Skip to contentSkip to conversation

x86: remove dupilcated #include

PATCH Applied Huang Weiyi 3 participants 6 messages Apr 6, 2009 – Apr 9, 2009

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 4/8 x86: remove dupilcated #include

Remove dupilcated #include in arch/x86/kernel/dumpstack.c.

---
 arch/x86/kernel/dumpstack.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)
arch/x86/kernel/dumpstack.c +0 -1
index 95ea5fa..6d7966d 100644
@@ -15,7 +15,6 @@
 #include <linux/bug.h>
 #include <linux/nmi.h>
 #include <linux/sysfs.h>
-#include <linux/ftrace.h>
 
 #include <asm/stacktrace.h>
 
Signature · 1 line
1.6.1.3
#1 lore raw
Quoted 12 lines · Huang Weiyi

Remove dupilcated #include in arch/x86/kernel/dumpstack.c.

---
 arch/x86/kernel/dumpstack.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)
arch/x86/kernel/dumpstack.c +0 -0
index 95ea5fa..6d7966d 100644

arch/x86/kernel/dumpstack.c line 19

Quoted 6 lines
arch/x86/kernel/dumpstack.c +0 -1
@@ -15,7 +15,6 @@
 #include <linux/bug.h>
 #include <linux/nmi.h>
 #include <linux/sysfs.h>
-#include <linux/ftrace.h>
 
 #include <asm/stacktrace.h>

Many of those include lines are probably unnecessary. Instead of these trivial patches causing churn, would you be interested in doing a comprehensive search to eliminate all the unused ones? That would be a real step forward. (and this holds for your other patches in this series too.)

The include files section of fault.c might be a good template to use:

#include <linux/magic.h> /* STACK_END_MAGIC */
#include <linux/sched.h> /* test_thread_flag(), ... */
#include <linux/kdebug.h> /* oops_begin/end, ... */
#include <linux/module.h> /* search_exception_table */
#include <linux/bootmem.h> /* max_low_pfn */
#include <linux/kprobes.h> /* __kprobes, ... */
#include <linux/mmiotrace.h> /* kmmio_handler, ... */
#include <linux/perf_counter.h> /* perf_swcounter_*(), ... */

#include <asm/traps.h> /* dotraplinkage, ... */
#include <asm/pgalloc.h> /* pgd_*(), ... */
#include <asm/kmemcheck.h> /* kmemcheck_*(), ... */

I was able to eliminate half of all include file lines there.

See latest -tip for that:

http://people.redhat.com/mingo/tip.git/README

Thanks,

Ingo

#2 lore raw
Quoted 40 lines · Ingo Molnar
Quoted 12 lines · Huang Weiyi

Remove dupilcated #include in arch/x86/kernel/dumpstack.c.

---
 arch/x86/kernel/dumpstack.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)
arch/x86/kernel/dumpstack.c +0 -0
index 95ea5fa..6d7966d 100644

arch/x86/kernel/dumpstack.c line 19

Quoted 6 lines
arch/x86/kernel/dumpstack.c +0 -1
@@ -15,7 +15,6 @@
 #include <linux/bug.h>
 #include <linux/nmi.h>
 #include <linux/sysfs.h>
-#include <linux/ftrace.h>
 
 #include <asm/stacktrace.h>

Many of those include lines are probably unnecessary. Instead of these trivial patches causing churn, would you be interested in doing a comprehensive search to eliminate all the unused ones? That would be a real step forward. (and this holds for your other patches in this series too.)

The include files section of fault.c might be a good template to use:

#include <linux/magic.h> /* STACK_END_MAGIC */
#include <linux/sched.h> /* test_thread_flag(), ... */
#include <linux/kdebug.h> /* oops_begin/end, ... */
#include <linux/module.h> /* search_exception_table */
#include <linux/bootmem.h> /* max_low_pfn */
#include <linux/kprobes.h> /* __kprobes, ... */
#include <linux/mmiotrace.h> /* kmmio_handler, ... */
#include <linux/perf_counter.h> /* perf_swcounter_*(), ... */

Quoted 5 lines

#include <asm/traps.h> /* dotraplinkage, ... */
#include <asm/pgalloc.h> /* pgd_*(), ... */
#include <asm/kmemcheck.h> /* kmemcheck_*(), ... */

I was able to eliminate half of all include file lines there.

