[PATCH] Fix compilation on gcc 3.4

All of lore.kernel.org
 help / color / mirror / Atom feed

* [PATCH] Fix compilation on gcc 3.4
@ 2004-01-14  8:37 Andi Kleen
  2004-01-14  8:47 ` Jakub Jelinek
  2004-01-14  8:49 ` Andrew Morton
  0 siblings, 2 replies; 6+ messages in thread
From: Andi Kleen @ 2004-01-14  8:37 UTC (permalink / raw)
  To: akpm; +Cc: jh, linux-kernel


The upcomming gcc 3.4 has a new inlining algorithm which sometimes
fails to inline some "dumb" inlines in the kernel. This sometimes leads
to undefined symbols while linking.

To make the kernel compile again this patch drops the always inline
for gcc 3.4.  The new algorithm should be good enough to do the right
thing on its own. 

-Andi

diff -u linux-34/include/linux/compiler-gcc3.h-o linux-34/include/linux/compiler-gcc3.h
--- linux-34/include/linux/compiler-gcc3.h-o	2003-09-28 10:53:23.000000000 +0200
+++ linux-34/include/linux/compiler-gcc3.h	2004-01-13 22:36:22.000000000 +0100
@@ -3,7 +3,9 @@
 /* These definitions are for GCC v3.x.  */
 #include <linux/compiler-gcc.h>
 
-#if __GNUC_MINOR__ >= 1
+/* gcc 3.4 has a new inlining algorithm and always_inline seems to 
+   do more harm than good now. */
+#if __GNUC_MINOR__ >= 1 && __GNUC_MINOR__ < 4
 # define inline		__inline__ __attribute__((always_inline))
 # define __inline__	__inline__ __attribute__((always_inline))
 # define __inline	__inline__ __attribute__((always_inline))




^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix compilation on gcc 3.4
  2004-01-14  8:37 [PATCH] Fix compilation on gcc 3.4 Andi Kleen
@ 2004-01-14  8:47 ` Jakub Jelinek
  2004-01-14  8:58   ` Andi Kleen
  2004-01-14  8:49 ` Andrew Morton
  1 sibling, 1 reply; 6+ messages in thread
From: Jakub Jelinek @ 2004-01-14  8:47 UTC (permalink / raw)
  To: Andi Kleen; +Cc: akpm, jh, linux-kernel

On Wed, Jan 14, 2004 at 09:37:00AM +0100, Andi Kleen wrote:
> 
> The upcomming gcc 3.4 has a new inlining algorithm which sometimes
> fails to inline some "dumb" inlines in the kernel. This sometimes leads
> to undefined symbols while linking.

It fails to inline routines with always_inline attribute?
That sounds like a gcc bug.  always_inline should mean inline always,
and issue error if for some reason it is impossible.

	Jakub

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix compilation on gcc 3.4
  2004-01-14  8:47 ` Jakub Jelinek
@ 2004-01-14  8:58   ` Andi Kleen
  2004-01-14 10:53     ` Jan Hubicka
  0 siblings, 1 reply; 6+ messages in thread
From: Andi Kleen @ 2004-01-14  8:58 UTC (permalink / raw)
  To: Jakub Jelinek; +Cc: Andi Kleen, akpm, jh, linux-kernel

On Wed, Jan 14, 2004 at 03:47:22AM -0500, Jakub Jelinek wrote:
> On Wed, Jan 14, 2004 at 09:37:00AM +0100, Andi Kleen wrote:
> > 
> > The upcomming gcc 3.4 has a new inlining algorithm which sometimes
> > fails to inline some "dumb" inlines in the kernel. This sometimes leads
> > to undefined symbols while linking.
> 
> It fails to inline routines with always_inline attribute?
> That sounds like a gcc bug.  always_inline should mean inline always,
> and issue error if for some reason it is impossible.

The problem is that there are some functions that are declared
inline in header files, but there is no body available. When they are
called this ends with an hard error in gcc. I started with fixing
them but eventually gave up because there were so many of them.

In addition you get tons of ugly warnings.

