Skip to contentSkip to conversation

mm: Support compiling out madvise and fadvise

PATCH Superseded Josh Triplett 3 participants 6 messages Sep 22, 2014 – Sep 25, 2014

Subsystems (2)
Memory management ABI/API
lore
Share & tools
lore

Josh Triplett

author
#0 lore raw

PATCH 1/1 mm: Support compiling out madvise and fadvise

Many embedded systems will not need these syscalls, and omitting them saves space. Add a new EXPERT config option CONFIG_ADVISE_SYSCALLS (default y) to support compiling them out.

bloat-o-meter:
add/remove: 0/3 grow/shrink: 0/0 up/down: 0/-2250 (-2250)
function old new delta

sys_fadvise64                                 57       -     -57
sys_fadvise64_64                             691       -    -691
sys_madvise                                 1502       -   -1502
Signed-off-by Josh Triplett <[email protected]>

---

Posting this for review. I can upstream this directly through my new tiny tree.

 init/Kconfig    | 10 ++++++++++
 kernel/sys_ni.c |  3 +++
 mm/Makefile     |  7 +++++--
 3 files changed, 18 insertions(+), 2 deletions(-)
init/Kconfig +10 -0
index e84c642..782a65b 100644
@@ -1537,6 +1537,16 @@ config AIO
 	  by some high performance threaded applications. Disabling
 	  this option saves about 7k.
 
+config ADVISE_SYSCALLS
+	bool "Enable madvise/fadvise syscalls" if EXPERT
+	default y
+	help
+	  This option enables the madvise and fadvise syscalls, used by
+	  applications to advise the kernel about their future memory or file
+	  usage, improving performance. If building an embedded system where no
+	  applications use these syscalls, you can disable this option to save
+	  space.
+
 config PCI_QUIRKS
 	default y
 	bool "Enable PCI quirk workarounds" if EXPERT
kernel/sys_ni.c +3 -0
index 391d4dd..d4709d4 100644
@@ -156,6 +156,9 @@ cond_syscall(sys_process_vm_writev);
 cond_syscall(compat_sys_process_vm_readv);
 cond_syscall(compat_sys_process_vm_writev);
 cond_syscall(sys_uselib);
+cond_syscall(sys_fadvise64);
+cond_syscall(sys_fadvise64_64);
+cond_syscall(sys_madvise);
 
 /* arch-specific weak syscall entries */
 cond_syscall(sys_pciconfig_read);
mm/Makefile +5 -2
index 632ae77..fe7a053 100644
@@ -3,7 +3,7 @@
 #
 
 mmu-y			:= nommu.o
-mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o madvise.o memory.o mincore.o \
+mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o memory.o mincore.o \
 			   mlock.o mmap.o mprotect.o mremap.o msync.o rmap.o \
 			   vmalloc.o pagewalk.o pgtable-generic.o
 

@@ -11,7 +11,7 @@ ifdef CONFIG_CROSS_MEMORY_ATTACH
 mmu-$(CONFIG_MMU)	+= process_vm_access.o
 endif
 
-obj-y			:= filemap.o mempool.o oom_kill.o fadvise.o \
+obj-y			:= filemap.o mempool.o oom_kill.o \
 			   maccess.o page_alloc.o page-writeback.o \
 			   readahead.o swap.o truncate.o vmscan.o shmem.o \
 			   util.o mmzone.o vmstat.o backing-dev.o \

@@ -28,6 +28,9 @@ else
 	obj-y		+= bootmem.o
 endif
 
+ifdef CONFIG_MMU
+	obj-$(CONFIG_ADVISE_SYSCALLS)	+= fadvise.o madvise.o
+endif
 obj-$(CONFIG_HAVE_MEMBLOCK) += memblock.o
 
 obj-$(CONFIG_SWAP)	+= page_io.o swap_state.o swapfile.o
Signature · 1 line
2.1.0
#1 lore raw
Quoted 3 lines · Josh Triplett

Many embedded systems will not need these syscalls, and omitting them saves space. Add a new EXPERT config option CONFIG_ADVISE_SYSCALLS (default y) to support compiling them out.

general question: if a user chooses CONFIG_ADVISE_SYSCALLS=n (or any config option related to tinyfication) and breaks the system/workload... will that be acceptable for a kernel pov? In other words, what's the degree of responsibility the user will have when choosing such builds?

