On Thu, 15 Apr 2021 01:55:22 PDT (-0700), Anup Patel wrote:

We should return a negative error code upon failure in

Not just should: the PLIC driver assumed this was the return flavor and will overflow a per_cpu because it doesn't detect the error. IIUC this will manifest if we have more harts that NR_CPUS, which is a possible configuration.

That said, it's gone unnoticed for ~3 years so I'm inclined to leave it out of fixes because I have nothing else and this late in the cycle is always a bit touchy. It's kind of splitting hairs as it'll just get backported, but post-rc8 is an odd spot to be in.

Is this actually manifesting as a crash for you?

Sorry I missed this last week.

riscv_hartid_to_cpuid() instead of NR_CPUS. This is also aligned with all uses of riscv_hartid_to_cpuid() which expect negative error code upon failure.

Fixes: 6825c7a80f18 ("RISC-V: Add logical CPU indexing for RISC-V")

Also

Fixes: f99fb607fb2b ("RISC-V: Use Linux logical CPU number instead of hartid")

which is what actually manifests the bug -- again, kind of splitting hairs as it's the patch after.

Cc: stable@vger.kernel.org Signed-off-by: Anup Patel anup.patel@wdc.com

---

arch/riscv/kernel/smp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/riscv/kernel/smp.c b/arch/riscv/kernel/smp.c index ea028d9e0d24..d44567490d91 100644 --- a/arch/riscv/kernel/smp.c +++ b/arch/riscv/kernel/smp.c @@ -54,7 +54,7 @@ int riscv_hartid_to_cpuid(int hartid) return i;

pr_err("Couldn't find cpu id for hartid [%d]\n", hartid);

• return i;

• return -ENOENT;

}

void riscv_cpuid_to_hartid_mask(const struct cpumask *in, struct cpumask *out)