-Andi

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix compilation on gcc 3.4
  2004-01-14  8:58   ` Andi Kleen
@ 2004-01-14 10:53     ` Jan Hubicka
  0 siblings, 0 replies; 6+ messages in thread
From: Jan Hubicka @ 2004-01-14 10:53 UTC (permalink / raw)
  To: Andi Kleen; +Cc: Jakub Jelinek, Andi Kleen, akpm, jh, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 1974 bytes --]

> On Wed, Jan 14, 2004 at 03:47:22AM -0500, Jakub Jelinek wrote:
> > On Wed, Jan 14, 2004 at 09:37:00AM +0100, Andi Kleen wrote:
> > > 
> > > The upcomming gcc 3.4 has a new inlining algorithm which sometimes
> > > fails to inline some "dumb" inlines in the kernel. This sometimes leads
> > > to undefined symbols while linking.
> > 
> > It fails to inline routines with always_inline attribute?
> > That sounds like a gcc bug.  always_inline should mean inline always,
> > and issue error if for some reason it is impossible.
> 
> The problem is that there are some functions that are declared
> inline in header files, but there is no body available. When they are
> called this ends with an hard error in gcc. I started with fixing
> them but eventually gave up because there were so many of them.
> 
> In addition you get tons of ugly warnings.

Yes, this is not really bug, but the simple fact that if compler never
see the function body, it can not inline it.  Old compilers accepted
this by mystake, as always_inline is supposed to cause hard error each
time function failes to be inlined.

I made testing with equivalent of Andi's patch and with my last fixes to
GCC code size estimate, there are 50 calls in the whole kernel where GCC
concludes that inlining is not good idea.  I plan to investigate these
too, but there is simple --param knob you may want to use if you really
think these are sane.

For a record, I made patch that fixes all the cases for my default
configuration.  You may consider it applying together with Andi's patch,
so -Winline output does not contain unnecesary complains.

I am also attaching the list of warnings I get during build.  The
parameter inlie-unit-growth is specifying the maximal growth (in
percent) number of instruction can increase within single source file by
inlining.  As Kernel does a lot of inlining and has small sources,
perhaps adjusting up from 50% makes sense.  Perhaps I will do this
default in GCC too.

Honza

[-- Attachment #2: kernelpa --]
[-- Type: text/plain, Size: 11774 bytes --]

Hi,
GCC 3.4 is now outputting error message when function marked always_inline
can't be inlined for wathever reason.  This include the case of function
body not being present in current compilation unit at all.  With current
always_inline hack in kernel, it is necessary to provide body to function
marked as inlined (otherwise the mark is ignored).
The patch removes such inline keywords that are ignored by older compilers
anyway.

Honza

diff -urp ./fs/ntfs.old/ntfs.h ./fs/ntfs/ntfs.h
--- ./fs/ntfs.old/ntfs.h	2004-01-13 00:57:33.000000000 +0100
+++ ./fs/ntfs/ntfs.h	2004-01-13 00:58:04.000000000 +0100
@@ -183,7 +183,7 @@ extern void post_write_mst_fixup(NTFS_RE
 /* From fs/ntfs/time.c */
 extern inline s64 utc2ntfs(const time_t time);
 extern inline s64 get_current_ntfs_time(void);
-extern inline time_t ntfs2utc(const s64 time);
+extern time_t ntfs2utc(const s64 time);
 
 /* From fs/ntfs/unistr.c */
 extern BOOL ntfs_are_names_equal(const uchar_t *s1, size_t s1_len,
diff -urp ./arch/i386.old/kernel/setup.c ./arch/i386/kernel/setup.c
--- ./arch/i386.old/kernel/setup.c	2004-01-13 15:58:40.000000000 +0100
+++ ./arch/i386/kernel/setup.c	2004-01-13 15:59:29.000000000 +0100
@@ -52,7 +52,7 @@
 
 int disable_pse __initdata = 0;
 
-static inline char * __init machine_specific_memory_setup(void);
+static char * __init machine_specific_memory_setup(void);
 
