Dmitry Vyukov

Please show the other changes where this is needed. Otherwise it's impossible to understand if this is the right approach or not. There should be some synchronization in the symbolizer,...

> Any further comment for this PR since there is no pending item? @dvyukov Sorry, did not realize this is ready for review. Please either comment something like "Please take...

> But I don't know if there's any way to explicitly wait until the unmount process is complete. remove_dir() should return before the next test is started. If the fs...

You are right. I forgot about MNT_DETACH. This is very old code that goes way back to almost first days of syzkaller: https://github.com/google/syzkaller/commit/1e06d2bafc There are mentioned of a number of...

My bet is that MNT_FORCE will hang due to reference leaks. We will also probably not detect it well, e.g. executor will be killed on timeout, but the device is...

Re LOOP_CTL_GET_FREE: IIRC it has own accounting of free/busy devices, which has nothing to do with actual use of devices. And also any code is free to just use /dev/loopN...

Right, it's an internal link. Sorry. I've tried to find any similar public statements, but it's very hard to google for something like "subsystems don't mark patches for stable" b/c...

Frankly I don't know the current status and if/how things have improved. Perhaps @gregkh and @sashalevin can share more on the current status of the stable process: - are there...

Removing more code duplication across vm implementations is good on its own. Re actual fix, I guess anything that fixes the problem with truncated output will do. We can sleep,...

`0x40086602` is `FS_IOC_SETFLAGS`: ``` sys/linux/fs_ioctl.txt.const:FS_IOC_SETFLAGS = 1074292226, 386:arm:1074030082, mips64le:ppc64le:2148034050 ``` Related to #477 and https://github.com/dvyukov/syzkaller/commit/cd4daf14452ee65caaf268f92e3242e338647533