[PATCH] unexport get_wchan

[PATCH] unexport get_wchan

[Christoph Hellwig]

only usedby procfs which certainly can't be modular


--- 1.38/arch/alpha/kernel/alpha_ksyms.c	2004-08-31 01:05:49 +02:00
+++ edited/arch/alpha/kernel/alpha_ksyms.c	2004-09-07 14:44:39 +02:00
@@ -258,8 +258,6 @@
 EXPORT_SYMBOL_NOVERS(memset);
 EXPORT_SYMBOL_NOVERS(memchr);
 
-EXPORT_SYMBOL(get_wchan);
-
 #ifdef CONFIG_ALPHA_IRONGATE
 EXPORT_SYMBOL(irongate_ioremap);
 EXPORT_SYMBOL(irongate_iounmap);
--- 1.39/arch/arm/kernel/process.c	2004-06-22 05:14:16 +02:00
+++ edited/arch/arm/kernel/process.c	2004-09-07 14:43:04 +02:00
@@ -452,4 +452,3 @@
 	} while (count ++ < 16);
 	return 0;
 }
-EXPORT_SYMBOL(get_wchan);
--- 1.2/arch/arm26/kernel/armksyms.c	2004-02-25 11:31:13 +01:00
+++ edited/arch/arm26/kernel/armksyms.c	2004-09-07 14:44:44 +02:00
@@ -219,8 +219,6 @@
 EXPORT_SYMBOL_NOVERS(__down_trylock_failed);
 EXPORT_SYMBOL_NOVERS(__up_wakeup);
 
-EXPORT_SYMBOL(get_wchan);
-
 #ifdef CONFIG_PREEMPT
 EXPORT_SYMBOL(kernel_flag);
 #endif
--- 1.4/arch/h8300/kernel/h8300_ksyms.c	2004-05-25 11:53:07 +02:00
+++ edited/arch/h8300/kernel/h8300_ksyms.c	2004-09-07 14:45:23 +02:00
@@ -58,8 +58,6 @@
 EXPORT_SYMBOL_NOVERS(memscan);
 EXPORT_SYMBOL_NOVERS(memmove);
 
-EXPORT_SYMBOL(get_wchan);
-
 /*
  * libgcc functions - functions that are used internally by the
  * compiler...  (prototypes are not correct though, but that
--- 1.63/arch/i386/kernel/i386_ksyms.c	2004-08-31 09:55:10 +02:00
+++ edited/arch/i386/kernel/i386_ksyms.c	2004-09-07 14:45:26 +02:00
@@ -167,8 +167,6 @@
 EXPORT_SYMBOL(screen_info);
 #endif
 
-EXPORT_SYMBOL(get_wchan);
-
 EXPORT_SYMBOL(rtc_lock);
 
 EXPORT_SYMBOL_GPL(set_nmi_callback);
--- 1.11/arch/m68k/kernel/m68k_ksyms.c	2003-04-01 00:29:49 +02:00
+++ edited/arch/m68k/kernel/m68k_ksyms.c	2004-09-07 14:45:33 +02:00
@@ -85,5 +85,3 @@
 EXPORT_SYMBOL_NOVERS(__down_failed_interruptible);
 EXPORT_SYMBOL_NOVERS(__down_failed_trylock);
 EXPORT_SYMBOL_NOVERS(__up_wakeup);
-
-EXPORT_SYMBOL(get_wchan);
--- 1.4/arch/m68knommu/kernel/m68k_ksyms.c	2004-05-10 15:51:25 +02:00
+++ edited/arch/m68knommu/kernel/m68k_ksyms.c	2004-09-07 14:45:37 +02:00
@@ -61,8 +61,6 @@
 EXPORT_SYMBOL_NOVERS(__down_failed_trylock);
 EXPORT_SYMBOL_NOVERS(__up_wakeup);
 
-EXPORT_SYMBOL(get_wchan);
-
 /*
  * libgcc functions - functions that are used internally by the
  * compiler...  (prototypes are not correct though, but that
===== arch/mips/kernel/process.c 1.14 vs edited =====
--- 1.14/arch/mips/kernel/process.c	2004-05-10 13:25:34 +02:00
+++ edited/arch/mips/kernel/process.c	2004-09-07 14:43:23 +02:00
@@ -340,5 +340,3 @@
 
