The Red Pill

How to climb Mountains

2007-01-26T01:02:00.000-08:00

26th Jan, its Indian and Australian Republic day, and today i was suppose to drive down to Harnai Beach for a quick *relaxation* holiday. But, things got screwed and here i am lying on my sofa sipping freshing milled-n-brewed starsbucks coffee and eating self made pasta cooked in red-wine. I know, not a very good combination but, had to manage with what was available at home :D

Hmm i have not been blogging lately, may be because my work or my high maintenance girlfriend takes most of my time. Or probably because my thoughts these days not "legal" enough to be blogged about openly :)

anyways, this entry is a modified version of a short reflection taken from book - "Like flowing River" by Paulo Coelho. It's quite relevant to the struggles of our lives, which makes it a good read ...

How to climb Mountains
_____________________

Choose the mountain you want to climb -
Don't be influenced by what other ppl say: `that one's prettier or easier`. Since you are going to put a lot of time and energy, choose your own mountain.

Find out how to reach the mountain -
Often you can see the mountain in the distance - beautiful and interesting. However when you try and reach it, what happens ? Its surrounded by roads/forests. Whats looks easy on map is way too complicated ? So you must try all paths are tracks, until one day you, you find yourself before the peak you want to climb.

Learn from someone who has been there before -
However unique you may think you are, there is always someone who has had the same dream before, and who would have left signs behind that will make the climb less arduous. Your climb is your responsibility, but never forget that other people's experiences are always helpful.

Respect you body -
You will only manage to climb the mountain if your give your body care it deserves. You have all the time life gives you, so don't demand too much from your body.

Respect your soul -
Don't keep repeating `I'm going to do it`. Your soul knows this already. What it needs to do is to use this long walk in order to grow, to reach out as far as the horizon, to touch the sky. Obsession will not help you, and will end up spoiling the pleasure of the climb. On the other hand don't snap your inner strength with `It's harder that i thought`.

Be prepared to go the extra mile -
The distance to the top of the mountain is always greater than you think. There is bound be a moment when what seemed close is still very far away.

Be joyful when you reach the top -
Very Important. Cry, clap your hands, shout out loud, that you made it !! let the wind cool your hot, blow the dust on your heart. What was one day only a dream, a distant vision, is now part of your life.

You made it, and this is good.

Make a promise -
Now that you have discovered a strength you did not even know you had, tell yourself that you would use it for the rest of your days.

Tell your story -
Tell your story and be an example to others.

---/

So, what do you say ? ready for that extra mile :-)

Eternal Sunshine of the Spotless Mind

2006-06-09T07:55:00.000-07:00

How happy is the blameless vestal's lot?
The world forgetting, by the world forgot.
Eternal Sunshine of the Spotless Mind.
Each prayer accepted and each wish resigned.

-- Alexander Pope, taken from its eponymn
( Movie: Eternal Sunshine of the Spotless Mind )

I counldn't understand the meaning properly untill i read Eloisa to Abelard http://en.wikipedia.org/wiki/Eloisa_to_Abelard

Towers of Hanoi

2006-04-11T02:34:00.000-07:00

"Hanoi" - as a close friend of mine calls himself, has started a new effort called "Towers of Hanoi".

The so called one man startup has started getting some cool techie work. After much running around doing some creepy service work and making small money .. some good work has found its way to Hanoi.

Of cource .. you must be thinking why the hell am i spending time writing such stupid blog. Well just to attract those "like-minded" few.

The work includes -
1. Moving the fast growing slibe.com to a OpenSSI Linux compute cluster with 4 nodes, which would be almost 10-15 times powerful than Old' Param 1000. ( work would start in June )

2. working with www.misoconcepts.com to research about HA clusters which can be virtualized just like VirtualFe. One very interesting objective of this work is to see if clusters can be extended/replicated beyond a normal LAN to take care of HA.

3. Working with some aquintance on "Distributed database Filesystem". This would ofcource be user space and would help the person host a online distributed storage solution with cost as low as 1$/GB. (work to start after misoconcepts)

And the best part is that he makes good money working on these. And I get free booze whenever i meet him.

cheers ... hic!!

Paradise Paradox

2006-04-10T01:18:00.000-07:00

Me and steve discusses this over booze .. but i guess this one is still worth blogging.

