btop
btop copied to clipboard
aristocratos
[BUG] segmentation fault after running btop over night
Describe the bug
I left a btop terminal window open overnight, trying to reproduce #316, but got a segmentation fault instead.
To Reproduce Start btop. Leave it on for a long time. Come back to a segementation fault.
Expected behavior
It shouldn't segfault
Screenshots
Info (please complete the following information):
- btop++ version: 1.2.13
- Binary: native from fresh Homebrew install
- (If compiled) Compiler and version:
- Architecture: arm64
- Platform: OsX
- Terminal used: Warp
- Font used: default
Additional context
contents of ~/.config/btop/btop.log
(Logger set to DEBUG in options):
2023/02/03 (23:34:51) | ===> btop++ v.1.2.13 2023/02/03 (23:34:51) | WARNING: No UTF-8 locale detected! Some symbols might not display correctly. 2023/02/03 (23:35:02) | INFO: Logger set to INFO 2023/02/03 (23:35:03) | INFO: Logger set to DEBUG 2023/02/03 (23:35:03) | INFO: Logger set to DEBUG 2023/02/04 (00:28:23) | ERROR: failed getting network interfaces 2023/02/04 (01:02:28) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (01:09:38) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (01:44:13) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (01:44:24) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (01:46:07) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (01:46:12) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (02:38:36) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (02:38:41) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (02:38:46) | ERROR: Stall in Runner thread, restarting! 2023/02/04 (02:38:51) | ERROR: Stall in Runner thread, restarting!
(The segfault, as per the screenshot, happened at 02:38:46)
Edit: running it for a third time, I got another error:
btop(50807,0x16de9f000) malloc: double free for ptr 0x117008a00 zsh: abort btop --debug
So it seems that btop stops working after some time due to some error and displays either:
- pointer being freed was not allocated
- segmentation fault
- double free for ptr
There's a memory leak somewhere which I'm trying to detect; it increases by 3MiB on each second, at least on my system [GNU / Linux Debian].
@aristocratos can you explain this excerpt to me, please?
static InputThr& instance() {
// intentional memory leak, to simplify shutdown process
static InputThr* input = new InputThr();
return *input;
}
It's from file btop_input.cpp if it helps.
valgrind detected it at line 128 and also
InputThr() : thr(run, this) {
}
from line 88.
Why would we want to intentionally leak memory?
Isn't there any other safer method to accomplish a shutdown process, in a more graceful way?
Please advice.
By the way, the leak takes place immediately with the following command, on my side: make DEBUG=true VERBOSE=true CXXFLAGS=-fsanitize=address,undefined,leak LDFLAGS=-fsanitize=address,undefined,leak ADDFLAGS="-fuse-ld=mold -ggdb3".
After it finishes the compilation, I filter for btop and within info, I can clearly see the memory leak taking place with every second that passes by; it increases by 3 MiB steadily.
@stefanos82
can you explain this excerpt to me, please?
See #227
By the way, the leak takes place immediately with the following command, on my side: make DEBUG=true VERBOSE=true CXXFLAGS=-fsanitize=address,undefined,leak LDFLAGS=-fsanitize=address,undefined,leak ADDFLAGS="-fuse-ld=mold -ggdb3".
Unless you are seeing this with sanitizers off not running in a debugger I would assume the sanitizing and debugger are creating the leak from trying to fix what looks like a leak but really isn't.
But I'm pretty sure this is unrelated to @SeseMueller's reported bug on MacOS arm64.
@SeseMueller What are your hardware specs?
See #227
Got it :+1:
Unless you are seeing this with sanitizers off not running in a debugger I would assume the sanitizing and debugger are creating the leak from trying to fix what looks like a leak but really isn't.
Yeah, it probably is this scenario because indeed I cannot see this behavior otherwise.
@SeseMueller What are your hardware specs?
I've got a Macbook Air M1 with 16 Gb. It might not be related to my specs, but I have no other hardware to test it currently.
I see the same bug / behavior on Ubuntu Server using btop 1.2.3
matt@lian:~/.config/btop$ neofetch
.-/+oossssoo+/-. matt@lian
`:+ssssssssssssssssss+:` ---------
-+ssssssssssssssssssyyssss+- OS: Ubuntu 22.04.2 LTS x86_64
.ossssssssssssssssssdMMMNysssso. Kernel: 5.15.0-60-generic
/ssssssssssshdmmNNmmyNMMMMhssssss/ Uptime: 2 days, 14 hours, 32 mins
+ssssssssshmydMMMMMMMNddddyssssssss+ Packages: 767 (dpkg), 5 (snap)
/sssssssshNMMMyhhyyyyhmNMMMNhssssssss/ Shell: bash 5.1.16
.ssssssssdMMMNhsssssssssshNMMMdssssssss. Resolution: 3440x1440
+sssshhhyNMMNyssssssssssssyNMMMysssssss+ Terminal: /dev/pts/0
ossyNMMMNyMMhsssssssssssssshmmmhssssssso CPU: Intel Xeon L5640 (12) @ 2.268GHz
ossyNMMMNyMMhsssssssssssssshmmmhssssssso GPU: AMD ATI Radeon RX 470/480/570/570X/580/580X/590
+sssshhhyNMMNyssssssssssssyNMMMysssssss+ Memory: 3537MiB / 11958MiB
.ssssssssdMMMNhsssssssssshNMMMdssssssss.
/sssssssshNMMMyhhyyyyhdNMMMNhssssssss/
+sssssssssdmydMMMMMMMMddddyssssssss+
/ssssssssssshdmNNNNmyNMMMMhssssss/
.ossssssssssssssssssdMMMNysssso.
