ramster: depend on CONFIG_NET for sock_* functions
PATCH 1/1 ramster: depend on CONFIG_NET for sock_* functions
Building ramster without CONFIG_NET would cause linkage issue due to missing sock_*() functions in cluster/tcp.c
--- drivers/staging/ramster/Kconfig | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
drivers/staging/ramster/Kconfig +1 -1
@@ -1,6 +1,6 @@ config RAMSTER bool "Cross-machine RAM capacity sharing, aka peer-to-peer tmem" - depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM + depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM && CONFIG_NET select LZO_COMPRESS select LZO_DECOMPRESS default n
Signature · 1 line
1.7.8.5
Quoted 1 line · Sasha Levin
Building ramster without
CONFIG_NETwould cause linkage issue due to missing
I'd say s/CONFIG_NET/NET/
Quoted 12 lines
sock_*() functions in
cluster/tcp.cSigned-off-by Sasha Levin <[email protected]>--- drivers/staging/ramster/Kconfig | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)drivers/staging/ramster/Kconfig +0
-0
drivers/staging/ramster/Kconfig line 3
Quoted 4 lines
drivers/staging/ramster/Kconfig +1
-1@@ -1,6 +1,6 @@ config RAMSTER bool "Cross-machine RAM capacity sharing, aka peer-to-peer tmem" - depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM + depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM && CONFIG_NET
s/CONFIG_NET/NET/ again, I'd say. How was this tested?
drivers/staging/ramster/Kconfig line 6
Quoted 3 lines
drivers/staging/ramster/Kconfig +0
-0select LZO_COMPRESS select LZO_DECOMPRESS default n
Paul Bolle
Quoted 12 lines · Paul Bolle
Quoted 1 line · Sasha Levin
Building ramster without
CONFIG_NETwould cause linkage issue due to missingI'd say s/CONFIG_NET/NET/
Quoted 7 lines
sock_*() functions in
cluster/tcp.cSigned-off-by Sasha Levin <[email protected]>---
drivers/staging/ramster/Kconfig | 2 +-1 files changed, 1 insertions(+), 1 deletions(-)
Quoted 11 lines
Quoted 9 lines
drivers/staging/ramster/Kconfig +0
-0@@ -1,6 +1,6 @@config RAMSTER
bool "Cross-machine RAM capacity sharing, aka peer-to-peer tmem"- depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM + depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM && CONFIG_NETs/CONFIG_NET/NET/ again, I'd say. How was this tested?
Hm... ran the build again with the patch and noticed it stopped complaining, but apparently it wasn't enough.
Quoted 12 lines · Sasha Levin
Quoted 10 lines · Paul Bolle
drivers/staging/ramster/Kconfig line 3
Quoted 7 lines · Sasha Levin
drivers/staging/ramster/Kconfig +1
-1@@ -1,6 +1,6 @@ config RAMSTER bool "Cross-machine RAM capacity sharing, aka peer-to-peer tmem" - depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM + depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM && CONFIG_NETs/CONFIG_NET/NET/ again, I'd say. How was this tested?
Quoted 2 lines
Hm... ran the build again with the patch and noticed it stopped complaining, but apparently it wasn't enough.
Well, my guess is that the errors you ran into weren't displayed because a dependency on an undefined Kconfig symbol (like CONFIG_NET is) will always fail. Ie, your patch made "RAM capacity sharing" simply disappear from the build.
Paul Bolle
Quoted 13 lines · Paul Bolle
Quoted 12 lines · Sasha Levin
Quoted 10 lines · Paul Bolle
Quoted 7 lines · Sasha Levin
drivers/staging/ramster/Kconfig +0
-0@@ -1,6 +1,6 @@config RAMSTER
bool "Cross-machine RAM capacity sharing, aka peer-to-peer tmem"- depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM + depends on (CLEANCACHE || FRONTSWAP) && CONFIGFS_FS=y && !ZCACHE && !XVMALLOC && !HIGHMEM && CONFIG_NETs/CONFIG_NET/NET/ again, I'd say. How was this tested?
Quoted 7 lines
Quoted 2 lines
Hm... ran the build again with the patch and noticed it stopped complaining, but apparently it wasn't enough.
Well, my guess is that the errors you ran into weren't displayed because a dependency on an undefined
Kconfigsymbol (likeCONFIG_NETis) will always fail. Ie, your patch made "RAM capacity sharing" simply disappear from the build.
Yup, which is not what I meant to do :)
Sorry about that.