I think you greatly overestimate what "seeing" means. They check the code for coding errors and functionality. They don't check it for vulnerabilities unless they are very obvious. They definitely don't check it for deliberately inserted vulnerabilities and backdoors, they have neither the resources nor the time.
If you say that people do check FOSS code for errors and functionality but no one tests for security then I agree with you. I guess no one puts the effort to check the vulnerabilities of said software as it is not is "responsibility" but I have to say again in the case of linux I find that hard to believe given how many fortune 500 companies use that software but none would put the effort to check the security. Stupid, isn't it?
And this code is submitted by people all over the world. Many if not most anonymous. Many what the Linux kernel maintainers themselves refer to as "drive-by contributors" who provide a single blurb of code to fix a bug or create a solution, then don't contribute again. Some contributors are teams associated with colleges. Which means that the people in them change all the time. It should be trivial for some foreign government hackers to set up a relationship with kernel maintainers using some state run college as a front end, keep submitting good code of high quality for a couple years, then carefully inject backdoors in one of the submissions.
As a none coder, those submissions are not reviewed before implemented in the source code?
I don't think you understand the difference between writing code, and securing it. Writhing is where the bugs and vulnerabilities are introduced. Bugs manifest themselves by unexpected code performance. Vulnerabilities are sitting there silently waiting to be discovered.
All of these developers are writing code. Mountains of it. Over 70 million lines of kernel code alone.
How many are securing the code that has already been written by others ? It's a painstakingly hard, boring, thankless work.
You are correct. i do not know. I just assumed there is some sort of "best practice" in writing code to make it none vulnerable. The code – I thought – was checked by project maintainers before if it adheres to those "best practices" . If you are thinking "this guy does not know what he is talking about" you are not wrong, I am just discussing my POV to my limited knowledge.
Of course they are. But not to fix them. A backdoor in a major OS is a goldmine for government agencies. The US government is sitting on many hacking tools and code it will not disclose, and will only quietly fix if the benefit of doing this outweighs the benefit of having access
There's a thriving market for selling code vulnerabilities.
Are you saying there are companies that did discover the vulnerability, and will sell that to payers? thats vile! Shouldn't this be illegal? its like selling a device to bypass metal detectors in an airpot. Linux foundation can buy this info though, and fix the vulnerabilities, I hope they have enough money from all those sponsors.
There's also a thriving commercial industry built around securing Windows, because that's where the money is.
So Windows is similarly unsecure ? Someone found the vulnerabilities in Windows?
There's no such industry built around securing Linux. The US gov't is trying to fix that.
I thought you said above the US gov. knows the vulnerabilities and willing to keep it that way for their advantage and not to fix the vulnerabilities in Linux.
Windows is not "closed source" in the sense that nobody but a few select MS employees see the code. They have bounty programs and have been partnering with security researchers for two decades now to find and fix bugs. They pay them to do it full time. If you have the right qualifications, you can apply, sign NDA, and get access.
And yes, Windows, despite having control over who contributes to the code, having internal QA procedures, and paying specialists to work full-time to secure the code, still has vulnerabilities. The simple logic says that Linux, which has little control over who contributes the code, doesn't have a team of highly paid security researchers, and is an extremely fragmented field with gazillion distros, should have far more of them.
And if there was a professional test to find vulnerabilities, there wouldn't be many. You can compare the numbers of vulnerabilities found and published, but it's not a good comparison since there's not a uniform approach to that.
Well noted. Makes sense.
As I said many times, "many eyes on the code" is a fallacy. Show me how many eyes have actually inspected the code. Have you looked at Linux kernel code ? Do you personally know anyone who is a qualified security researcher and inspected Linux kernel, or drivers, or programs ? Writing code is fun, securing code is boring.
Honestly, no, but I assumed Linus and the guys are doing that work.