 /*
  * Machine setup..
diff -urp ./kernel.old/fork.c ./kernel/fork.c
--- ./kernel.old/fork.c	2004-01-13 00:51:45.000000000 +0100
+++ ./kernel/fork.c	2004-01-13 00:51:54.000000000 +0100
@@ -417,7 +417,7 @@ struct mm_struct * mm_alloc(void)
  * is dropped: either by a lazy thread or by
  * mmput. Free the page directory and the mm.
  */
-inline void __mmdrop(struct mm_struct *mm)
+void __mmdrop(struct mm_struct *mm)
 {
 	BUG_ON(mm == &init_mm);
 	mm_free_pgd(mm);
diff -urp ./include.old/asm/apic.h ./include/asm/apic.h
--- ./include.old/asm/apic.h	2004-01-13 00:38:37.000000000 +0100
+++ ./include/asm/apic.h	2004-01-13 00:39:10.000000000 +0100
@@ -85,7 +85,7 @@ extern void disable_lapic_nmi_watchdog(v
 extern void enable_lapic_nmi_watchdog(void);
 extern void disable_timer_nmi_watchdog(void);
 extern void enable_timer_nmi_watchdog(void);
-extern inline void nmi_watchdog_tick (struct pt_regs * regs);
+extern void nmi_watchdog_tick (struct pt_regs * regs);
 extern int APIC_init_uniprocessor (void);
 extern void disable_APIC_timer(void);
 extern void enable_APIC_timer(void);
Only in ./include/asm-i386: .string.h.swp
diff -urp ./include.old/asm-i386/apic.h ./include/asm-i386/apic.h
--- ./include.old/asm-i386/apic.h	2004-01-13 00:38:37.000000000 +0100
+++ ./include/asm-i386/apic.h	2004-01-13 00:39:10.000000000 +0100
@@ -85,7 +85,7 @@ extern void disable_lapic_nmi_watchdog(v
 extern void enable_lapic_nmi_watchdog(void);
 extern void disable_timer_nmi_watchdog(void);
 extern void enable_timer_nmi_watchdog(void);
-extern inline void nmi_watchdog_tick (struct pt_regs * regs);
+extern void nmi_watchdog_tick (struct pt_regs * regs);
 extern int APIC_init_uniprocessor (void);
 extern void disable_APIC_timer(void);
 extern void enable_APIC_timer(void);
Only in ./include/linux: .reiserfs_fs.h.swp
diff -urp ./include.old/linux/bio.h ./include/linux/bio.h
--- ./include.old/linux/bio.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/linux/bio.h	2004-01-13 02:35:21.000000000 +0100
@@ -231,13 +231,13 @@ extern void bio_put(struct bio *);
 
 extern void bio_endio(struct bio *, unsigned int, int);
 struct request_queue;
-extern inline int bio_phys_segments(struct request_queue *, struct bio *);
-extern inline int bio_hw_segments(struct request_queue *, struct bio *);
+extern int bio_phys_segments(struct request_queue *, struct bio *);
+extern int bio_hw_segments(struct request_queue *, struct bio *);
 
-extern inline void __bio_clone(struct bio *, struct bio *);
+extern void __bio_clone(struct bio *, struct bio *);
 extern struct bio *bio_clone(struct bio *, int);
 
-extern inline void bio_init(struct bio *);
+extern void bio_init(struct bio *);
 
 extern int bio_add_page(struct bio *, struct page *, unsigned int,unsigned int);
 extern int bio_get_nr_vecs(struct block_device *);
diff -urp ./include.old/linux/compile.h ./include/linux/compile.h
--- ./include.old/linux/compile.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/linux/compile.h	2004-01-13 19:04:06.000000000 +0100
@@ -1,7 +1,7 @@
-/* This file is auto generated, version 0 */
+/* This file is auto generated, version 4 */
 #define UTS_MACHINE "i386"
-#define UTS_VERSION "#0 Tue Jan 13 00:36:03 CET 2004"
-#define LINUX_COMPILE_TIME "00:36:03"
+#define UTS_VERSION "#4 Tue Jan 13 19:04:06 CET 2004"
+#define LINUX_COMPILE_TIME "19:04:06"
 #define LINUX_COMPILE_BY "root"
 #define LINUX_COMPILE_HOST "ibm"
 #define LINUX_COMPILE_DOMAIN "home.cz"
diff -urp ./include.old/linux/efi.h ./include/linux/efi.h
--- ./include.old/linux/efi.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/linux/efi.h	2004-01-13 00:39:55.000000000 +0100
@@ -297,8 +297,8 @@ extern u64 efi_mem_attributes (unsigned 
 extern void efi_initialize_iomem_resources(struct resource *code_resource,
 					struct resource *data_resource);
 extern efi_status_t phys_efi_get_time(efi_time_t *tm, efi_time_cap_t *tc);
-extern unsigned long inline __init efi_get_time(void);
-extern int inline __init efi_set_rtc_mmss(unsigned long nowtime);
+extern unsigned long __init efi_get_time(void);
+extern int __init efi_set_rtc_mmss(unsigned long nowtime);
 extern struct efi_memory_map memmap;
 
