kernel.org, VMware Inc [off-list ref], linux-snps-arc@lists.infradead.org, mgorman@suse.de, jacob.jun.pan@linux.intel.com, Arnd Bergmann [off-list ref], ulli.kroll@googlemail.com, vgupta@kernel.org, linux-clk@vger.kernel.org, josh@joshtriplett.org, rostedt@goodmis.org, rcu@vger.kernel.org, Borislav Petkov [off-list ref], bcain@quicinc.com, tsbogend@alpha.franken.de, linux-parisc@vger.kernel.org, sudeep.holla@arm.com, shawnguo@kernel.org, davem@davemloft.net, kirill.shutemov@linux.intel.com, dalias@libc.org, tony@atomide.com, amakhalov@vmware.com, bjorn.andersson@linaro.org, "H. Peter Anvin" [off-list ref], sparclinux@vger.kernel.org, linux-hexagon@vger.kernel.org, linux-riscv@lists.infradead.org, Isaku Yamahata [off-list ref], anton.ivanov@cambridgegreys.com, jonas@southpole.se, yury.norov@gmail.com, Richard Weinberger [off-list ref], X86 ML [off-list ref], linux@armlinux.org.uk, Ingo Molnar [off-list ref], Albert Ou [off-list ref], paulmck@ker
 nel.org, hca@linux.ibm.com, stefan.kristiansson@saunalahti.fi, openrisc@lists.librecores.org, Paul Walmsley [off-list ref], linux-tegra@vger.kernel.org, Namhyung Kim [off-list ref], Andy Shevchenko [off-list ref], jpoimboe@kernel.org, Juergen Gross [off-list ref], monstr@monstr.eu, linux-mips@vger.kernel.org, Palmer Dabbelt [off-list ref], Anup Patel [off-list ref], ink@jurassic.park.msu.ru, Johannes Berg [off-list ref], linuxppc-dev@lists.ozlabs.org
Errors-To: linuxppc-dev-bounces+linuxppc-dev=archiver.kernel.org@lists.ozlabs.org
Sender: "Linuxppc-dev" [off-list ref]

On Wed, Jun 8, 2022 at 10:48 PM Peter Zijlstra [off-list ref] wrote:

Now that arch_cpu_idle() is expected to return with IRQs disabled,
avoid the useless STI/CLI dance.

Per the specs this is supposed to work, but nobody has yet relied up
this behaviour so broken implementations are possible.

I'm totally newbie here.

The point of safe_halt() is that STI must be used and be used
directly before HLT to enable IRQ during the halting and stop
the halting if there is any IRQ.

In TDX case, STI must be used directly before the hypercall.
Otherwise, no IRQ can come and the vcpu would be stalled forever.

Although the hypercall has an "irq_disabled" argument.
But the hypervisor doesn't (and can't) touch the IRQ flags no matter
what the "irq_disabled" argument is.  The IRQ is not enabled during
the halting if the IRQ is disabled before the hypercall even if
irq_disabled=false.

The "irq_disabled" argument is used for workaround purposes:
https://lore.kernel.org/kvm/c020ee0b90c424a7010e979c9b32a28e9c488a51.1651774251.git.isaku.yamahata@intel.com/ (local)

Hope my immature/incorrect reply elicits a real response from
others.

Thanks
Lai