Paradise: any place of complete bliss and delight and peace.

If there is indeed such a place called paradise then everybody in that place would be happy. But the meaning of happiness for me is definitely not that of Multi's or to talk of someone famous say Bill Gates. Then how can both of us be happy living together.

Also if such a way of living which could make every righteous soul happy after death .. why wouldn't they strive to achieve this when alive... why do they have to die for it :-)

So, all we can conclude is that either paradise is a hoax promised to man by God to keep him/her on right track OR its a dream or a feeling of virtual reality which is visualized or realized by each individual soul differently.

Think about it.

A brain in drain

2006-04-05T13:03:00.000-07:00

My mind is in a state of utter unrest.

A friend of mine once told me that IIT helped me develop wings and left me on a runway. Now I would be a soab if I couldn't break all fetters and takeoff.

Although veritas has been fair to me in giving me good work. I think i am capable of doing much better work and to me its like pumping diesel into aircraft engine.

Even since i have met bhatia .. the !dea of a startup has struck me like a leach. And now neither can i feed this parasite nor can i take it off.

Plus for this endeavor i would require a year long supply of -
1. Moolah
2. Brains

After some months of struggle I think the first one can be taken care of. But the people that i trust are either too busy or take it as yet another booze talk.

But i have hope, coz i know whenever i have been in such situation i have managed to come out well.

Yaaran da Tushon

2006-04-03T12:09:00.001-07:00

I got my well awaited laptop HP dv8000z -

AMD Turion x64 2 Gigz
17" brightview
RAM - 512 MB
Bus 1600 Mgz
V. Ram - 128 MB
Chipset - ATI Xpress 200M
Wifi 802.11 a/b/g
DVD +/- RW
7.8 lb

I am a sony fan and wanted to buy a VAIO but .. it would have costed me almost twice. In case of camera/walkmans this is acceptable .. but for something you just assemble why the hell should you charge double.

Plus no one offered a x64 based PC.

I hate Thinkpads for their poor graphics and looks. Dell is good .. but somewhat not that sturdy.

My wishlist of gadgets include -
1. Sony Dig-SLR (expected to hit market in may '06) - $1000
2. Sony PS3 - $500
3. Samsung X730 phone - $300
4. OpenMosix Cluster ;-)

As kids you love toys ... and when you grow up the toys just become a little expensive.

M$ Interview

2006-04-03T10:34:00.000-07:00

I like giving interviews .. simply because they keep me updated of latest techs and to put it bluntly my contemporary cost. But I was at M$ for a different reason .. to know if M$ is really as bad as the open source community claim.

This was an "only IITians" walkin at pune. The first day, I went there wearing a "Got Linux" t-shirt and ppl there didn't seem to mind it. The HRrahulpc@microsoft.com) welcomed me and showed me a room make myself comfortable.

My first interviewer was wearing a pink shirt with "Microsoft" imprinted on the collar. He started of asking my work and then started asking me some stupid questions which he called "puzzles". They included all kind of questions like "what is a link list" and "how can you make it circular". Finally he convinced himself that i wasn't probably as stupid as i look and gave me some more questions. He demanded some "tight code" for -

1. Reverse a link list
2. Sort a given string.

This reminded me of my worst prof ever "Hon. Dr. P. K. Das" of IIT Guwahati. I still respect that man ... coz he did teach me one thing - "There would always be one person, much more stupid than you and yet sitting on your head if not making more money than you."

I asked this interviewer again .. if i was being interviwed for a Q.A job, but he told me that i was being considered for a R&D job for the mobile computing group. For the second question I tried to use "String Tries", but very soon came to know that i am crossing my interviewers intelligence level.

Anyways this interview got over and I met this guy dressed smartly and looking at my t-shirt .. as if some stupid has entered the room. He gave me a lecture about "Enterprise Software Development" and how Linux lacks it.

Rest of the interview was as stupid as the previous one .. write a code, find mistakes in it and write test cases. I tried to suggest that probably .. asking me some analytical questions should be better. But i guess i was taken for a fool.

Next day i decided not to go for an interview .. but the HR gave me a call that I should come as my interviews were going *very good*. And stupid of me that i again went there.

But, this was a good day. I met the manager of "mobile computing group". I tried explainig him that i wasn't been tested me for my skill sets. He gave me a puzzle which was similiar to the Chinese Postman problem. I liked it.