I assume you are aware that when you minimize the # of include file in the various .c files, then you implicitly add dependency on the includes the individual .h files have.

If we for example killed the include of <linux/list.h> from <linux/module.h> with this approach we would likely see random build failures.

It is good to have it cleaned up but we should keep a critical eye on the header files we use. Sometimes these are the right target
to clean up but as this is a bit more complicated or at least requires more build testing to do right.

Sam

Acked-by H. Peter Anvin
#3 lore raw
Quoted 45 lines · Sam Ravnborg
Quoted 44 lines · Ingo Molnar
Quoted 12 lines · Huang Weiyi

Remove dupilcated #include in arch/x86/kernel/dumpstack.c.

---
 arch/x86/kernel/dumpstack.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)
arch/x86/kernel/dumpstack.c +0 -0
index 95ea5fa..6d7966d 100644

arch/x86/kernel/dumpstack.c line 19

Quoted 6 lines
arch/x86/kernel/dumpstack.c +0 -1
@@ -15,7 +15,6 @@
 #include <linux/bug.h>
 #include <linux/nmi.h>
 #include <linux/sysfs.h>
-#include <linux/ftrace.h>
 
 #include <asm/stacktrace.h>

Many of those include lines are probably unnecessary. Instead of these trivial patches causing churn, would you be interested in doing a comprehensive search to eliminate all the unused ones? That would be a real step forward. (and this holds for your other patches in this series too.)

The include files section of fault.c might be a good template to use:

#include <linux/magic.h> /* STACK_END_MAGIC */
#include <linux/sched.h> /* test_thread_flag(), ... */
#include <linux/kdebug.h> /* oops_begin/end, ... */
#include <linux/module.h> /* search_exception_table */
#include <linux/bootmem.h> /* max_low_pfn */
#include <linux/kprobes.h> /* __kprobes, ... */
#include <linux/mmiotrace.h> /* kmmio_handler, ... */
#include <linux/perf_counter.h> /* perf_swcounter_*(), ... */

#include <asm/traps.h> /* dotraplinkage, ... */
#include <asm/pgalloc.h> /* pgd_*(), ... */
#include <asm/kmemcheck.h> /* kmemcheck_*(), ... */

Quoted 5 lines
Quoted 1 line

I was able to eliminate half of all include file lines there.

I assume you are aware that when you minimize the # of include file in the various .c files, then you implicitly add dependency on the includes the individual .h files have.

Yes. Look at the commit -tip that does the above change (also attached below):

a2bcd47: x86/mm: further cleanups of fault.c's include file section

that commit uncovered a masked-until-then dependency bug in one of the x86 include files.

More automation to discover include file dependency bugs would be nice though, it's hard to get these things right and it all needs a strong testing infrastructure.

Ingo

----------------->

Quoted 1 line

From a2bcd4731f77cb77ae4b5e4a3d7f5471cf346c33 Mon Sep 17 00:00:00 2001

From: Ingo Molnar <[email protected]>
Date: Sun, 29 Mar 2009 23:47:48 +0200
Subject: [PATCH] x86/mm: further cleanups of fault.c's include file section

Impact: cleanup

Eliminate more than 20 unnecessary #include lines in fault.c

Also fix include file dependency bug in asm/traps.h. (this was masked before, by implicit inclusion)

LKML-Reference: <[email protected]>

---
 arch/x86/include/asm/traps.h |    1 +
 arch/x86/mm/fault.c          |   42 +++++++++---------------------------------
 2 files changed, 10 insertions(+), 33 deletions(-)
arch/x86/include/asm/traps.h +1 -0
index 0d53425..37fb07a 100644
@@ -2,6 +2,7 @@
 #define _ASM_X86_TRAPS_H
 
 #include <asm/debugreg.h>
+#include <asm/siginfo.h>			/* TRAP_TRACE, ... */
 
 #ifdef CONFIG_X86_32
 #define dotraplinkage
arch/x86/mm/fault.c +10 -34

50 lines not on this page · raw message

#4 lore raw
Quoted 5 lines
Quoted 4 lines

I assume you are aware that when you minimize the # of include file in the various .c files, then you implicitly add dependency on the includes the individual .h files have.