Thanks,
Davidlohr

author
#2 lore raw
Quoted 9 lines · Davidlohr Bueso
Quoted 3 lines · Josh Triplett

Many embedded systems will not need these syscalls, and omitting them saves space. Add a new EXPERT config option CONFIG_ADVISE_SYSCALLS (default y) to support compiling them out.

general question: if a user chooses CONFIG_ADVISE_SYSCALLS=n (or any config option related to tinyfication) and breaks the system/workload... will that be acceptable for a kernel pov? In other words, what's the degree of responsibility the user will have when choosing such builds?

It's hidden behind EXPERT for exactly that reason: if you turn it off, and your userspace needs it and can't cope with ENOSYS, you get to keep all the pieces. Only turn it off if you know your userspace doesn't use it.

The same thing goes for several other such options, such as UID16, SYSCTL_SYSCALL, SGETMASK_SYSCALL, and USELIB.

- Josh Triplett

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

#3 lore raw
In reply to Josh Triplett
Quoted 19 lines · Josh Triplett
mm/Makefile +2 -2
@@ -3,7 +3,7 @@
 #
 
 mmu-y			:= nommu.o
-mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o madvise.o memory.o mincore.o \
+mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o memory.o mincore.o \
 			   mlock.o mmap.o mprotect.o mremap.o msync.o rmap.o \
 			   vmalloc.o pagewalk.o pgtable-generic.o
 

@@ -11,7 +11,7 @@ ifdef CONFIG_CROSS_MEMORY_ATTACH
 mmu-$(CONFIG_MMU)	+= process_vm_access.o
 endif
 
-obj-y			:= filemap.o mempool.o oom_kill.o fadvise.o \
+obj-y			:= filemap.o mempool.o oom_kill.o \
 			   maccess.o page_alloc.o page-writeback.o \
 			   readahead.o swap.o truncate.o vmscan.o shmem.o \
 			   util.o mmzone.o vmstat.o backing-dev.o \

mm/Makefile else · line 33

Quoted 6 lines
mm/Makefile +3 -0
@@ -28,6 +28,9 @@ else
 	obj-y		+= bootmem.o
 endif
 
+ifdef CONFIG_MMU
+	obj-$(CONFIG_ADVISE_SYSCALLS)	+= fadvise.o madvise.o
+endif

That makes fadvise MMU-only, but I don't see why it should be.

Was that intentional?

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

author
#4 lore raw
Quoted 23 lines · Johannes Weiner
Quoted 22 lines · Josh Triplett
mm/Makefile +2 -2
@@ -3,7 +3,7 @@
 #
 
 mmu-y			:= nommu.o
-mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o madvise.o memory.o mincore.o \
+mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o memory.o mincore.o \
 			   mlock.o mmap.o mprotect.o mremap.o msync.o rmap.o \
 			   vmalloc.o pagewalk.o pgtable-generic.o
 

@@ -11,7 +11,7 @@ ifdef CONFIG_CROSS_MEMORY_ATTACH
 mmu-$(CONFIG_MMU)	+= process_vm_access.o
 endif
 
-obj-y			:= filemap.o mempool.o oom_kill.o fadvise.o \
+obj-y			:= filemap.o mempool.o oom_kill.o \
 			   maccess.o page_alloc.o page-writeback.o \
 			   readahead.o swap.o truncate.o vmscan.o shmem.o \
 			   util.o mmzone.o vmstat.o backing-dev.o \

@@ -28,6 +28,9 @@ else
 	obj-y		+= bootmem.o
 endif
Quoted 7 lines

mm/Makefile else · line 33

Quoted 3 lines
mm/Makefile +3 -0
+ifdef CONFIG_MMU
+	obj-$(CONFIG_ADVISE_SYSCALLS)	+= fadvise.o madvise.o
+endif

That makes fadvise MMU-only, but I don't see why it should be.