And then asked me a mix of good and stupid questions like -
1. What is the common filesystem on linux.
i said .. ext2
2. which company makes it.
well ... it can be the guy next door also.
3. Whats the max path length in ext2
4096 ( he must be googling ;-)
4. How are acls or xattrs organized on disk in ext2.
There are common blocks on the disk and different inodes point to same.
5. How is OS security integrated with filesystem security.
6. Does filesystem security depend upon support from processor.
yes.

anyways he made me an offer (some 8+ lakh annum).. and i told him frankly that if the work is going to me same as the skillsets i have been tested upon, I would not be happy doing it and he can hire a person half the cost as mine to do the same. I walked out.

Thoughout the interview i could feel that these guys dont know anything ouside M$ and they are anything but IITians.

Then HR requested me to meet one of the directors of technology, this guy was smart. He was ready to listen .. i showed him my blog and told him what veritas tested me for. He gave me to design a scheduler and then tested me on a graph problem.

All went well and he talked to me about a position in the Windows Server group and asked me to Hyd for some more interviews.

And then in next few days ... I faced the might of HR. I got a call from Vijaya (Vijaya Talluri v-vijtal@microsoft.com) from Microsoft asking me to join them in Hyd for an interview. I told them my apprehentions about taking a leave .. but she insisted me to take one. Then finally i decided to take a leave .. and she sent me details of flight and accomodations. I recieved a little under a dozen more ppl for the same.

I planned o fly on a friday at 5:30 and about 10 AM same day i get a call from her saying the position i was looking for i filled so she would .. help me get in touch with another HR person.

This is idiotic new character in the story .. calls me at 3PM asking me to prepare for an interview at 6PM. I told her .. that even if I want to I can't take calls during a flight. And then she quite easily says .. that i should drop the idea about flight and take a leave the following week. This fumed me .. and i blasted off .. and to my utter surprise she simply dropped the phone over me.

Next i get a call from Rahul apologizin for the same and told me that I recieved a 5 star "must hire rating" from that director and thats the reason M$ wants to take me in... Finally that stupid HR again called up acting as if nothing has happened.

This time ... i dropped the phone. Thrice.

And, they never spammed again.

Ethics of writing a Signal Handler

2005-10-26T01:23:00.000-07:00

Signal handlers are generally used incorrectly in the code. I am writing some guidelines for writing a signal handler (This involved comments from Boris and also some hacking with apache.)

1. A signal should do the minimum, like just setting/unsetting a variable. Most often signal handlers interupt execution of a syscal (if its in a interuptable sleep) , and this syscal restarts when signal handler finishes. Now If this signal handler calls the same syscal which was interupted .. we have a "futex" ( I have had some real bad experiences with this .. trust me).

Boris - "On the other hand, if we expect some code to execute right away,
we need to also arrange any blocking system calls to be interrupted by the
signal. By default, some calls like read() are automatically restarted
after the signal handler returns but this behaviour can be changed with
sa_flags in sigaction()."

2. Earlier the implementation of exclog had same sig-handler for both SIGTERM and SIGHUP but If apache calls SIGHUP and then SIGTERM (which it does) in a very short span of time, sig handler for SIGHUP will get interupetd by sig handler of SIGTERM !! .. and as boris explains it "BOOM!! Anything from lock up to infinite loop to smoke and fire may happen."

3. Linux automatically blocks the signal which caused a signal handler to be invoked during the execution of that handler BUT it doesn't block any other signals that may be using the same handler. Luckily, this is easy to achieve with sa_mask in sigaction().

4. In addition to point 1. Sometime when you modify a variable in a signal handler which is later being checked in the main loop for any changes, it is always advisable to declare this global variable as volatile. Coz the compiler seeing that variable is only check in main loop, may optimize so that the vaiable is not loaded from memory everytime its checked. Hence changes by signal handler may not be visible in the loop.

From Apache's point of view -
When Apache is stopping it sends a SIGHUP first and then SIGTERM to all its children ( thats what was also visible in apache modules). And most of the modules catch both these signals to clean up stuff. So, I am not sure why SIGHUP is handled my modules, it can be handled using SIG_ING. For that matter why SIGHUP and SIGTERM are handled when the process is going to die anyways .. and you can't do anything about it.

