<html><head></head><body><div>Hi,</div><div><br></div><div>On Tue, 2022-01-25 at 07:05 +0000, peter green wrote:</div><blockquote type="cite" style="margin:0 0 0 .8ex; border-left:2px #729fcf solid;padding-left:1ex"><div>On 25/01/2022 06:14, David Bannon wrote:<br></div><blockquote type="cite" style="margin:0 0 0 .8ex; border-left:2px #729fcf solid;padding-left:1ex"><div>that would also cover the situation that now applies to eg x86-64 and Arm too where hardening does not work with a statically linked binary, you need to manually force it to be a dynamic link first.<br></div><div><br></div><div>Your question ? Personally I see little benifit in hardening on a single user, private system. But agree that its a very good thing on what we generally call a server. We should be able to do it ! <br></div></blockquote><div> ><br></div><div>To me it's less about the system and more about the program. There are two key questions.<br></div><div><br></div><div>1. To what extent is the program used to process untrusted data. The bottom line with compilers<br></div><div>and related tools is that most of the time people use them on a codebase they plan to execute,<br></div><div>so there is little to be gained by attacking them.<br></div><div><br></div><div>2. To what extent does the language and programming style help avoid the kind of screwups that<br></div><div>lead to hardening being created in the first place. I'd say in this regard borland style pascal<br></div><div>is better than C, possiblly slightly worse than modern C++, much worse than rust.<br></div><div><br></div><div>How many pascal programs in Debian are there that do not link against the (dynamic) C library<br></div><div>for one reason or another *and* are likely to be used to process untrusted data?<br></div><div>Is the inability to harden static binaries really that big a deal?</div></blockquote><div>Personally, I'd prefer robustness against performance, but have no clue what kind of attacks we may encounter if lacking hardening.</div><div>I don't have time, for now, to query this, and taking into accounts remarks and questions above, I would say let's keep this as is?</div><div>No override, live with it until we get a clear decision on whether we should do it or not. It will not matter too much as anyway there are less than 1k systems installing FPC, so not a nice target for an attacker.<br class="Apple-interchange-newline">-- <br></div><pre style="caret-color: rgb(46, 52, 54); color: rgb(46, 52, 54); font-variant-caps: normal; letter-spacing: normal; orphans: auto; text-align: start; text-indent: 0px; text-transform: none; widows: auto; word-spacing: 0px; -webkit-tap-highlight-color: rgba(0, 0, 0, 0.4); -webkit-text-stroke-width: 0px; text-decoration: none;">Cheers,
Abou Al Montacir</pre></body></html>