Benchmarking the reentrant result showed it to be around 5% slower.
Now I'm trying to redo it again but this time scripting the refactoring using sparse https://github.com/lucvoo/sparse to parse and using it's error messages with with line/column to guide the refactoring, I already got an initial script that performs some initial transformations and is repeatable, but more work need to be done, mainly enhance/extend the info that sparse provide while parsing the code.
gdb -batch -ex "info variables" -ex quit --args binary-to-examine
Also, until very recently, a lot compilers/platforms were unable to handle thread_local variables larger than a pointer size making it difficult to retrofit a lot of old code.
* ideally, thread-local variables should not be directly exposed at all
* variables intended to be dynamically scoped should be explicitly declared using the `dynamic` keyword at top level, using a thread-local variable (possibly of a different type) under the hood (contrary to languages that use dynamic scoping for all variables, this is opt-in, explicit, and more efficient. Note also that many languages do not have real thread-locals, they just pass the thread ID to a map, thus have atrocious performance.)
* reading a dynamically-scoped variable just accesses the thread-local, after unwrapping the container if needed. If the top level declaration didn't have an initializer, maybe this should throw an error?
* mutations to the dynamically-scoped variable require an explicit keyword, and involve saving the old value as if in a block-scoped local variable, and restoring it on unwind. If there's more than one mutation in a single block this can be optimized out.
* note that "assign a value wholesale" is only the most basic mutation; sometimes a stack is wanted (in which case the saved local value is just the depth) or a set (in which case the saved local value is the key to remove; this needs to be conditional)
* there should be escape hatches to manually do a save, restore, or write-without-saving. These should be used only to implement nontrivial mutations in the library or to port code using unsafe thread-locals.
(an alternate approach would be to write code as if it were passing arguments to functions, but internally change them to writes to thread-local variables (only if not passed from a parameter of the same name) purely for register-pressure-optimization reasons)
Also, if you pass a param, then it can be shared.
There's a perfectly cromulent register just begging to be used; the circuitry has already been paid for, generating heat whether you like it or not, what magic are you afraid of here?
> Also, if you pass a param, then it can be shared.
Maybe, but if you design for sharing you'll never use your program might be bigger and slower as a result. Sometimes that matters.
> There's a perfectly cromulent register just begging to be used; [...] what magic are you afraid of here?
Most of the magic is not when using the thread-local variable, but when allocating it. When you declare a "static __thread char *p", how do you know that for instance this is located at the 123th word of the per-thread area? What if that declaration is on a dynamic library, which was loaded late (dlopen) into the process? What about threads which were started before that dynamic library was loaded, and therefore did not have enough space in their per-thread area for that thread-local variable, when they call into code which references it? What happens if the thread-local variable has an initializer?
The documentation at https://gcc.gnu.org/onlinedocs/gcc/Thread-Local.html links to a 81-page document describing four TLS access models, and that's just for Unix-style ELF; Windows platforms have their own complexities (which IIRC includes a per-process maximum of 64 or 1088 TLS slots, with slots above the first 64 being handled in a slightly different way).
Believe it or not, it's exactly the same way it knows that "static char p" is located at the 123rd word of the data section: The linker does it!
> What if that declaration is on a dynamic library, which was loaded late (dlopen) into the process?
Same thing, except the linker is dynamic.
> What about threads which were started before that dynamic library was loaded, and therefore did not have enough space in their per-thread area for that thread-local variable, when they call into code which references it? What happens if the thread-local variable has an initializer?
Seriously? I mean, you probably do have enough space because address space is huge, but dlopen should move TLS.
> links to a 81-page document describing four TLS access models, and that's just for Unix-style ELF
Just because someone can write 81-pages about something doesn't mean that it takes 81-pages to understand something.
For my first magic trick, I'd like to make something not equal to itself:
foo("bar") == foo("bar")
> false // Magic!
foo[world1]("bar") == foo[world2]("bar")
> false // Perfectly reasonable behaviour
assertEquals ( "12:23", foo[time1]("bar") )
> Fail!
The way I fix it is to just make the implicit explicit instead, by turning them into parameters.
assertEquals ( "12:23", foo[](time1, "bar") )
> Pass!