Why buy a life insurance for self .. if you are going to die ;-).

What if two SIGTERMs appear together, it would cause a futex, and the process will never die.

Veritas Interview Part II

2005-10-21T01:44:00.000-07:00

I am just summing up some more qus which veritas asked me .. during all those technicle rounds ... Some of these were what i ans irrespective of whether they were asked or not :D

They told me that they just want my approach .. Even if my ans are incorrect.

here it goes ...

Q1. Would the TLBs be synchronized in SMP environment. And why.
Ans: I struggled and then said No. Obviously both CPU would get different set of instructions and there is not sense in syncronizing them.

Q2 How does computer boot up. ( I myself pulled the guy to this qus ).
This was interesting ..
Ans: Algo for x86
1. BIOS dos POST (some hardware checks) and Loads MBR (512Kb) at some fixed addr i think .. 0xfffffff0
2. sends RESET on the BUS. CPU starts execution from 0xfffffff0
3. lilo or whatever is loaded into MBR searches for Linux kernel's compressed image.
4. Loads kernel and the TRICK is.. that BIOS can only address physical addresses and even the CPU while booting knows only physical addresses (Since its registers like GDTR/LDTR and all are not set to traslate virtual-physical addresses)
At one point linux has to set the resgisters make CPU start executing in "protected mode" i.e 32bit virtaul addr. mode .. but since after that point the kernel execution path has to change coz kernel is %eip is poining to kernel code as per physical addresses. So before booting CPU into virtual addr mode .. kernel copies itself to another location from where execution would start when CPU comes to protected mode.

Q3. What do you call a daemon process. Who cleans up all the zombie processes.
Ans: Daemon is a program whose parent is already dead and init owns it now. And also it doesn't hold any tty/pts. Daemon is issentially a zombie process. And since it doesn't know if its parent would is dead .. its waiting for its parent to call wait on it. (Cos if parent calls wait and child is dead .. there would be kernel OOPs).

So init cleans all listed zombie processes.

Q4. What is cache snopping.
Ans: No idea ...

Q5. So if i say Its syncing the L1/L2 caches in SMP. So why and how is the done.
Ans: (I could only answered it when the guy asked me in second interview) You sync them coz in SMP even if the OS changes something in memory its no sure if the CPU is realy writing to the mem or just L1/L2 cache. This syncing is done through inter processor interupts. These interrupts can be raised writing to cr0 (control) register.

Q6. How are interrupts delivered to different CPUs in SMP.
Ans: APIC - advanced progm Inter. Contr. or something like that is persent in x86 (p3 onwards) you can program it to mask of some int. for some CPU. OR you can program it to give int. to CPU with lowest priority of task.

I dont know about 8259 or whatvere the int. controller was that in 386-P1.

Q7. Explain what is the layout of Ext2 filesystem.
Ans. Harddisk is divided into parts ( i dont know the actual term) and each part has itz own superblock, disk bitmap, inode bitmap, intode table , data blocks. Same copy of superblock is maintained by each segment so that if the first one gets corrupt then the others can be read and recovered. Bitmap cntains the bitmap of the data/inode. So 0 in bitmap would mean that corresponding block in the inode/data blocks is valid. It is used to quickly delete files.

Q8.How would you write a recursive function in assembly.
Ans: In normal functions you
push %ebp
mov %esp, %ebp .. so that you acn access the arguments pushed to the stack using %ebp
So Arg1 can be 8(%ebp) and Arg2 at 12(%ebp).
I told him about the recursive function .. The only trick there was that you needed to pop %ebp on every return and move %ebp into %esp.

Q9. How would you query for a graph in another graph. ( This he asked after i told him about Glassbox stuff).
Ans: Well given you have starting-points/root of the graph same you can identify the graph in O(n2) but if thats not give its a NP complete problem.

Finally i realised that they are just testing for basic skills ... I mean if these guys really wanted to screw me .. they could easily have.

Finally .. today i got the offer. Itz good. I am joining the filesystem group .. most probabbly the AIX/Linux team. This group has arround 35 dev members and most of them Phd. So wish me luck.

Veritas Interview Part I

2005-10-17T02:58:00.000-07:00

This is bloody cool .. believe me ... And this time i didnt have to use "Towers of Honoi" Trick

Interview -

Q1: So you have worked with filesystems, tell me about the work.
Ans: I have just sent out patches for Unionfs, a vfs-layer fanout file-system. But thats my hobby My work at Ensim is in system level.