 #ifdef CONFIG_EFI
diff -urp ./include.old/linux/elevator.h ./include/linux/elevator.h
--- ./include.old/linux/elevator.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/linux/elevator.h	2004-01-13 02:38:19.000000000 +0100
@@ -96,9 +96,9 @@ extern elevator_t iosched_as;
 
 extern int elevator_init(request_queue_t *, elevator_t *);
 extern void elevator_exit(request_queue_t *);
-extern inline int elv_rq_merge_ok(struct request *, struct bio *);
-extern inline int elv_try_merge(struct request *, struct bio *);
-extern inline int elv_try_last_merge(request_queue_t *, struct bio *);
+extern int elv_rq_merge_ok(struct request *, struct bio *);
+extern int elv_try_merge(struct request *, struct bio *);
+extern int elv_try_last_merge(request_queue_t *, struct bio *);
 
 /*
  * Return values from elevator merger
diff -urp ./include.old/linux/ide.h ./include/linux/ide.h
--- ./include.old/linux/ide.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/linux/ide.h	2004-01-13 03:02:22.000000000 +0100
@@ -1417,12 +1417,12 @@ typedef struct pkt_task_s {
 	void			*special;
 } pkt_task_t;
 
-extern inline u32 ide_read_24(ide_drive_t *);
+extern u32 ide_read_24(ide_drive_t *);
 
-extern inline void SELECT_DRIVE(ide_drive_t *);
-extern inline void SELECT_INTERRUPT(ide_drive_t *);
-extern inline void SELECT_MASK(ide_drive_t *, int);
-extern inline void QUIRK_LIST(ide_drive_t *);
+extern void SELECT_DRIVE(ide_drive_t *);
+extern void SELECT_INTERRUPT(ide_drive_t *);
+extern void SELECT_MASK(ide_drive_t *, int);
+extern void QUIRK_LIST(ide_drive_t *);
 
 extern void ata_input_data(ide_drive_t *, void *, u32);
 extern void ata_output_data(ide_drive_t *, void *, u32);
diff -urp ./include.old/linux/reiserfs_fs.h ./include/linux/reiserfs_fs.h
--- ./include.old/linux/reiserfs_fs.h	2004-01-13 00:38:37.000000000 +0100
+++ ./include/linux/reiserfs_fs.h	2004-01-13 01:07:58.000000000 +0100
@@ -1780,26 +1780,26 @@ int reiserfs_convert_objectid_map_v1(str
 
 /* stree.c */
 int B_IS_IN_TREE(const struct buffer_head *);
-extern inline void copy_short_key (void * to, const void * from);
-extern inline void copy_item_head(struct item_head * p_v_to, 
+extern void copy_short_key (void * to, const void * from);
+extern void copy_item_head(struct item_head * p_v_to, 
 								  const struct item_head * p_v_from);
 
 // first key is in cpu form, second - le
-extern inline int comp_keys (const struct key * le_key, 
+extern int comp_keys (const struct key * le_key, 
+		      const struct cpu_key * cpu_key);
+extern int  comp_short_keys (const struct key * le_key, 
 			     const struct cpu_key * cpu_key);
-extern inline int  comp_short_keys (const struct key * le_key, 
-				    const struct cpu_key * cpu_key);
-extern inline void le_key2cpu_key (struct cpu_key * to, const struct key * from);
+extern void le_key2cpu_key (struct cpu_key * to, const struct key * from);
 
 // both are cpu keys
-extern inline int comp_cpu_keys (const struct cpu_key *, const struct cpu_key *);
-extern inline int comp_short_cpu_keys (const struct cpu_key *, 
+extern int comp_cpu_keys (const struct cpu_key *, const struct cpu_key *);
+extern int comp_short_cpu_keys (const struct cpu_key *, 
 				       const struct cpu_key *);
 extern inline void cpu_key2cpu_key (struct cpu_key *, const struct cpu_key *);
 
 // both are in le form
-extern inline int comp_le_keys (const struct key *, const struct key *);
-extern inline int comp_short_le_keys (const struct key *, const struct key *);
+extern int comp_le_keys (const struct key *, const struct key *);
+extern int comp_short_le_keys (const struct key *, const struct key *);
 