Quoted 7 lines

Yes. Look at the commit -tip that does the above change (also attached below):

a2bcd47: x86/mm: further cleanups of fault.c's include file section

that commit uncovered a masked-until-then dependency bug in one of the x86 include files.

We have plenty of header files that needs other header files to be included in order to build.

I tried to do a simple build check with a i386 defconfig. It revealed that the appended list of header files did not build.

What I did was to create a simple:

module.h.c file containing:
#include <linux/module.h>

And then I used kbuild to build the file,
adding "obj-y += module.h.o" to Kbuild.

I did so for all header files in include/linux/ and result was that 172 header files failed to build.

Some of these was expected since they are not supposed to be included direct. But the rest should build have build with no errors?

I can automate the above included some way to say which header files we should NOT build.
But eliminating the includes that are not needed is a bigger challenge and also the biggest win :-(
Is it worthwhile to make the header files buildable alone?

Sam

include/linux/ac97_codec.h
include/linux/adfs_fs.h
include/linux/adfs_fs_i.h
include/linux/adfs_fs_sb.h
include/linux/arcdevice.h
include/linux/atalk.h
include/linux/ata_platform.h
include/linux/ath9k_platform.h
include/linux/atmel_pwm.h
include/linux/atm_zatm.h
include/linux/bit_spinlock.h
include/linux/bsg.h
include/linux/cgroup_subsys.h
include/linux/chio.h
include/linux/coda_fs_i.h
include/linux/coda_psdev.h
include/linux/compiler-gcc3.h
include/linux/compiler-gcc4.h
include/linux/compiler-gcc.h
include/linux/compiler-intel.h
include/linux/console_struct.h
include/linux/cramfs_fs_sb.h
include/linux/cryptohash.h
include/linux/cyclades.h
include/linux/cycx_drv.h
include/linux/cycx_x25.h
include/linux/dca.h
include/linux/dio.h
include/linux/dirent.h
include/linux/dlm_netlink.h
include/linux/dlm_plock.h
include/linux/dm9000.h
include/linux/eeprom_93cx6.h
include/linux/efs_vh.h
include/linux/elfcore-compat.h
include/linux/eventpoll.h
include/linux/ext2_fs.h
include/linux/ext3_fs.h
include/linux/ext3_fs_i.h
include/linux/ext3_fs_sb.h
include/linux/f75375s.h
include/linux/fault-inject.h
include/linux/fiemap.h
include/linux/filter.h
include/linux/flat.h
include/linux/fscache-cache.h
include/linux/fs_struct.h
include/linux/fs_uart_pd.h
include/linux/genalloc.h
include/linux/hayesesp.h
include/linux/hid-debug.h
include/linux/hiddev.h
include/linux/hpet.h
include/linux/hp_sdc.h
include/linux/htirq.h
include/linux/hwmon-sysfs.h
include/linux/hwmon-vid.h
include/linux/i2c-ocores.h
include/linux/i2c-pnx.h
include/linux/ibmtr.h
include/linux/if_ec.h
include/linux/if_ppp.h
include/linux/if_tunnel.h
include/linux/init_ohci1394_dma.h
include/linux/ioc3.h
include/linux/iommu.h
include/linux/iommu-helper.h
include/linux/ip6_tunnel.h
include/linux/ipv6_route.h
include/linux/irda.h
include/linux/irq_cpustat.h
include/linux/iscsi_ibft.h
include/linux/isdn_ppp.h
include/linux/istallion.h
include/linux/ixjuser.h
include/linux/jhash.h
include/linux/kdev_t.h
include/linux/kernelcapi.h
include/linux/kmalloc_sizes.h
include/linux/kvm_host.h
include/linux/kvm_para.h
include/linux/leds-bd2802.h
include/linux/leds_pwm.h
include/linux/libps2.h
include/linux/license.h
include/linux/llc.h
include/linux/lmb.h
include/linux/lzo.h
include/linux/maple.h
include/linux/mbcache.h
include/linux/mbus.h
include/linux/mc6821.h
include/linux/mm_inline.h
include/linux/mpage.h
include/linux/mroute6.h
include/linux/mv643xx_eth.h
include/linux/ncp_fs_i.h
include/linux/netrom.h
include/linux/nfs4_mount.h
include/linux/nfs_fs_sb.h
include/linux/nfs.h
include/linux/nfs_page.h
include/linux/nfs_xdr.h
include/linux/nilfs2_fs.h
include/linux/nls.h
include/linux/n_r3964.h
include/linux/nsc_gpio.h
include/linux/nubus.h
include/linux/of_device.h
include/linux/of.h
include/linux/of_platform.h
include/linux/of_spi.h
include/linux/page_cgroup.h
include/linux/pagevec.h
include/linux/parport_pc.h
include/linux/parser.h
include/linux/patchkey.h
include/linux/pci-acpi.h
include/linux/pcieport_if.h
include/linux/pci_hotplug.h
include/linux/pda_power.h
include/linux/phonet.h
include/linux/phy_fixed.h
include/linux/pipe_fs_i.h
include/linux/pm_wakeup.h
include/linux/prio_tree.h
include/linux/qnx4_fs.h
include/linux/ramfs.h
include/linux/rcuclassic.h
include/linux/rcupreempt.h
include/linux/reiserfs_fs_i.h
include/linux/reiserfs_fs_sb.h
include/linux/reiserfs_xattr.h
include/linux/romfs_fs.h
include/linux/rose.h
include/linux/rwsem-spinlock.h
include/linux/scc.h
include/linux/scx200_gpio.h
include/linux/sdla.h
include/linux/selection.h
include/linux/selinux.h
include/linux/serial167.h
include/linux/serialP.h
include/linux/sh_intc.h
include/linux/signalfd.h
include/linux/slab_def.h
include/linux/slob_def.h
include/linux/slub_def.h
include/linux/smb_fs_sb.h
include/linux/smb_mount.h
include/linux/spinlock_api_smp.h
include/linux/spinlock_api_up.h
include/linux/spinlock_types_up.h
include/linux/spinlock_up.h
include/linux/splice.h
include/linux/srcu.h
include/linux/stallion.h
include/linux/svga.h
include/linux/swapops.h
include/linux/sysv_fs.h
include/linux/task_io_accounting.h
include/linux/tick.h
include/linux/tty_driver.h
include/linux/tty_flip.h
include/linux/un.h
include/linux/usb_usual.h
include/linux/videotext.h
include/linux/virtio_net.h
include/linux/vmstat.h
include/linux/vt_buffer.h
include/linux/x25.h
include/linux/zorro.h

