unfortunately these are not the topic to digest in a mail, I recommend you reading UTLK book (http://shop.oreilly.com/product/9780596005658.do). Still...

will it be maped with vm_area struct ? Yes if it is accessed via mmap system call.

what is the relation between page-cache and file operation? file operations for data access like read/write will look into page-cache first before going to disk.

-Rajat On Thu, Jan 3, 2013 at 1:01 PM, horseriver <horserivers@gmail.com> wrote:

On Thu, Jan 03, 2013 at 12:48:01PM +0530, Rajat Sharma wrote:

...

Never heard of page-cache?

will it be maped with vm_area struct ? what is the relation between page-cache and file operation?

thanks!

...

On Thu, Jan 3, 2013 at 12:29 PM, horseriver <horserivers@gmail.com>

wrote:

...

...

hi:

when one file is opened , does its data be put into memory ? and all operation on this file

will be implemented by operation on its mapping memory area ?

thanks!

_______________________________________________ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

_______________________________________________ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

in this vm_operations_struct , there are open/close functions , are there necessary relations between file operations and this struct ? well not really for open/close of vm_ops are of interest to filesystems, but page fault handler and making page writable is where filesystem come into picture. Have a look at ext4_file_vm_ops, it implements operations of interest.

static const struct vm_operations_struct ext4_file_vm_ops = { .fault = filemap_fault, .page_mkwrite = ext4_page_mkwrite, } Note that only filesystem knows how to fill up this page. On Thu, Jan 3, 2013 at 3:39 PM, horseriver <horserivers@gmail.com> wrote:

On Thu, Jan 03, 2013 at 01:16:06PM +0530, Rajat Sharma wrote:

...

...

will it be maped with vm_area struct ? Yes if it is accessed via mmap system call.

you know that , in the struct vm_area_struct,there is a struct vm_operations_struct * vm_ops;

in this vm_operations_struct , there are open/close functions , are there necessary relations between file

operations and this struct ?

thanks!

...

...

what is the relation between page-cache and file operation? file operations for data access like read/write will look into page-cache first before going to disk.

-Rajat

_______________________________________________ Kernelnewbies mailing list Kernelnewbies@kernelnewbies.org http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies

Default read/write inerfaces are better suited for sequential read/write within your program. Although you can seek to any location within the file, you still have overhead to issue system calls to get data. However mmap allows you to map a section of file into program address space. Now if your access patter is rather random, modifying few bytes here and there, but at random offset. You just get continous memory array, which is much easier than issuing read and write at different file offsets. A most common and mandatory use case of mmap is in mapping executable binary program image, and libraries into process address spaces. Access pattern for a program is not sequential, you can have multiple jump (if, else, for loop), so it is better suited with mmap. It is read only and private memory mapping, Any modifications you do will create a COW page which if private to your process, so that is another advantage of mmap which is completely transparent to user mode. If your filesystem does not support, this basic mmap mode, you can not execute a binary file stored in this filesystem, unless you copy it to some other filesystem which does. Apart from this, read more about mmap from UTLK book. -Rajat On Fri, Jan 11, 2013 at 2:44 AM, horseriver <horserivers@gmail.com> wrote:

hi:

these two wayes of operating one file :

1.use open/write interface call .

2.mmap this file into memory , then access this memory area and do r/w .

what is the essential difference between this teo wayes?

thanks!

Default read/write inerfaces does not move file's data to process address space ? Yes it does, in either case, user space memory has to be in process address space. But Difference is in the access pattern. With read/write, you demand for it explicitly through a system call, hence application is more involved here. While for mmap access, all transfer happens in application-unaware mechanism, with page-fault handlers inside kernel. Application just access mapped buffers like memory array and magic happens inside kernel as you keep on accessing bytes.

Best way to find out difference is to try out writing a simple mmap program. -Rajat On Fri, Jan 11, 2013 at 3:31 AM, horseriver <horserivers@gmail.com> wrote:

On Fri, Jan 11, 2013 at 12:39:26PM +0530, Rajat Sharma wrote:

...

Default read/write inerfaces are better suited for sequential read/write within your program. Although you can seek to any location within the file, you still have overhead to issue system calls to get data. However mmap allows you to map a section of file into program address space.

Default read/write inerfaces does not move file's data to process address space ?

when r/w a file descript which returnd by open() , how do the file data move from one place to another place ?

For each time the write function being called , will kernel call filesystem's driver's write to respond ?? In my opinion,kernel will passed a buffer's head address which is passed form user-layer into driver,then driver will fill this buffer with file's data which is got by filesystem's read operation ?

Am I right?

Thanks!

...

-Rajat

On Fri, Jan 11, 2013 at 2:44 AM, horseriver <horserivers@gmail.com>

wrote:

...

...

hi:

these two wayes of operating one file :

1.use open/write interface call .

2.mmap this file into memory , then access this memory area and do

r/w .

...

...

what is the essential difference between this teo wayes?

thanks!