Hacker News new | ask | show | jobs
by akerl_ 46 days ago
Again, the current flow was that the researchers only reported the vuln to kernel devs, who then committed a fix with no flag that it was a security fix.

A proposal where there is an unmarked commit and then anybody tells anybody anything about the fix including a security remediation is strictly more information being disclosed into the world.

Also you’re just wrong about the ability of bad actors to identify vuln remediations (and consequently vulns) by looking at commits to major projects. I don’t know what else to say here other than that this is happening, and is easily attainable via the current combination of human expertise and automated tools.

1 comments

> Again, the current flow was that the researchers only reported the vuln to kernel devs

That's what happened in this case, but the current idea/expectation (according to what was linked in the comment chain I replied to) seems to be that the researcher would email the distro maintainers with information:

> > Notify security@kernel.org, linux-distros@vs.openwall.org and relevant maintainers of the vulnerability; establishing details, embargo period, CVE request and possible fix

This is the process I'm suggesting could make more sense if it was instead the kernel maintainers alerting distro maintainers (with no more detail than necessary) after a fix has made it in. Should also be less fallible than relying on the researcher to do so.

> Also you’re just wrong about the ability of bad actors to identify vuln remediations (and consequently vulns) by looking at commits to major projects. I don’t know what else to say here other than that this is happening, and is easily attainable via the current combination of human expertise and automated tools

I don't deny that bad actors can figure out vulnerabilities from tracking commits, but removing or refactoring some subsystem does not immediately give all bad actors a full list of vulnerabilities with the old version. Developing an abusable exploit chain (as may be shown by the researcher to justify high priority patching) is also not necessarily trivial even if they do figure out an issue that was fixed.