Basic Working difference between Windows Copy and linux copy

[Prasanjit Shome]

can anyone tell me what is the basic working difference between a Windows copy and linux copy ?

Thanxs in Advance

[Steven E. Protter]

Shalom,

On the pure OS level, its not that different. Its just i/o at a certain block size.

What skews performance data however is the GUI windows has to support and work with. Linux overhead is lower so generally things like file copy will work more quickly. There are of course exceptions to every general rule.

SEP

[Alzhy]

You will need to expound on what you mean by Windows Copy sir. Because there are multiple ways Copies (file/tree) copies are done on Windows and there could be some basic differences between Linux and Windows.

Windows Explorer (aka File Manager or File Explorer) copies in the fashion of Copy/Cut and Paste could be similar to a Linux GUI aided copy (via Nautilus or KDE's File Manager or equiv).

On the Windows Command Line (aka DOS or CMD32/64) - you have COPY.EXE and XCOPY.EXE. On Linux Command Line - you have a "cp" which performs the functionaility of both COPY.EXE and XCOPY.EXE. You also have "dd" for copying large files and which could be faster as it can do Direct IO (cacheless wriets). Both cp, XCOPY/COPY.EXE likely use the same basic tenets of FileIO. For tree/recursive copies -- both likely uses arrays to recurse the copy and which could have the same limitations in dealing with so called gazillion nested trees.

So there ya go - basic working differences between the 2 platforms.

And Shalom.

[dirk dierickx]

how many ways are there to copy a file? the process will be much the same between any platform.

things that make the difference are the filesystem used, the OS IO layer, which are ofcourse different and contribute to a difference in performance (but not in the actual copy process itself).

[TwoProc]

By Windows copy, he means at most a copy command executed via Windows software command, on a Windows file system. At the very least, he means a copy command on a Windows file system. Not that it matters that much, if it is on a Windows file system, whatever method the copy, underneath the covers, you're stilling using Windows OS file I/O primitives to get the work done (yeah, I wanted to say DOS at this point, but I held back).

:-)

[Alzhy]

TwoProc,

I dont think Windows Explorer (aka File Explorer)
copies" on windows use the same "primitives" as the DOS/CMD32/CMD64 - COPY.EXE or XCOPY.EXE.


Just like Nautilus or FIle manager on UNIX/Linux - GUI driven file copies could be different IMHO, I still remember my memory use increasing when using a GUI copy under Wndows or UNIX/Linux GUI. (the case of those gazillionm files and trees filesystems).

[bullz]

haa haa i like the quest.

If i am not wrong, windows will not check for free space available before copying data.

i am not sure...... :(

[Lucifer Megacruel]

Hi,

Windows generally used to use the FAT file system , ie FAT file system is very easy to implement and understand. A windows disk looks like this (Taking the example of a floppy since it is the easiest to understand )

[Boot Block][Fat table1][Fat table2][Root Directory] [ Rest of Data]

The boot block is 512 bytes in length and should have a signature 0x55aa (remember intel x86 is little endian)

a boot sector would look some thing like things
org 0x7c00
jmp start

; BIOS Parameter block

start:
; enable a20 line
; set up a dummy gdt
; and enter protected
; read few clusters using bios 13 h service ( basically load the loader into memory)
; jump to the os loader code

times db 512-($-$$) 0 ; Pad rest of the bytes with zero

dw 0xaa55

BPB contains media related information. Each file has a entry in the root directory area called the directory entry. The directory entry in contains a field called the first cluster number
FatTable1[cluster_number] will give the next cluster number.

So to copy a file , you basically need to create a directory entry and set the fat table accordingly with new set of clusters containing the duplicate data of the first file :) .


Linux supports many file systems as long as you implement the vfs interface you can implement your own fs as well. Which FS are you talking about here ?

Truly Evil
Lucifer Megacruel

[Alzhy]

Lucifer Megacruel.

Truly Evil.

Lol! ;^))

You made my day man!

[Lucifer Megacruel]

Just for fun Alzhy :)