Question about flashing nand image on qcom arm

[mbanders]

Someone has suggested that we should use $filesize for the line nand write in the function below.

They suspected (but didn't prove) that if you do nand write larger than the file size, then it will be filled with junk. Does that sound possible?

Maybe $filesize was tried for a while, but wasn't as good as using that calculated hsize?

    def nand_flash_bin(self, addr, size, src):
        # make sure we round writes up to the next sector size
        hsize = hex((((int(size, 0) - 1) / self.flash_block_size) + 1) * self.flash_block_size)

        self.sendline("nand erase %s %s" % (addr, size))
        self.expect("OK", timeout=90)
        self.expect(self.uprompt)
        self.sendline("nand write $fileaddr %s %s" % (addr, hsize))
        self.expect("OK", timeout=90)
        self.expect(self.uprompt)

[mattsm]

The write could probably use the actual size. The erase probably need to erase the entire block because some fs will scan the entire block and give up if it sees old data. Although I'd think garbage could have the same problem.