Q2. What entry points into the VFS you would have to trap, to handle this situation.
And: Everything including open, create, readdir .. etc. The only execeptions are sendfile and mmap and some others.

Q3. Can't you just pass the lower level dentries to mounted filesystem.
And: Its a union so you have to union directory entries hence create new interposed inodes and dentries. And Also coz you have to maintain different caches at different levels. The the files you craete/delete over the union are not there on the lower branches. And you don't want writes to the lower branches to affect the union.

Q4. What could be the implecations of nfs over unionfs. Tell me how mmap works over NFS.
Ans: mmap i am not sure, But in general NFS relies upon persistant inodes. i.e the inode numbers should be same over remounts. Since unionfs creates vnodes on the fly it can't ensure same inode numbers over the remounts. Although I had thought of a patch for computing inode numbers as a hash of the file path. but it would be a problem for hardlinks.

Q5. How would you write a disk defragment software for ext3. Are there any benifits in doing that.
Ans: It would be good for any filesystem to have disk defragmentation as it brings all the file blocks together. It saves -
1. reading time
2. inode table space. As inodes would not be dublicated over multiple inode blocks.

Algo:
1. Lock superblock
2. Aquire semaphore over inode.
2. Everytime you move the blocks ensure that file cache is distoyed. This can be done by increasing the superblock generation number. or otherwise...
3. Update inode tables
4. Release locks.

Q6. How does a fs say ext3 serializes the read/write on same byte over same file.
Ans: In uniprocessor it already serialized. In SMP I guess the bus controller ensures that. I think so ....

Q7. Whats the difference between Semaphores and Spin Locks ?
Semaphores sleep while waiting. Spin locks spin in a while(1) loop. Spin locks make no sence in uniprocessor as there is only on thread running.

In SMP .. you cant implement semaphores without using spin locks. In SMP data-structure of semaphore itself becomes a critical section. ..You have to ensure local interupts are disabled and SMP waits over spin lock.

There is an overhead in semaphores of acquiring a spin lock and then enqueuing the task struct and the dequeuing it. Better use a spin lock if you know wait would not be much.

Q8. In which parts of the kernel code you can't use semaphores.
Ans: Interrupt handlers. They obviously can't sleep over locks.

Q9. In SMP environment would the CPU cache cause a problem in spin locks. How would you avoid it.
Ans: Cache in L1/L2 may become dirty in one CPU. So other CPU must be informed about it. x86 uses interprocessor interrupts to flush the cache. I think c0 control register is used for it.

Typical use case is after using a spinlock. The spinning condition is almost always in cache and a spinlock changes the value for which the other CPU is waiting/spinning for.

Q10. How would you make sure the instructions are serialized and there is no optimization by the compiler/CPU over the instructions.
Ans: Use something called memory barriers. In x86 use a prefix instruction called "lock". Which requests the CPU to do all thats written before this point before executing the instruction.

In compiler you could append volatile before asm qualifier.

Q11. How do you insure such things in C. Give me an example of a possible scenario.
Ans: Same volatile. Typical use case is when a global variable changes in signal handler. If that variable doesn't directly change in the function. The compiler makes it constant i.e reads the variable value from the memory and copies to the DS (Data segment) and never refers back to the changed value.

Putting a volatile qualifier tells gcc not to make such assumptions.

Q12. Do you know about itanium architechture? In a project in veritas, We had problems with spin locks over different itanium CPUs in same machine. One lock was increasing the value and the other decreasing the value. How would you avoid race conditions in this case. Are there any instructions x86/ia64 insures for this.
Ans: x86 gives atomic instructions to test_and_set the value in the memory. I guess BTS, BTC inctrutions can work well over SMP... and also those 'X' instruction for byte manipulation like CMPXCHG etc.

Q13. How POSIX recomends locks in system level. Tell me the rules.
Ans: Posix recomends mandatory and advisory locks. In mandatory locks every system call has to wait for a lock (if it is acquired). Generally syscall send a signal to the lock owner and waits for a lease time. After that they forcefully take the lock.

