > take a look http://peter-bochs.googlecode.com , It is a debugger, I > made it for my operating system development. I want it to support / > proc. > In my debugger, I can read anywhere in memory, I can also read the CPU > status. But I don't know where to find the /proc information in memory > (perhaps it doesn't exist in memory). No, it does not. > > For operating system development, only reading the memory for debug is > not enough. I need to read the data-structure(data structure for > paging information, segmentation information, process information or > etc... ) of kernel. But for linux, I don't know where is those > structures in memory. Have you looked at the kernel *source code*. To misquote Obewan Knobie: "Use the Source, Peter". http://www.kernel.org/ > > If the /proc is not hold in memory or unable to decode it. I am going > to write a small standard about "how to put and decode kernel debug > information in memory/harddisk". If your operating system is follow > this small-standard, then you can use my peter-bochs debugger to debug > your operating system in high-level way. "high-level way" mean not > just reading some bytes in memory, it means allows us to read the data > structure of your kernel in real time (forgive my bad English). > > Also, I am applying a mPhil degree. If I use this as a research topic, > is it too naive? > > Nice guys, thanks. Here is a good forum. > -- Robert Heller -- 978-544-6933 Deepwoods Software -- Download the Model Railroad System http://www.deepsoft.com/ -- Binaries for Linux and MS-Windows heller(a)deepsoft.com -- http://www.deepsoft.com/ModelRailroadSystem/ |