fread() on a bad disk?

Quote:

---

Originally Posted by brewbuck

Then you need to write the code correctly, using non-blocking reads and such. The OS is capable of telling you when a disk is pulled, learn how to do that.

---

very hostile.....thanks for the help?

Quote:

---

Originally Posted by annied

very hostile.....thanks for the help?

---

Sorry, didn't mean that to be hostile at all. I'm just saying that doing blocking reads is not the right way to do this.

Quote:

---

Originally Posted by brewbuck

Sorry, didn't mean that to be hostile at all. I'm just saying that doing blocking reads is not the right way to do this.

---

My apologizes, guess that's why email/internet is tough to hear tone in someone's voice. :)

Thanks for your help, I'll try to figure out the better way to do this.

Quote:

---

Originally Posted by annied

My apologizes, guess that's why email/internet is tough to hear tone in someone's voice. :)

Thanks for your help, I'll try to figure out the better way to do this.

---

Id' suggest, for starters, to use fcntl on the open fd to put it into a non-blocking mode. And investigate how to get disk status from the kernel. Probably via the /proc filesystem somehow. Been a while since I wrote that sort of code.

Quote:

---

Originally Posted by brewbuck

Id' suggest, for starters, to use fcntl on the open fd to put it into a non-blocking mode. And investigate how to get disk status from the kernel. Probably via the /proc filesystem somehow. Been a while since I wrote that sort of code.

---

thanks much for the pointers. Hopefully I can report back with what worked.

I would suggest you look at using open() and read(), as opposed to the standard fopen/fread approach. You'll have a lot more scope for controlling what is going on, with what is after all rather system specific to begin with.

Quote:

---

Originally Posted by Salem

I would suggest you look at using open() and read(), as opposed to the standard fopen/fread approach. You'll have a lot more scope for controlling what is going on, with what is after all rather system specific to begin with.

---

I had tried an open() (read-only) and read() and got the same result....

You also need to put the stream in non-blocking mode. Salem's point is that this is impossible in the first place with fread.

Quote:

---

Originally Posted by CornedBee

You also need to put the stream in non-blocking mode. Salem's point is that this is impossible in the first place with fread.

---

any idea how I would go about putting it in a non-blocking mode?
I've been looking around but haven't had much success....

Passing O_NONBLOCK in mode to open() sounds like an option.

As already mentioned - http://www.rt.com/man/fcntl.2.html

Quote:

---

Originally Posted by CornedBee

Passing O_NONBLOCK in mode to open() sounds like an option.

---

was just looking at that, going to try it but been reading that its not recommended...thanks

tried with the open()/read() and the open with O_NONBLOCK but no luck.
wondering if adding O_NDELAY would help.

You could start another thread to monitor the disks or something? Very OS specific, but so is your problem.

Quote:

---

Originally Posted by Salem

As already mentioned - http://www.rt.com/man/fcntl.2.html

---

tried with fcntl(), I got passed the read() but then hang on the close(fd).

I'm sure I missed something here. Can someone have a look?
Not that familiar with fcntl()

Code:

---


        int            fd_flags = 0;

        fd = open(fullpath, O_RDONLY|O_NONBLOCK|O_NDELAY);
        if (!fd) {
                fprintf(stderr, "ERROR: Can't open device %s
\n",
                        fullpath);
                return 0;
        }
        fd_flags = fcntl(fd, F_GETFL, 0);
        fcntl(fd, F_SETFL, fd_flags | O_NONBLOCK);

        ret = read(fd, buffer, sizeof(buffer));
        close(fd);

---