 //
 // get key version from on disk key - kludge
@@ -1834,7 +1834,7 @@ int search_by_key (struct super_block *,
 int search_for_position_by_key (struct super_block * p_s_sb, 
 								const struct cpu_key * p_s_cpu_key, 
 								struct path * p_s_search_path);
-extern inline void decrement_bcount (struct buffer_head * p_s_bh);
+extern void decrement_bcount (struct buffer_head * p_s_bh);
 void decrement_counters_in_path (struct path * p_s_search_path);
 void pathrelse (struct path * p_s_search_path);
 int reiserfs_check_path(struct path *p) ;
@@ -1916,7 +1916,7 @@ void sd_attrs_to_i_attrs( __u16 sd_attrs
 void i_attrs_to_sd_attrs( struct inode *inode, __u16 *sd_attrs );
 
 /* namei.c */
-inline void set_de_name_and_namelen (struct reiserfs_dir_entry * de);
+void set_de_name_and_namelen (struct reiserfs_dir_entry * de);
 int search_by_entry_key (struct super_block * sb, const struct cpu_key * key, 
 			 struct path * path, 
 			 struct reiserfs_dir_entry * de);
@@ -2037,8 +2037,8 @@ int balance_internal (struct tree_balanc
                       struct buffer_head **);
 
 /* do_balance.c */
-inline void do_balance_mark_leaf_dirty (struct tree_balance * tb, 
-					struct buffer_head * bh, int flag);
+void do_balance_mark_leaf_dirty (struct tree_balance * tb, 
+				 struct buffer_head * bh, int flag);
 #define do_balance_mark_internal_dirty do_balance_mark_leaf_dirty
 #define do_balance_mark_sb_dirty do_balance_mark_leaf_dirty
 
diff -urp ./include.old/linux/sched.h ./include/linux/sched.h
--- ./include.old/linux/sched.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/linux/sched.h	2004-01-13 00:50:45.000000000 +0100
@@ -669,7 +669,7 @@ static inline int capable(int cap)
 extern struct mm_struct * mm_alloc(void);
 
 /* mmdrop drops the mm and the page tables */
-extern inline void FASTCALL(__mmdrop(struct mm_struct *));
+extern void FASTCALL(__mmdrop(struct mm_struct *));
 static inline void mmdrop(struct mm_struct * mm)
 {
 	if (atomic_dec_and_test(&mm->mm_count))
diff -urp ./include.old/net/tcp.h ./include/net/tcp.h
--- ./include.old/net/tcp.h	2004-01-13 00:38:36.000000000 +0100
+++ ./include/net/tcp.h	2004-01-13 15:37:24.000000000 +0100
@@ -738,7 +738,7 @@ DECLARE_SNMP_STAT(struct tcp_mib, tcp_st
 #define TCP_ADD_STATS_BH(field, val)	SNMP_ADD_STATS_BH(tcp_statistics, field, val)
 #define TCP_ADD_STATS_USER(field, val)	SNMP_ADD_STATS_USER(tcp_statistics, field, val)
 
-extern inline void		tcp_put_port(struct sock *sk);
+extern void			tcp_put_port(struct sock *sk);
 extern void			tcp_inherit_port(struct sock *sk, struct sock *child);
 
 extern void			tcp_v4_err(struct sk_buff *skb, u32);

[-- Attachment #3: oo --]
[-- Type: text/plain, Size: 6718 bytes --]

fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/lockd/xdr.c:48: warning: inlining failed in call to 'nlm_decode_cookie': --param inline-unit-growth limit reached
fs/nfsd/nfsxdr.c:139: warning: inlining failed in call to 'encode_fattr': --param max-inline-insns-single limit reached
fs/nfsd/nfsxdr.c:139: warning: inlining failed in call to 'encode_fattr': --param max-inline-insns-single limit reached
fs/nfsd/nfsxdr.c:139: warning: inlining failed in call to 'encode_fattr': --param max-inline-insns-single limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/string.h:216: warning: inlining failed in call to '__constant_memcpy': --param inline-unit-growth limit reached
include/asm/string.h:216: warning: inlining failed in call to '__constant_memcpy': --param inline-unit-growth limit reached
include/asm/string.h:216: warning: inlining failed in call to '__constant_memcpy': --param inline-unit-growth limit reached
include/asm/string.h:216: warning: inlining failed in call to '__constant_memcpy': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/asm/uaccess.h:497: warning: inlining failed in call to 'copy_from_user': --param inline-unit-growth limit reached
include/linux/slab.h:78: warning: inlining failed in call to 'kmalloc': --param inline-unit-growth limit reached

