Believe this would be discovered much sooner had smap been enabled. (SMAP prevents the kernel from accessing user land, and would thus cause bcopy to crash in the normal not exploit case, which somebody might fix. Or just run sparc64 or another arch without user land mappings.) But this depends on somebody going down the bad path.
Believe this would be discovered much sooner had smap been enabled. (SMAP prevents the kernel from accessing user land, and would thus cause bcopy to crash in the normal not exploit case, which somebody might fix. Or just run sparc64 or another arch without user land mappings.) But this depends on somebody going down the bad path.
SMAP is enabled; as you point out someone would first have had to take the bad path, encounter the crash, and submit a report.
It does have a name, it’s CVE-2019-5602.