view the rest of the comments
Technology
This is the official technology community of Lemmy.ml for all news related to creation and use of technology, and to facilitate civil, meaningful discussion around it.
Ask in DM before posting product reviews or ads. All such posts otherwise are subject to removal.
Rules:
1: All Lemmy rules apply
2: Do not post low effort posts
3: NEVER post naziped*gore stuff
4: Always post article URLs or their archived version URLs as sources, NOT screenshots. Help the blind users.
5: personal rants of Big Tech CEOs like Elon Musk are unwelcome (does not include posts about their companies affecting wide range of people)
6: no advertisement posts unless verified as legitimate and non-exploitative/non-consumerist
7: crypto related posts, unless essential, are disallowed
You're not that far off, everything you've said about shared memory is true of writable shared memory, which is what is usually meant when talking about "shared memory" in user space.
When a program loads a .so a version of that file is loaded in the program's virtual memory space so technically that same .so file exists in multiple virtual spaces, but that's only a trick used by your MMU (memory management unit), the .so file is only loaded once in actual memory:
Let's imagine some program loads opencl.so and no one else on the system uses it. The kernel will check all needed libs and loads that one at physical address 0x700. After loading it up it talks to the MMU and creates a virtual address space for said program, telling the hardware to create a read-only link between virtual address 0x100 and physical address 0x700, "loading" the library in virtual memory (without actually making a copy). If some other program then needs opencl.so the kernel will recognize that lib as already loaded so skips ahead to linking virtual adress 0x200 (or whatever really, virtual address space is specific to one instance of one program, it could be 0x100 again it doesn't matter) to physical address 0x700, once again loading it in program memory without making a copy.
Ahh, okay. I think the main confusion is I was kind of ignoring the code portion of the library, which is easily deduplicated like you said.
Any allocations done by the library running (opencl.so in this example) will still be separate per process though, unless the program is passing mutable shared memory buffers around, which could be the case if there's IPC going on.
The point about memory usage being hard to measure is definitely true, especially since linux won't actually consume memory until a write is made to the block, unlike Windows which preemptively allocates the full requested size.