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix compilation on gcc 3.4
  2004-01-14  8:37 [PATCH] Fix compilation on gcc 3.4 Andi Kleen
  2004-01-14  8:47 ` Jakub Jelinek
@ 2004-01-14  8:49 ` Andrew Morton
  2004-01-14  9:04   ` Andi Kleen
  1 sibling, 1 reply; 6+ messages in thread
From: Andrew Morton @ 2004-01-14  8:49 UTC (permalink / raw)
  To: Andi Kleen; +Cc: jh, linux-kernel

Andi Kleen <ak@muc.de> wrote:
>
> 
> The upcomming gcc 3.4 has a new inlining algorithm which sometimes
> fails to inline some "dumb" inlines in the kernel. This sometimes leads
> to undefined symbols while linking.

That's a compiler bug, surely.

> To make the kernel compile again this patch drops the always inline
> for gcc 3.4.  The new algorithm should be good enough to do the right
> thing on its own. 

Are you sure?  Without the always_inline stuff we were seeing 100 different
copies of __constant_c_and_count_memset and friends in the vmlinux symbol
table.  It was really silly.

After applying this patch and building with gcc-3.4, how does `size
vmlinux' compare with the same kernel built with gcc-3.3, minus the patch?

^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [PATCH] Fix compilation on gcc 3.4
  2004-01-14  8:49 ` Andrew Morton
@ 2004-01-14  9:04   ` Andi Kleen
  0 siblings, 0 replies; 6+ messages in thread
From: Andi Kleen @ 2004-01-14  9:04 UTC (permalink / raw)
  To: Andrew Morton; +Cc: Andi Kleen, jh, linux-kernel

On Wed, Jan 14, 2004 at 12:49:44AM -0800, Andrew Morton wrote:
> Andi Kleen <ak@muc.de> wrote:
> >
> > 
> > The upcomming gcc 3.4 has a new inlining algorithm which sometimes
> > fails to inline some "dumb" inlines in the kernel. This sometimes leads
> > to undefined symbols while linking.
> 
> That's a compiler bug, surely.

See my mail to Jakub.

> 
> > To make the kernel compile again this patch drops the always inline
> > for gcc 3.4.  The new algorithm should be good enough to do the right
> > thing on its own. 
> 
> Are you sure?  Without the always_inline stuff we were seeing 100 different
> copies of __constant_c_and_count_memset and friends in the vmlinux symbol
> table.  It was really silly.
> 

That was with the old inliner. The new one is completely different
and should do better, especially with -funit-at-a-time. In fact there were
 still a few no inlines for __builtin_constant_p stuff, but Jan fixed them
 now (it was one small left over buglet in the inlining
heuristics). Should be fixed soon in gcc mainline.

I have another patch to compile with -Winline (submitted shortly)
which catches all these issues easily.

> After applying this patch and building with gcc-3.4, how does `size
> vmlinux' compare with the same kernel built with gcc-3.3, minus the patch?

I can only compare to a 3.3-hammer branch compiler, which is a kind
of mixture between 3.3 and 3.4 and already has an older version of the
inliner. 

Better is to use -Winline. You get a warning then
There was only one warning for copy_*_user stuff somewhere in XFS because
it bloated an ioctl function considerably.  There are many other
inlining warnings too, but these seemed to be mostly kernel bugs
(stuff declared inline in headers without bodies etc.)
I think fixing copy_*_user with extern inline would be better so that a 
out of line copy_from_user is used in this case (I can prepare a patch for 
that if you're interested) 

-Andi

^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2004-01-14 10:53 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2004-01-14  8:37 [PATCH] Fix compilation on gcc 3.4 Andi Kleen
2004-01-14  8:47 ` Jakub Jelinek
2004-01-14  8:58   ` Andi Kleen
2004-01-14 10:53     ` Jan Hubicka
2004-01-14  8:49 ` Andrew Morton
2004-01-14  9:04   ` Andi Kleen

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.