Unhide DEBUG_RODATA if EXPERT, even if DEBUG_KERNEL=n
PATCH 1/1 Unhide DEBUG_RODATA if EXPERT, even if DEBUG_KERNEL=n
CONFIG_DEBUG_RODATA defaults to y. Embedded systems might want to disable this to save space; however, disabling it requires enabling CONFIG_DEBUG_KERNEL to see it. Unhide it for CONFIG_EXPERT=y as well, so it shows up as an option for embedded users without having to turn on CONFIG_DEBUG_KERNEL.
---
On a related note, I noticed while making this patch that DEBUG_RODATA_TEST defaults to y, as of commit 72370f2a5b227bd3817593a6b15ea3f53f51dfcb; however, the description says "If in doubt, say N". Should the default change, or should the description? DEBUG_RODATA_TEST seems like the kind of thing that should default to n.
arch/x86/Kconfig.debug | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/x86/Kconfig.debug +1 -1
@@ -81,7 +81,7 @@ config X86_PTDUMP config DEBUG_RODATA bool "Write protect kernel read-only data structures" default y - depends on DEBUG_KERNEL + depends on DEBUG_KERNEL || EXPERT ---help--- Mark the kernel read-only data as write-protected in the pagetables, in order to catch accidental (and incorrect) writes to such const
Signature · 1 line
1.7.5.3
arch/x86/Kconfig.debug config X86_PTDUMP · line 84
Quoted 8 lines · Josh Triplett
arch/x86/Kconfig.debug +1
-1@@ -81,7 +81,7 @@ config X86_PTDUMP config DEBUG_RODATA bool "Write protect kernel read-only data structures" default y - depends on DEBUG_KERNEL + depends on DEBUG_KERNEL || EXPERT
This is doing it the wrong way around. The right solution is what i wrote for verbose-debug: if we select DEBUG_KERNEL from EXPERT then you will avoid having to expand || EXPERT conditions into every default-enabled debug facility ...
My suggested solution will be a lot easier to maintain as well as under your scheme, because with your scheme we'd have to:
- add || EXPERT every time a debug facility is enabled by default
- remove || EXPERT every time a debug facility is switched to off-by-default
If EXPERT selects DEBUG_KERNEL then we always get the right behavior, without polluting debug Kconfig's with EXPERT logic ...
Thanks,
Ingo