/g/ - syscall of the day: readlink(at) - Technology

Anonymous

syscall of the day: readlink(a(...) 03/14/26(Sat)13:04:34 No.108370984

File: 1771973239227176.png (316 KB, 1080x996)

syscall of the day: readlink(at) Anonymous 03/14/26(Sat)13:04:34 No.108370984 Archived

previous: >>108361692
#define __NR_readlink            89
#define __NR_readlinkat                267
https://man7.org/linux/man-pages/man2/readlink.2.html

tl;dr:
read the contents of a symbolic link

what a disgusting syscall LMFAO
>readlink() does not append a terminating null byte to buf.
>The required size for the buffer can be obtained from the stat.st_size value returned by a call to lstat(2) on the link. However, the number of bytes written by readlink() and readlinkat() should be checked to make sure that the size of the symbolic link did not increase between the calls.
genuinely horrifying...
i don't have much else to say about this one. it's just bad

relevant resources:
man man
man syscalls
https://man7.org/linux/man-pages/
https://linux.die.net/man/
https://elixir.bootlin.com/linux/
https://elixir.bootlin.com/musl/
https://elixir.bootlin.com/glibc/

Anonymous
03/14/26(Sat)13:27:03 No.108371125

Anonymous 03/14/26(Sat)13:27:03 No.108371125

>>108370984
LMAO. Who designed that shit?

Anonymous
03/14/26(Sat)13:47:59 No.108371245

Anonymous 03/14/26(Sat)13:47:59 No.108371245

>>108370984
>The required size for the buffer can be obtained from the stat.st_size value returned by a call to lstat(2) on the link.
At least they're limited to a few hundred bytes, so you can just use a buffer larger than the limit.

Anonymous
03/14/26(Sat)13:50:21 No.108371269

Anonymous 03/14/26(Sat)13:50:21 No.108371269

>>108371245
they are not technically limited to that

Anonymous
03/14/26(Sat)14:22:42 No.108371460

Anonymous 03/14/26(Sat)14:22:42 No.108371460

>>108370984
Whats a syscall

Anonymous
03/14/26(Sat)14:24:52 No.108371473

Anonymous 03/14/26(Sat)14:24:52 No.108371473

What is so horrifying about the lack of null terminator? I don't get it.

Anonymous
03/14/26(Sat)14:26:15 No.108371481

Anonymous 03/14/26(Sat)14:26:15 No.108371481

>>108371460
a method to access kernel resources/apis from userspace

Anonymous
03/14/26(Sat)14:28:19 No.108371489

Anonymous 03/14/26(Sat)14:28:19 No.108371489

>>108371481
use case?

Anonymous
03/14/26(Sat)14:35:30 No.108371532

Anonymous 03/14/26(Sat)14:35:30 No.108371532

>>108371125
Autists who don't understand kernelland locks.

Anonymous
03/14/26(Sat)14:40:51 No.108371561

Anonymous 03/14/26(Sat)14:40:51 No.108371561

>>108371473
Things like null terminators and zero indexing frighten retards.

Anonymous
03/14/26(Sat)14:41:46 No.108371567

Anonymous 03/14/26(Sat)14:41:46 No.108371567

>>108371473
Tons of functions expect NUL-terminated strings. If a function doesn't write them, then you have to write them yourself. If you don't know how long the string is, and there is no reliable way to estimate the size (because locks aren't a thing in Linux-autismo land), then you have to zero out the buffer IN ITS ENTIRETY first and THEN issue the call.

Anonymous
03/14/26(Sat)14:41:46 No.108371568

Anonymous 03/14/26(Sat)14:41:46 No.108371568

>>108371481
oh cool thank you!

Anonymous
03/14/26(Sat)14:47:01 No.108371607

Anonymous 03/14/26(Sat)14:47:01 No.108371607

>>108371567
>then you have to zero out the buffer IN ITS ENTIRETY first
There is absolutely no reason to do this, unless you're trying to waste electricity and your user's time.

Anonymous
03/14/26(Sat)14:48:14 No.108371615

Anonymous 03/14/26(Sat)14:48:14 No.108371615

>>108371607
>what's a race condition
Just be quiet.

Anonymous
03/14/26(Sat)14:53:43 No.108371650

Anonymous 03/14/26(Sat)14:53:43 No.108371650

memset(str, '\0', sizeof(str));
omg make it stop make it stooopp

Anonymous
03/14/26(Sat)14:55:35 No.108371662

Anonymous 03/14/26(Sat)14:55:35 No.108371662

>>108371615
>int size = readlink(path, buf, bufsiz);
>buf[size] = 0;
Only in a schizo's mind does this cause a race condition.

Anonymous
03/14/26(Sat)14:55:57 No.108371663

Anonymous 03/14/26(Sat)14:55:57 No.108371663

>>108371650
Congrats, you've just thrashed a quarter of your L1 cache.

