[PATCH] MIPS: OCTEON: program L2C_CTL on CNF71XX too

[Orgad Shaneh <orgads@gmail.com>]

Since commit bd978130307c ("MIPS: OCTEON: program L2C_CTL, not L2C_CFG,
on OCTEON II"), cvmx_helper_initialize_packet_io_global() writes L2C_CTL
instead of L2C_CFG, because L2C_CFG does not exist on OCTEON II and
writing it hangs a CN6335. The test is OCTEON_CN6XXX, which leaves the
CNF71XX - also an OCTEON II, as OCTEON_IS_OCTEON2() and the rest of this
file treat it - writing the register that commit says is not there.

Use OCTEON_IS_OCTEON2(). OCTEON III (CN7XXX) still takes the L2C_CFG
path, as before: I have no such part and do not know which register it
wants.

Fixes: bd978130307c ("MIPS: OCTEON: program L2C_CTL, not L2C_CFG, on OCTEON II")
Assisted-by: LLM
Signed-off-by: Orgad Shaneh <orgads@gmail.com>
---
Found by the Sashiko review bot on the original series. By inspection
only: I have no CNF71XX board to test on. The Fixes: hash is the one in
mips-next.

diff --git a/arch/mips/cavium-octeon/executive/cvmx-helper.c b/arch/mips/cavium-octeon/executive/cvmx-helper.c
index 107b1be..c02448b 100644
--- a/arch/mips/cavium-octeon/executive/cvmx-helper.c
+++ b/arch/mips/cavium-octeon/executive/cvmx-helper.c
@@ -1028,7 +1028,7 @@ int cvmx_helper_initialize_packet_io_global(void)
 	 * to the cores. This avoids conditions where IO blocks might
 	 * be starved under very high L2 loads.
 	 */
-	if (OCTEON_IS_MODEL(OCTEON_CN6XXX)) {
+	if (OCTEON_IS_OCTEON2()) {
 		union cvmx_l2c_ctl l2c_ctl;
 
 		l2c_ctl.u64 = cvmx_read_csr(CVMX_L2C_CTL);