Zack's Kernel News

Protecting Future Feet

When compiling software of a certain size, you typically compile a bunch of smaller files and then link the results together into the final executable file. This is a simple and easy process that definitely contains no psychotic nightmare lunacy whatsoever – except for most of it.

One aspect of linking involves assigning memory addresses to all the functions and variables and whatnot that are perhaps defined in one file but need to be referenced by code in another file. It gets even more fun when you want to do this while simultaneously hiding these addresses from potentially hostile code that might be running on the same system, looking for security holes. But that's another story.

A while back, Linus Torvalds stubbed his toe on the linker when he was struggling to identify the start and end of particular sections of code to be linked so that he could store certain data in some sections and other data in other sections. After Linus kicked that brick wall for a while, Rasmus Villemoes explained to him that there was a simple and automatic way to do what he wanted: "If an output section's name is the same as the input section's name and is representable as a C identifier, then the linker will automatically [provide] two symbols: __start_SECNAME and __stop_SECNAME, where SECNAME is the name of the section. These indicate the start address and end address of the output section respectively. Note: most section names are not representable as C identifiers because they contain a '.' character." Linus had a big laugh at himself over that one, because all his struggles could be solved with a single character. But that's yet another story.

[...]