Was that intentional?

No. Fixed in v2; will send out momentarily. Thanks!

- Josh Triplett

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>

author
#5 lore raw

PATCH v2 1/1 mm: Support compiling out madvise and fadvise

In reply to Josh Triplett

Many embedded systems will not need these syscalls, and omitting them saves space. Add a new EXPERT config option CONFIG_ADVISE_SYSCALLS (default y) to support compiling them out.

bloat-o-meter:
add/remove: 0/3 grow/shrink: 0/0 up/down: 0/-2250 (-2250)
function old new delta

sys_fadvise64                                 57       -     -57
sys_fadvise64_64                             691       -    -691
sys_madvise                                 1502       -   -1502
Signed-off-by Josh Triplett <[email protected]>

---

v2: Don't make fadvise depend on CONFIG_MMU.

 init/Kconfig    | 10 ++++++++++
 kernel/sys_ni.c |  3 +++
 mm/Makefile     |  8 ++++++--
 3 files changed, 19 insertions(+), 2 deletions(-)
init/Kconfig +10 -0
index e84c642..782a65b 100644
@@ -1537,6 +1537,16 @@ config AIO
 	  by some high performance threaded applications. Disabling
 	  this option saves about 7k.
 
+config ADVISE_SYSCALLS
+	bool "Enable madvise/fadvise syscalls" if EXPERT
+	default y
+	help
+	  This option enables the madvise and fadvise syscalls, used by
+	  applications to advise the kernel about their future memory or file
+	  usage, improving performance. If building an embedded system where no
+	  applications use these syscalls, you can disable this option to save
+	  space.
+
 config PCI_QUIRKS
 	default y
 	bool "Enable PCI quirk workarounds" if EXPERT
kernel/sys_ni.c +3 -0
index 391d4dd..d4709d4 100644
@@ -156,6 +156,9 @@ cond_syscall(sys_process_vm_writev);
 cond_syscall(compat_sys_process_vm_readv);
 cond_syscall(compat_sys_process_vm_writev);
 cond_syscall(sys_uselib);
+cond_syscall(sys_fadvise64);
+cond_syscall(sys_fadvise64_64);
+cond_syscall(sys_madvise);
 
 /* arch-specific weak syscall entries */
 cond_syscall(sys_pciconfig_read);
mm/Makefile +6 -2
index 632ae77..2ad574d 100644
@@ -3,7 +3,7 @@
 #
 
 mmu-y			:= nommu.o
-mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o madvise.o memory.o mincore.o \
+mmu-$(CONFIG_MMU)	:= fremap.o gup.o highmem.o memory.o mincore.o \
 			   mlock.o mmap.o mprotect.o mremap.o msync.o rmap.o \
 			   vmalloc.o pagewalk.o pgtable-generic.o
 

@@ -11,7 +11,7 @@ ifdef CONFIG_CROSS_MEMORY_ATTACH
 mmu-$(CONFIG_MMU)	+= process_vm_access.o
 endif
 
-obj-y			:= filemap.o mempool.o oom_kill.o fadvise.o \
+obj-y			:= filemap.o mempool.o oom_kill.o \
 			   maccess.o page_alloc.o page-writeback.o \
 			   readahead.o swap.o truncate.o vmscan.o shmem.o \
 			   util.o mmzone.o vmstat.o backing-dev.o \

@@ -28,6 +28,10 @@ else
 	obj-y		+= bootmem.o
 endif
 
+obj-$(CONFIG_ADVISE_SYSCALLS)	+= fadvise.o
+ifdef CONFIG_MMU
+	obj-$(CONFIG_ADVISE_SYSCALLS)	+= madvise.o
+endif
 obj-$(CONFIG_HAVE_MEMBLOCK) += memblock.o
 
 obj-$(CONFIG_SWAP)	+= page_io.o swap_state.o swapfile.o
Signature · 1 line
2.1.1