Anonymous
03/14/26(Sat)14:57:58 No.108371679

Anonymous 03/14/26(Sat)14:57:58 No.108371679

>>108371662
Talent is hitting a mark others miss.
Genius is hitting a mark others don't even see.

You do neither.

Anonymous
03/14/26(Sat)14:59:33 No.108371685

Anonymous 03/14/26(Sat)14:59:33 No.108371685

>>108371662
Has to be bufsz-1 or you will overflow if the link is exactly bufsz bytes. This kind of demonstrates why this interface should have probably just NUL terminated the destination buffer instead of making each caller do this dance.

Anonymous
03/14/26(Sat)15:00:32 No.108371697

Anonymous 03/14/26(Sat)15:00:32 No.108371697

what is lil bro yapping about?

Anonymous
03/14/26(Sat)15:00:50 No.108371700

Anonymous 03/14/26(Sat)15:00:50 No.108371700

>>108371663
more like 1% of it for 10 nano seconds

Anonymous
03/14/26(Sat)15:02:46 No.108371711

Anonymous 03/14/26(Sat)15:02:46 No.108371711

>>108371685
Point taken, but it's still not a race condition.

Anonymous
03/14/26(Sat)15:05:08 No.108371731

Anonymous 03/14/26(Sat)15:05:08 No.108371731

>>108371697
Don't worry, you couldn't understand it anyway. Just go play with some toy language.

>>108371700
Oops, not enough memory. Go reallocate.

Oops, not enough memory again. Go reallocate.

Oops, not enough memory again. Go reallocate.

Oops, not enough memory again. Go reallocate.

Oops, not enough memory again. Go reallocate.

Anonymous
03/14/26(Sat)15:07:49 No.108371748

Anonymous 03/14/26(Sat)15:07:49 No.108371748

>>108371731
Is the 8 bit cache in the room with us right now?

Anonymous
03/14/26(Sat)15:10:50 No.108371762

Anonymous 03/14/26(Sat)15:10:50 No.108371762

>>108371748
>8-bit cache
Found your problem.

Anonymous
03/14/26(Sat)15:13:49 No.108371780

Anonymous 03/14/26(Sat)15:13:49 No.108371780

>>108370984
What speaks against rewriting it?

Anonymous
03/14/26(Sat)15:14:56 No.108371784

Anonymous 03/14/26(Sat)15:14:56 No.108371784

simd schizo redditor out of his depth again meltie and emacs screenshot incoming

Anonymous
03/14/26(Sat)15:16:35 No.108371791

Anonymous 03/14/26(Sat)15:16:35 No.108371791

>>108371731
>Oops, not enough memory again. Go reallocate.
Not an issue if you start with a big enough buffer.

Anonymous
03/14/26(Sat)15:17:12 No.108371793

Anonymous 03/14/26(Sat)15:17:12 No.108371793

>>108371784
Speak English

Anonymous
03/14/26(Sat)15:20:16 No.108371816

Anonymous 03/14/26(Sat)15:20:16 No.108371816

>>108371784
Hey, it's not my fault you're only alive because you technically qualify for human rights, thus making killing you illegal.

Anonymous
03/14/26(Sat)18:39:45 No.108373064

Anonymous 03/14/26(Sat)18:39:45 No.108373064

bampu

Anonymous
03/14/26(Sat)22:33:18 No.108374267

Anonymous 03/14/26(Sat)22:33:18 No.108374267

seems like there is more interest in fighting than in syscalls...

Anonymous
03/14/26(Sat)22:42:35 No.108374308

Anonymous 03/14/26(Sat)22:42:35 No.108374308

>>108374267
This is what happens when you lay out bait for the registry schizo (if that's who this is).
At least the permissions ones tomorrow should be interesting. I predict anger at there not being enough granularity in permissions, and my issue with them that I'll mention then.

Anonymous
03/14/26(Sat)23:37:29 No.108374547

Anonymous 03/14/26(Sat)23:37:29 No.108374547

>>108374308
>obvious persecution complex suggests schizophrenia
>projects it onto others
It makes so much sense now.

Anonymous
03/14/26(Sat)23:51:04 No.108374602

Anonymous 03/14/26(Sat)23:51:04 No.108374602

>>108374547
>people getting sick of your bullshit
>they must have schizophrenia
Of course, there's no other explanation for this

Anonymous
03/15/26(Sun)00:02:17 No.108374636

Anonymous 03/15/26(Sun)00:02:17 No.108374636

>>108374602
Yes, and there is absolutely nothing you can do about it.

Nothing.

Anonymous
03/15/26(Sun)07:14:00 No.108376198

Anonymous 03/15/26(Sun)07:14:00 No.108376198

retarded OP