ARM: 9140/1: allow compile-testing without machine record

A lot of randconfig builds end up not selecting any machine type at
all. This is generally fine for the purpose of compile testing, but
of course it means that the kernel is not usable on actual hardware,
and it causes a warning about this fact.

As most of the build bots now force-enable CONFIG_COMPILE_TEST for
randconfig builds, use that as a guard to control whether we warn
on this type of broken configuration.

We could do the same for the missing-cpu-type warning, but those
configurations fail to build much earlier.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>

diff --git a/arch/arm/kernel/vmlinux-xip.lds.S b/arch/arm/kernel/vmlinux-xip.lds.S
index 5013682..97adb4b 100644 (file)
--- a/arch/arm/kernel/vmlinux-xip.lds.S
+++ b/arch/arm/kernel/vmlinux-xip.lds.S
@@ -162,7 +162,9 @@ SECTIONS
  * binutils is too old (for other reasons as well)
  */
 ASSERT((__proc_info_end - __proc_info_begin), "missing CPU support")
+#ifndef CONFIG_COMPILE_TEST
 ASSERT((__arch_info_end - __arch_info_begin), "no machine record defined")
+#endif
 
 #ifdef CONFIG_XIP_DEFLATED_DATA
 /*

diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
index 20c4f6d..f02d617 100644 (file)
--- a/arch/arm/kernel/vmlinux.lds.S
+++ b/arch/arm/kernel/vmlinux.lds.S
@@ -174,6 +174,8 @@ __start_rodata_section_aligned = ALIGN(__start_rodata, 1 << SECTION_SHIFT);
  * binutils is too old (for other reasons as well)
  */
 ASSERT((__proc_info_end - __proc_info_begin), "missing CPU support")
+#ifndef CONFIG_COMPILE_TEST
 ASSERT((__arch_info_end - __arch_info_begin), "no machine record defined")
+#endif
 
 #endif /* CONFIG_XIP_KERNEL */