-+sssssssssssssssssyyyssss+-
`:+ssssssssssssssssss+:`
.-/+oossssoo+/-.
023/03/01 (20:32:22) | ===> btop++ v.1.2.3 2023/03/01 (20:32:22) | ERROR: Stall in Runner thread, restarting!
I have same bug after leaving it on for a long time on macos 12.6 (hw spec: mac m1 pro arm64/16gb ram/1tb ssd nvme). Btop version: 1.2.13 installed with brew package manager.
Same issue here, usually segfaults on startup (though goes accompanied with eg. "terminate called without an active exception", will update once it happens again)
btop 1.2.13 neofetch:
[root@Rory@Rory-desktop ~]$ neofetch
-` root@Rory@Rory-desktop
.o+` ----------------------
`ooo/ OS: Arch Linux x86_64
`+oooo: Host: <redacted for length>
`+oooooo: Kernel: 6.2.11-zen1-1-zen
-+oooooo+: Uptime: 18 hours, 24 mins
`/:-:++oooo+: Packages: 1844 (pacman), 70 (nix-user)
`/++++/+++++++: Shell: bash 5.1.16
`/++++++++++++++: Resolution: 3840x2160, 1920x1080
`/+++ooooooooooooo/` WM: bspwm
./ooosssso++osssssso+` Theme: Adwaita [GTK2/3]
.oossssso-````/ossssss+` Icons: Adwaita [GTK2/3]
-osssssso. :ssssssso. Terminal: kitty
:osssssss/ osssso+++. Terminal Font: monospace 11.0
/ossssssss/ +ssssooo/- CPU: AMD Ryzen 5 3600 (12) @ 3.600GHz
`/ossssso+/:- -:/+osssso+- GPU: NVIDIA GeForce RTX 2060 Rev. A
`+sso+:-` `.-/+oso: GPU: NVIDIA GeForce GTX 760 OEM
`++:. `-/+/ Memory: 19998MiB / 40053MiB
.` `/
Same here btop 1.2.13 installed through brew macOS Montery 12.6.1 M1 Pro 32Gb Ram iTerm2 v3.4.19 zsh 5.8.1 (x86_64-apple-darwin21.0)
Runs for about a minute and then crashes
Same issue, it occurs pretty much randomly for me, sometimes after just a few minutes but on other occasions it runs for hours. btop 1.12.13 installed through pacman
-` rex@archrex
.o+` -----------
`ooo/ OS: Arch Linux x86_64
`+oooo: Kernel: 6.3.4-arch2-1
`+oooooo: Uptime: 32 mins
-+oooooo+: Packages: 1050 (pacman)
`/:-:++oooo+: Shell: bash 5.1.16
`/++++/+++++++: Resolution: 1920x1200, 2560x1440, 1920x1200
`/++++++++++++++: DE: Plasma 5.27.5
`/+++ooooooooooooo/` WM: KWin
./ooosssso++osssssso+` Theme: [Plasma], Breeze [GTK2/3]
.oossssso-````/ossssss+` Icons: [Plasma], breeze-dark [GTK2/3]
-osssssso. :ssssssso. Terminal: kitty
:osssssss/ osssso+++. CPU: AMD Ryzen 5 5600X (12) @ 3.700GHz
/ossssssss/ +ssssooo/- GPU: NVIDIA GeForce RTX 3070 Ti
`/ossssso+/:- -:/+osssso+- Memory: 6951MiB / 23939MiB
`+sso+:-` `.-/+oso:
`++:. `-/+/
.` `/
Any update? I suspect this might be just the same issue with #765.
For me the startup of btop just becomes unreasonably slow, and crashes in a minute. I thought it was macOS breaking something but a clean reboot did not fix it (see uptime below).
That being said, such things did not happen to me until I upgraded to macOS Sonoma 14.4. Could be their aggressive memory-violation strategy uncovering the issue. I'd love to help testing.
..' 45gfg9@45gfg9-MacBook-Pro
,xNMM. -------------------------
.OMMMMo OS: macOS Sonoma 14.4.1 23E224 arm64
lMM" Host: MacBook Pro (16-inch, 2021)
.;loddo:. .olloddol;. Kernel: 23.4.0
cKMMMMMMMMMMNWMMMMMMMMMM0: Uptime: 10 mins
.KMMMMMMMMMMMMMMMMMMMMMMMWd. Packages: 345 (brew), 44 (brew-cask)
XMMMMMMMMMMMMMMMMMMMMMMMX. Shell: zsh 5.9
;MMMMMMMMMMMMMMMMMMMMMMMM: Display (Color LCD): 3456x2234 @ 120Hz (as 1728x1117) [Built-in]
:MMMMMMMMMMMMMMMMMMMMMMMM: DE: Aqua
.MMMMMMMMMMMMMMMMMMMMMMMMX. WM: Quartz Compositor
kMMMMMMMMMMMMMMMMMMMMMMMMWd. WM Theme: Multicolor (Dark)
'XMMMMMMMMMMMMMMMMMMMMMMMMMMk Font: .AppleSystemUIFont [System], Helvetica [User]
'XMMMMMMMMMMMMMMMMMMMMMMMMK. Cursor: Fill - Black, Outline - White (32px)
kMMMMMMMMMMMMMMMMMMMMMMd Terminal: iTerm 3.4.23
;KMMMMMMMWXXWMMMMMMMk. Terminal Font: SFMono-Regular (13pt)
"cooc*" "*coo'" CPU: Apple M1 Max (10) @ 3.23 GHz
GPU: Apple M1 Max (32) [Integrated]
Memory: 14.80 GiB / 64.00 GiB (23%)
Swap: Disabled
Disk (/): 1.03 TiB / 1.81 TiB (57%) - apfs [Read-only]
