Fix NULL ptr deref in ipc_describe_workspace

[TheBlueMatt]

When hot-plugging monitor(s) ipc_json_describe_workspace can end up dereferencing a NULL ptr in wlr_output->name. Here we simply check that it is set.

While this is probably a race and just checking that it is not NULL before accessing it won't fully solve the race, it should make it substantially more rare.

Mostly addresses #8747

[emersion]

I would prefer to find the root cause of this NULL pointer instead of adding a workaround.

[TheBlueMatt]

Sure, don't disagree, but absent someone having time to dig into it probably worth at least making sway stable.

[TheBlueMatt]

Any update here?

[emersion]

Would still prefer to find the root cause rather than merge this.

[TheBlueMatt]

If no one has time to dig into the root cause it still might presumably be preferrable to merge this so that at least sway is stable on impacted devices. A broken sway seems worse than a workaround for a bug.

[emersion]

If a workaround is merged, it removes any incentive to find the root cause.

[TheBlueMatt]

If impacted users migrate off of sway, it also removes any incentive to fix the issue :p

[kennylevinsen]

If impacted users migrate off of sway, it also removes any incentive to fix the issue :p

The fact remains that adding a NULL check for a field that can never be NULL isn't a fix. The value should be set by the backend at the very start, and when destroyed it is simply free'd, not NULL'd.

It could possibly be a use-after-free/dangling pointer of sway_output or wlr_output, rather than being related to the name field itself. Reproducing the issue with address sanitizer enabled might help pinpoint things. That is done by specifying -Db_sanitize=address to meson, and will cause a report to be logged to stderr when a memory snafu is detected, and might give a clue as to what's happening.

(I don't have a reproduction of this issue myself.)

[emersion]

@njdom24, that sounds like a different issue. My guess would be that con->pending.workspace is NULL.