 	return pc;
 }
-
-EXPORT_SYMBOL(get_wchan);
--- 1.60/arch/ppc/kernel/ppc_ksyms.c	2004-06-25 14:52:53 +02:00
+++ edited/arch/ppc/kernel/ppc_ksyms.c	2004-09-07 14:45:43 +02:00
@@ -295,7 +295,6 @@
 void ppc_irq_dispatch_handler(struct pt_regs *, int);
 EXPORT_SYMBOL(ppc_irq_dispatch_handler);
 EXPORT_SYMBOL(tb_ticks_per_jiffy);
-EXPORT_SYMBOL(get_wchan);
 EXPORT_SYMBOL(console_drivers);
 #ifdef CONFIG_XMON
 EXPORT_SYMBOL(xmon);
--- 1.42/arch/ppc64/kernel/ppc_ksyms.c	2004-06-24 10:55:56 +02:00
+++ edited/arch/ppc64/kernel/ppc_ksyms.c	2004-09-07 14:45:46 +02:00
@@ -157,7 +157,6 @@
 
 EXPORT_SYMBOL(timer_interrupt);
 EXPORT_SYMBOL(irq_desc);
-EXPORT_SYMBOL(get_wchan);
 EXPORT_SYMBOL(console_drivers);
 
 EXPORT_SYMBOL(tb_ticks_per_usec);

Re: [PATCH] unexport get_wchan

[Alan Cox]

On Maw, 2004-09-07 at 15:45, Christoph Hellwig wrote:
> only usedby procfs which certainly can't be modular

And multiple out of tree debuggers. Christoph - do some basic homework.
Also save the "so merge it" crap for someone else before you come up
with that comment. Given the choice between a few exports of relevant
functionality and merging some of the open source stuff that uses it we
are far far better off with the export.

Plus Linus has a random personal and unreasonable hatred of debugging
tools so they'll never get merged anyway without forking the kernel.

Alan

Re: [PATCH] unexport get_wchan

[Alan Cox]

On Maw, 2004-09-07 at 19:11, Christoph Hellwig wrote:
> Which debuging tool?  Both kdb and xmon don't use it.

You broke my kgdb 8) 

Re: [PATCH] unexport get_wchan

[Christoph Hellwig]

On Tue, Sep 07, 2004 at 06:07:49PM +0100, Alan Cox wrote:
> On Maw, 2004-09-07 at 15:45, Christoph Hellwig wrote:
> > only usedby procfs which certainly can't be modular
> 
> And multiple out of tree debuggers. Christoph - do some basic homework.
> Also save the "so merge it" crap for someone else before you come up
> with that comment. Given the choice between a few exports of relevant
> functionality and merging some of the open source stuff that uses it we
> are far far better off with the export.
> 
> Plus Linus has a random personal and unreasonable hatred of debugging
> tools so they'll never get merged anyway without forking the kernel.

Which debuging tool?  Both kdb and xmon don't use it.

Re: [PATCH] unexport get_wchan

[Christoph Hellwig]

On Tue, Sep 07, 2004 at 06:29:18PM +0100, Alan Cox wrote:
> On Maw, 2004-09-07 at 19:11, Christoph Hellwig wrote:
> > Which debuging tool?  Both kdb and xmon don't use it.
> 
> You broke my kgdb 8) 

Well, you kdb cotains quite a few other patches to core code already,
no?  And it could add back a single EXPORT_SYMBOL in a common place
instead of duplicated over every architecture..

Re: [PATCH] unexport get_wchan

[Arjan van de Ven]

[-- Attachment #1: Type: text/plain, Size: 264 bytes --]

On Tue, 2004-09-07 at 19:29, Alan Cox wrote:
> On Maw, 2004-09-07 at 19:11, Christoph Hellwig wrote:
> > Which debuging tool?  Both kdb and xmon don't use it.
> 
> You broke my kgdb 8) 

kgdb surely uses the kallsyms stuff instead... far more reliable...


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [PATCH] unexport get_wchan

[Andreas Dilger]

[-- Attachment #1: Type: text/plain, Size: 816 bytes --]

On Sep 07, 2004  20:49 +0200, Christoph Hellwig wrote:
> Well, you kdb cotains quite a few other patches to core code already,
> no?  And it could add back a single EXPORT_SYMBOL in a common place
> instead of duplicated over every architecture..

I think most projects' goals is to reduce the number of patches they need
to make to the kernel instead of increasing them...  If you think it is
ugly to have multiple EXPORT_SYMBOLS() then it should just be moved to
the line after the function definition.  Having to add another configure
rule to check for now-unexported symbols isn't my idea of time well spent
(not that I care about this one in particular).

Cheers, Andreas
--
Andreas Dilger
http://sourceforge.net/projects/ext2resize/
http://members.shaw.ca/adilger/             http://members.shaw.ca/golinux/


[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]