I'm trying to manipulate file from kernel module that is loaded at boot time (late_initcall() ).

yes i know that access to 'user space' from 'kernel space' is not the beast practice, but i have no choice.

for now im checking if the file exist...

However, if the filesystem is not mounted yet, so the check is not valid.
also i check if i can read /home but it's not beast practice as well

Is there a way to check if the F.S is mounted by the time the module is call?

The model can wait for the filesystem but, what is the right way to check that?

except of reading or writing a file.

Thanks

Take a look at the mountpoint command. It's normally used in bash script with an if statement:
See man mountpoint for more details on this command.

Here is a command sequence that will mount a partition and give an error message and quit if the mount command failed.


# Mount the usb hard drive
if ! (mount -t ext4 -L Memorex.2141 /media/usb.Memorex.2141);
then
echo "mount failed for usb thumb drive Memorex.2141"
exit
fi

Hi thanks for the answer.
but there is no bash shell or whats ever in kernel module at boot time.
it's pure c language.
so whats I'm trying to find out. is, what struct (in c language, kernel space ) is valid for check that mount is happened (rootfs).
Such as 'struct super_block' that located at linux/fs.h

p.s the file is located on the rootfs at /my_folder/my_file

Redesign it.

Expose your updates via /sys and provide a userspace program to do the update to the actual file. /proc/mounts might be populated early enough for you - but if you can wait, it doesn't matter. For fstab based mountpoints, you'll have to wait.