Q14. Explain segmentation in Linux wrt to x86.
Ans: Linux doesn't use segmentation of x86. The 16 segment selectors (DS,SS, ES) all point to GDT tables with have initialized the segment descripter with base addreess as 0x0000

Q15. Explain paging with respect to x86.
Ans: Explained.

Q16. Can I do everything in software or hardware support is necessary.
Ans: Everything but the page cache misses. (Answered after a lot of struggle)

Q17. We wanted to have more than 8KB serialized physical memory allocation for a process in Linux. Is it easy/difficult.
Ans: Sorry :D

Q18. I have a machine with infinite RAM. Do i still near virtual memory.
Ans: Yes, to have on demand pagging.

Q19. And if don't i want on demand fecthing.
Ans: When you would have to statically compile all the code. There would be no dinamic compilation.

Q20. Have you heard about. Paging extentions in x86 (P4 onwards). Explain.
Ans: Fortunately yes. You basically use a PDTR to address 64 GB of ram with page size as 2MB. P4 gives 36 address lines inspite of 32 in normal case.

Q21. Will that have good bad effect on TLB hits.
Ans: (With lots of struggle) TLB will improve as it can adress moe mem.

Q22. I have an adjacency list of a graph. And for each node i want to sort the neighbours.
1 -> 5 , 3 , 2
2 -> 1, 4, 3
3 -> 2 , 1 , 5

becomes ..
1-> 2, 3, 5
2 -> 1, 3, 4
3-> 1, 2, 5
Solve.

Ans: (With lotz of struggle) After sorting the first list you can make out that if
1-> 2, 3, 5 then all 2,3,&5 will have first element as 1. And so on.

Q23. I have overlapping line segments and then i give you a point. You gotto tell me how many lines would have this point.

Ans: Used R-trees ( the guy gave some hints )

Just waiting for results now .... :D

Dream ...

2005-05-30T07:57:00.000-07:00

The child is grown, the dream is gone.
I have become comfortably numb...

My dad is serving Indian Air Force ... I look upon him with great respect. As i child i also cherished a dream .. a dream to join the forces, but....

.

Now, What I want ... is i guess what every man wants - Everything.

<-- Incomplete -->

Dogde This

2005-05-25T02:16:00.000-07:00

Sometime back I use to like the TOI caption - "No preaching. Just News"... I liked that. So I am trying to blog a unaltered and unbiased opion of the movie - "Matrix".

If nothing else .. i kinda liked the names and dialogues ..

dodge this ....

I liked the way .. characters were named in matrix. All the characters in the movie (except Tank, mouse and Dozer) were children of the matrix, they didn't have any real world names. So (in the story) probably they themselves were responsible for these symbolic names.

Some quick googling showed gave me these results -

1. Neo - Greek word `Neo` meaning New. Anagram of the word "One". Interesting

2. Morpheus - God of Dreams (Greek Phil.), son of Thanatos (god of Death) and the brother of Hypnos (god of Sleep).

3. Trinity - The holy Trinity, from the bible (Father , son and the holy ghost)

4. Oracle - Pretty straight forward .. a prophet/authoritative person who divines the future

5. Zion - A holy city , place (Jewish belief)

6. Merovingian - The Merovingians were a royal bloodline that ruled from early on after Christ and carried through their rulership of southern and central Europe(now France), until 680 AD. They claim direct kinship with Christ through his bloodline. Interesting ...

7. Persephone - Greek mythology; Daughter of Demeter (I think), grabbed by Hades and forced to live in the underworld... Kind of how the Merovingian has a hold over her until she defies him.

8. Epoch - Is time(UTC) since January 1st, 1970. Atleast thats what time.h (glibc2) says.

9. Niobe - In shakespearean texts Nirobi/Niobe was a symbol of grief, she lost 13 children. There is another myth to do with human arrogance, Noibe dares to claim herself to be better than a goddess. She was married to Amphion son of Zeus. Niobe had twelve children with Amphion and when boasting about them to Leto who only had two, brought the wrath of Apollo down on her and he punished this insult to his mother by shooting with arrows all of Niobe's children.

10. Nebuchadnezzar - He was a Babylonian ruler. The Hanging Gardens of Babylon were built by King Nebuchadnezzar II about 600 BC. The garden was his palace basically.

