[PATCH] RISC-V: Provide a more helpful error message on invalid ISA strings

[Palmer Dabbelt]

Right now we provide a somewhat unhelpful error message on systems with
invalid error messages, something along the lines of

[    0.000000] CPU with hartid=0 is not available
[    0.000000] ------------[ cut here ]------------
[    0.000000] kernel BUG at arch/riscv/kernel/smpboot.c:174!
[    0.000000] Kernel BUG [#1]
[    0.000000] Modules linked in:
[    0.000000] CPU: 0 PID: 0 Comm: swapper Not tainted 6.4.0-rc1-00096-ge0097d2c62d5-dirty #1
[    0.000000] Hardware name: Microchip PolarFire-SoC Icicle Kit (DT)
[    0.000000] epc : of_parse_and_init_cpus+0x16c/0x16e
[    0.000000]  ra : of_parse_and_init_cpus+0x9a/0x16e
[    0.000000] epc : ffffffff80c04e0a ra : ffffffff80c04d38 sp : ffffffff81603e20
[    0.000000]  gp : ffffffff8182d658 tp : ffffffff81613f80 t0 : 000000000000006e
[    0.000000]  t1 : 0000000000000064 t2 : 0000000000000000 s0 : ffffffff81603e80
[    0.000000]  s1 : 0000000000000000 a0 : 0000000000000000 a1 : 0000000000000000
[    0.000000]  a2 : 0000000000000000 a3 : 0000000000000000 a4 : 0000000000000000
[    0.000000]  a5 : 0000000000001fff a6 : 0000000000001fff a7 : ffffffff816148b0
[    0.000000]  s2 : 0000000000000001 s3 : ffffffff81492a4c s4 : ffffffff81a4b090
[    0.000000]  s5 : ffffffff81506030 s6 : 0000000000000040 s7 : 0000000000000000
[    0.000000]  s8 : 00000000bfb6f046 s9 : 0000000000000001 s10: 0000000000000000
[    0.000000]  s11: 00000000bf389700 t3 : 0000000000000000 t4 : 0000000000000000
[    0.000000]  t5 : ffffffff824dd188 t6 : ffffffff824dd187
[    0.000000] status: 0000000200000100 badaddr: 0000000000000000 cause: 0000000000000003
[    0.000000] [<ffffffff80c04e0a>] of_parse_and_init_cpus+0x16c/0x16e
[    0.000000] [<ffffffff80c04c96>] setup_smp+0x1e/0x26
[    0.000000] [<ffffffff80c03ffe>] setup_arch+0x6e/0xb2
[    0.000000] [<ffffffff80c00384>] start_kernel+0x72/0x400
[    0.000000] Code: 80e7 4a00 a603 0009 b795 1097 ffe5 80e7 92c0 9002 (9002) 715d
[    0.000000] ---[ end trace 0000000000000000 ]---
[    0.000000] Kernel panic - not syncing: Fatal exception in interrupt

This adds a warning for the cases where the ISA string isn't valid.  It's still
above the BUG_ON cut, but hopefully it's at least a bit easier for users.

Signed-off-by: Palmer Dabbelt <palmer at rivosinc.com>
---
I haven't tried this yet, but Conor posted the log as we were discussing
the DT deprecation at
<https://lore.kernel.org/all/20230629-angled-gallantly-8fe7451a25fa@spud/>.
---
 arch/riscv/kernel/cpu.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/arch/riscv/kernel/cpu.c b/arch/riscv/kernel/cpu.c
index a2fc952318e9..3af2d214ce21 100644
--- a/arch/riscv/kernel/cpu.c
+++ b/arch/riscv/kernel/cpu.c
@@ -66,11 +66,15 @@ int riscv_early_of_processor_hartid(struct device_node *node, unsigned long *har
 		return -ENODEV;
 	}
 
-	if (IS_ENABLED(CONFIG_32BIT) && strncasecmp(isa, "rv32ima", 7))
+	if (IS_ENABLED(CONFIG_32BIT) && strncasecmp(isa, "rv32ima", 7)) {
+		pr_warn("CPU with hartid=%lu does not support rv32ima", *hart);
 		return -ENODEV;
+	}
 
-	if (IS_ENABLED(CONFIG_64BIT) && strncasecmp(isa, "rv64ima", 7))
+	if (IS_ENABLED(CONFIG_64BIT) && strncasecmp(isa, "rv64ima", 7)) {
+		pr_warn("CPU with hartid=%lu does not support rv64ima", *hart);
 		return -ENODEV;
+	}
 
 	return 0;
 }
-- 
2.40.1