Date: Fri, 18 Jul 2025 09:38:41 +0000 From: bugzilla-noreply@freebsd.org To: virtualization@FreeBSD.org Subject: [Bug 288237] bhyve does not boot Rocky Linux 10 from ISO Message-ID: <bug-288237-27103-kZItlCya8p@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-288237-27103@https.bugs.freebsd.org/bugzilla/> References: <bug-288237-27103@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=288237 --- Comment #13 from Zhenlei Huang <zlei@FreeBSD.org> --- (In reply to Konstantin Belousov from comment #12) > No, if the guest issues undefined instruction, it must be reflected as the > corresponding exception (#UD) back to the guest. It is up to the guest code > to handle it or not. > Typically the loaders are very naive in handling CPU exceptions, most often > they do not even install exception handlers. This is even more true for UEFI > boot envs, where UEFI started having dedicated debugger protocol. > In other words, if loader issued an instruction that is not supported by the > CPU (and bhyve does not emulate unsupported instructions), the result is > typically hang or silent reboot. Can I conclude that, only when VM-Exit events occurs, then the VMM can take over the CPU ( so it is exiting a virtual execution mode ) ? I did a quick look at Intel’s Software Developer Manual outlines 64 “Basic Exit Reasons”, it appears an undefined instruction do not cause VM-Exit. So if the guest ( either the loader or kernel ) do not handle the event of undefined instruction ( print to console or hand it over to VMM explicitly, say via vmexit ), the VMM is hopeless to have a chance to log the `undefined instruction` event ? -- You are receiving this mail because: You are the assignee for the bug.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-288237-27103-kZItlCya8p>