package/glibc: only supports ARC little-endian since glibc 2.41

Since upstream commit
https://sourceware.org/git/?p=glibc.git;a=commit;h=4ff55d08df0e6f11cbd217f4dca4532bd7d9330d,
glibc no longer supports ARC big-endian, and the configure step fails
with:

checking for egrep -e... (cached) /bin/grep -E
checking for big endian... yes
configure: error: big-endian not supported
make: *** [package/pkg-generic.mk:263: /home/buildroot/instance-0/output-1/build/glibc-2.41-70-g1502c248d58cb99a203731707987a4342926e830/.stamp_configured] Error 1

This commit was first merged in glibc 2.41, so this issue is
applicable to both master (glibc 2.42) and 2025.02.x (glibc 2.41).

Fixes:

  https://autobuild.buildroot.net/results/87e4b84f3aee64d041c3b845cb003e6b90649ffb/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>

package/glibc/Config.in

@@ -21,7 +21,7 @@ config BR2_PACKAGE_GLIBC_ARCH_SUPPORTS
 	default y if BR2_sparc64
 	default y if BR2_x86_64
 	default y if BR2_microblaze
-	default y if BR2_arc && BR2_ARC_ATOMIC_EXT && !BR2_arc750d && !BR2_arc770d
+	default y if BR2_arcle && BR2_ARC_ATOMIC_EXT && !BR2_arc750d && !BR2_arc770d
 	depends on !BR2_POWERPC_CPU_HAS_SPE
 	# glibc needs atomic instructions, and does not support
 	# single-precision floating point ABIs (ilp32f and lp64f)