Nebuchadnezzar was a king who had a dream inspired by god which showed him a giant which was composed in 4 different metals gold, iron etc. then, the giant exploded, and a mountain grow up from it,taking all over the world. that dream was supposedly an indication of the 4 kindoms that ruled the world (last one was romans) .it is said that there won't be anymore kingdom trying to take over the world before the apocalypse.

Where others see Coincidence,
I see Consequence.

If you know something more ... put in the comments.

Information Security

2004-12-11T23:20:00.000-08:00

Have you ever got this feeling of being *watched* at an ATM, Airport or a Govt Building. If not try this ... generally all ATMs have (convex-)mirrors installed on top of the machines ( ICICI ATMs always have round convex mirror with slightly-curved triangular wooden frame). Why do they have to be convex ??? Well, put your finger on them .. and you would see "No" gap between your finger and its image. Yes, it means exactly what you are thinking. You been watched.

The govt says ..it has a right to interrogate ..the right of information ..the right to voyeur !!
"Well then who decides .. Who keeps all the information" - Kevin Mitnick.

The Orange Book

All those interested in aspects of info-security this book (or rather a draft) published in 1983 - "The Orange Book" is a must read. This book "defined" security and its policies in true sense. The main idea was that the security should not be left on the shoulders of application writers and its users, but governed by a central security policy. It should take care of both .. vulnerabilities in ptrace (left open by application writer) and ignorant users .. leaving faulty permissions here and there ( leaving a ignorant vunerabilty as it is called).

M$ widows security and ACL model is also inspired from similiar ideas .. And it would totally unjust for the linux fanatics to say that .. its not secure. M$ windows is just a bad implementation of a good design. Who can deny the micro-kernel architechture if WinNT and there ACL based security model.

The best thing about linux is that it doesn't take/includes anything in its face value. Take for example Solaris .. opensolaris code could easily show you have they have nt left any single complex algorithm in their design. Be it there scheduler or VFS. But even back in 1994 Linux 1.0 was atleast 3 times faster than Solaris.

On the contrary linux (atlest the kernel .. glibc sucks at times ) include in its code only whats works and works efficiently.

NSA - SELinux

Based on the "Orange Book" guidelines NSA designed the - "SELinux - Security Enhanced Linux". Some moron mailed on the NSA SELinux mailing list asking if NSA people are also integrating biometrics in SELinux ... The answer was very clear "SELInux is for authorization and not authentication".

Authentication deals with "*Who and When* should a person should be allowed to enter a system.
Authorization on the other hand answers - "What privileges should be given to the Subject just logged in".

SELinux ... takes on the system after PAM is done with the work. It examines not the subject but its actions on objects in system. Objects could be a file .. or even some other subject. It is a very good implementation of DTE (Domain Type Enforcement) and RBAC (Role Based Access Control) both described in The Orange book.

DTE understands only domains (Processes) and Types (Objects like files.) and defines rules as to which domain it allowed to act-upon a particular type. RBAC on the contrary understands only users and the roles which they can take up.

So, here is the complete pic.

User -> Role 1(admin) AND Role 2(apache)

Role 1 (admin) --- can enter domain --- Admin --- can read or exec
Role 2 (apache) --- can enter domain --- Apache -- can exec --- type 1 (binary files)
Role 2 (apache) --- can enter domain --- Apache -- can read --- type 3 (user files)

But to the very nature of centralization .. it is prone to disasters. One wrong policy can blow off the entire system.

And you though your system is secure.

About Me

2004-10-23T18:46:00.000-07:00

herez a FAQ about me ..

Name : Jaspreet Singh
Age : 23 (when i wrote this )
Sex : Personal Qus sorry
Profesion: Geek ( employed )
Religion : Geek Orthodox
Birthday : April 12, `82
Sexual Orientation : Not discovered yet !!
Lisense : GPL v2.0

what do you do ?? .. i live
what do you do for living ?? i work as software engineer for Veritas.
what insterests you ?? Open source , Linux , Music and Movies
how do you look like ?? earch cracking material
where do you see yourself after 10 years ?? .. doing scoba diving at bora-bora island.
why this change in blog ?? posts about my non-geek self is on The Blue Pill . The non hierarchical namespace of blogger has forced me to do so ... This blog would only be for geeky stuff.

Dont walk as if you own the world ..
Walk as if you dont care who owns it.
Jaspreet Singh