Re: [LTP] [PATCH] mmap/munmap /dev/zero

[CAI Q. <ca...@re...>]

----- "Mike Frysinger" <va...@ge...> wrote:

> On Tuesday, October 12, 2010 22:06:29 CAI Qian wrote:
> > ----- "Mike Frysinger" wrote:
> > > On Tuesday, October 12, 2010 10:07:16 CAI Qian wrote:
> > > > +	 x = mmap("/dev/zero", SIZE+SIZE-4096, PROT_READ|PROT_WRITE,
> > > > +		 MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
> > > 
> > > uhh, have you even checked this test ?  this mmap() makes
> absolutely
> > > no sense on so many levels.
> > 
> > Can you elaborate?
> 
> so you havent actually looked at the syscalls made by the application
> and the 
> memory maps created nor used mmap() in detail before ?  well, let's
> start with 
> the mmap() man page:
> 
> 	void *mmap(void *addr, size_t length, int prot, int flags,
> 		int fd, off_t offset);
> 
> "addr" doesnt sound like "file name" to me, and the description
> agrees:
> 
> 	If addr is not NULL, then the kernel takes it as a hint about where
> to
> 	place the mapping; on Linux, the mapping will be created at a nearby
> page
> 	boundary.  The address of the new mapping is returned as the result
> of
> 	the call.
> 
> so basically you told the kernel to create a mapping in/near your
> .rodata 
> section (since that is the address of the constant string
> "/dev/zero").  the 
> contents of that pointer (which is not of type "char *") have
> absolutely no 
> meaning to the kernel.
> 
> then there are the flags you're using:
> 
> 	MAP_ANONYMOUS
> 		The mapping is not backed by any file; its contents are initialized
> 		to zero.  The fd and offset arguments are ignored; however, some
> 		implementations require fd to be -1 if MAP_ANONYMOUS (or MAP_ANON)
> 		is specified, and portable applications should ensure this.
> 
> so not only are you not getting a mapping from /dev/zero, you're just
> 
> allocating a random piece of memory that the kernel has guaranteed
> will be 
> zeroed out for you.  so any attempts to verify the contents are zero
> *because 
> the data came from /dev/zero* fail.
> 
> finally, there's the fd which you're passing as "-1":
> 
> 	The contents of a file mapping (as opposed to an anonymous mapping;
> see
> 	MAP_ANONYMOUS below), are initialized using length bytes starting at
> 	offset "offset" in the file (or other object) referred to by the
> file
> 	descriptor "fd".
> 
> "-1" is not a valid fd, so there's no way it could possibly be from
> /dev/zero.
> 
> so we're back where we started: this test makes no sense as written
> and you 
> apparently havent verified it "works" beyond "is the exit status 0
> when it 
> finished".
I see. MAP_ANONYMOUS is not needed. It certainly want to use int fd. v3 followed.
> 
> > > why are you using _exit() ?
> > 
> > It is used to terminate a child.
> 
> i meant why arent you using exit() ?  why do you need _exit()
> semantics ?
I don't think that too much. I think it can use exit() too.
> -mike