#5 lore raw
Quoted 34 lines · Sam Ravnborg
Quoted 5 lines
Quoted 4 lines

I assume you are aware that when you minimize the # of include file in the various .c files, then you implicitly add dependency on the includes the individual .h files have.

Quoted 7 lines

Yes. Look at the commit -tip that does the above change (also attached below):

a2bcd47: x86/mm: further cleanups of fault.c's include file section

that commit uncovered a masked-until-then dependency bug in one of the x86 include files.

We have plenty of header files that needs other header files to be included in order to build.

I tried to do a simple build check with a i386 defconfig. It revealed that the appended list of header files did not build.

What I did was to create a simple:

module.h.c file containing:
#include <linux/module.h>

And then I used kbuild to build the file,
adding "obj-y += module.h.o" to Kbuild.

I did so for all header files in include/linux/ and result was that 172 header files failed to build.

Some of these was expected since they are not supposed to be included direct. But the rest should build have build with no errors?

Quoted 5 lines

I can automate the above included some way to say which header files we should NOT build.
But eliminating the includes that are not needed is a bigger challenge and also the biggest win :-(
Is it worthwhile to make the header files buildable alone?

I think it is definitely worthwile, for a couple of reasons:

- a data type definition that does not even build if included in a .c file is broken, almost by definition. If it happens to build in its current include file environment that's just a hidden bug, nothing more.

- having such a check would simplify header dependency hell removal efforts _immensely_. I could remove a suspect looking #include line from a header and build with your check - that would be a pretty good mechanism to drive header file simplification efforts. (Human review goes only so far and it has failed to reach its goal in the past 15 years, as it only addressed the most obvious excesses.)

If we are too broken right now then maybe make it only available to someone explicitly interested in it, via make CONFIG_DEBUG_HEADER_DEPENDENCIES=y or so?

Ingo