I feel like thread-local takes this magic to the next level:
assertEquals ( three(), one() + two() )
assertEquals ( three[thread m](), one[thread n]() + two[thread o]() )
Magic just means "I don't understand this"
And whilst I don't think it's that complicated, I also don't think you really need to understand how TLS works to use it any more than you need to understand auto mechanics to drive a car. It's faster than walking.
> I feel like thread-local takes this magic to the next level:
> assertEquals ( three(), one() + two() )
> is really:
> assertEquals ( three[thread m](), one[thread n]() + two[thread o]() )
So I think it's more like:
assertEquals ( thread[o][three](), thread[o][one]() + thread[o][two]() )
assertEquals ( data[three](), data[one]() + data[two]() )
- TLS is indexed off of a base address, just like data.
- The thread "o" is going to be the same in each expression, so if thread[o] can live in a special-purpose register like %fs then it won't take a general-purpose register. If data can live in a special-purpose register...
Perhaps with a better picture you too will consider TLS is basically the same as non-TLS, except it's sometimes faster because, well, special circuitry.
If anything accessing via "data" is the weird one because even though it looks like you could just set data=0, you'd find almost everyone does it indirectly via %rip, and a variable "foo" is going to be %rip(-5000) in one place and %rip(-4000) in another, but would always be %fs:-26(0) if it were TLS.
Another pitfall with these is with thread-stealing concurrent schedulers - e.g. your worker thread now waits on something, and the scheduler decides to reuse the current thread for another worker - what is the meaning of thread_local there?
Another one would be coroutines (though haven't used them a lot in C/C++).
Like, you could easily write your compiler to do not have to rely on such machinery
Meanwhile they add complexity and decrease quality of error messages (in cpp)
Besides, there is more than one programming language, so that's something we have to deal with somehow.
And to be fair, merging modules in the compiler, as you go by, while not that difficult, is just annoying. If you link them properly together, into big amalgamated text/rodata/data sections, then you need to apply relocations (and have them in the first place). If you just place them next to each other, then you have to organize the inter-module calls via some moral equivalent of GOT/PLT. In any case, all this logic really doesn't have much to do with code generation proper, it's administrativia — and logic for dealing with has already been written for you and packed in the so called "link editor".
Your favorite toolchain will often include archives and objects that you might take for granted like crt0.o, init.o, fini.o, libgcc/clang_rt.builtins and more.
The compiler's design is simplified by not having to resolve references among symbols. The assembler can do this for references within a section and linkers can do it among sections. Linkers might have to add trampolines/thunks for relocations that span a distance longer than the opcode could reach. Loaders do this symbol resolution and relocation at runtime.
Then you can easily generate e.g web assembly from your own lang input without linker and it will run fine
Aint the same viable for x86?
Note that the nontrivial C program I showed above was not "hello world". Emitting output is yet more complicated still.
> you can easily generate e.g web assembly from your own lang input without linker
IIRC some things that generate wasm generate object files (e.g. C translation units). Those wasm objects require a linker to resolve symbol references to create the executable.
> it will run fine
...on a wasm virtual machine. IIRC Java doesn't have anything like a traditional linker either.
> Aint the same viable for x86?
I mean, you can do anything you want inventing a new compiler that does much more than traditional compilers. But consider perhaps that there's a lot of value in the status quo that you might appreciate more if you were more familiar with the design of a toolchain.
It's not that linkers are an absolute drop-dead necessity, it's that having one allows you to have distinct assembler and compiler functions that are much simpler to design by focusing on one kind of transformation.
I guess rustc detects the situations in which the linker would throw an error and then throws its own error preemptively. It leads to a much better user experience than the C++ one, since the error messages produced by the linker are always unnecessarily terrible
Pretty much. The crucial difference between C and Rust which enables Rust to do this sort of detection is that in Rust, the extern things are anchored in modules (crates? whatever), and so when you import things, you have to say where you are importing them from.
extern void *magic_init(int);
extern void *magic_stuff(void*, const char*, int);
extern void magic_fini(void*);
use crate_of_magic::{init, stuff, fini};
You need a linker as soon as you are dealing with either multiple languages in one project (say, C++ and ASM) or if you include other libraries.