Skip to contentSkip to conversation

s390: Remove obsolete bust_spinlock() implementation

PATCH In review Petr Mladek 5 participants 5 messages Nov 22, 2018 – Nov 27, 2018

Subsystems (2)
s390 s390 MM
lore
Share & tools
lore

Petr Mladek

author
#0 lore raw

PATCH 1/1 s390: Remove obsolete bust_spinlock() implementation

The commit cefc8be82403cf ("Consolidate bust_spinlocks()") kept the s390-specific implementation because of the absence of CONFIG_VT. In fact, the only difference was calling console_unblank() instead of unblank_screen().

The common implementation in lib/bust_spinlocks.c started to call unblank_screen() explicitly since the commit b61312d353da187 ("oops handling: ensure that any oops is flushed to the mtdoops console").

As a result, the custom implementation is not longer necessary. And we could get all the other improvements of the common implementation for free.

Signed-off-by Petr Mladek <[email protected]>
---
 arch/s390/mm/fault.c | 24 ------------------------
 1 file changed, 24 deletions(-)
arch/s390/mm/fault.c +0 -24
index 2b8f32f56e0c..11613362c4e7 100644
@@ -81,30 +81,6 @@ static inline int notify_page_fault(struct pt_regs *regs)
 	return ret;
 }
 
-
-/*
- * Unlock any spinlocks which will prevent us from getting the
- * message out.
- */
-void bust_spinlocks(int yes)
-{
-	if (yes) {
-		oops_in_progress = 1;
-	} else {
-		int loglevel_save = console_loglevel;
-		console_unblank();
-		oops_in_progress = 0;
-		/*
-		 * OK, the message is on the console.  Now we call printk()
-		 * without oops_in_progress set so that printk will give klogd
-		 * a poke.  Hold onto your hats...
-		 */
-		console_loglevel = 15;
-		printk(" ");
-		console_loglevel = loglevel_save;
-	}
-}
-
 /*
  * Find out which address space caused the exception.
  * Access register mode is impossible, ignore space == 3.
Signature · 1 line
2.13.7
#1 lore raw

Please remove the weak attribute from the generic implementation as well now that the last override is gone.

#2 lore raw
In reply to Petr Mladek
Quoted 5 lines · Petr Mladek

The commit cefc8be82403cf ("Consolidate bust_spinlocks()") kept the s390-specific implementation because of the absence of CONFIG_VT. In fact, the only difference was calling console_unblank() instead of unblank_screen().

Quoted 8 lines

The common implementation in lib/bust_spinlocks.c started to call unblank_screen() explicitly since the commit b61312d353da187 ("oops handling: ensure that any oops is flushed to the mtdoops console").

As a result, the custom implementation is not longer necessary. And we could get all the other improvements of the common implementation for free.

I believe I sent a similar patch several weeks ago and it's in s390 patch queue as of now, waiting for the next merge window.

lkml.kernel.org/r/20181025081108.GB26561@osiris

	-ss
#3 lore raw
Quoted 15 lines · Sergey Senozhatsky
Quoted 5 lines · Petr Mladek

The commit cefc8be82403cf ("Consolidate bust_spinlocks()") kept the s390-specific implementation because of the absence of CONFIG_VT. In fact, the only difference was calling console_unblank() instead of unblank_screen().

Quoted 8 lines

The common implementation in lib/bust_spinlocks.c started to call unblank_screen() explicitly since the commit b61312d353da187 ("oops handling: ensure that any oops is flushed to the mtdoops console").

As a result, the custom implementation is not longer necessary. And we could get all the other improvements of the common implementation for free.

Quoted 5 lines

I believe I sent a similar patch several weeks ago and it's in s390 patch queue as of now, waiting for the next merge window.

lkml.kernel.org/r/20181025081108.GB26561@osiris

Yes, it will be added soon to the features branch of the s390/linux.git repository on kernel.org and then hit linux-next.

#4 lore raw
Quoted 16 lines · On Fri, 23 Nov 2018 08:02:12 +0100 Heiko Carstens <[email protected]>…
Quoted 15 lines · Sergey Senozhatsky
Quoted 13 lines · Petr Mladek

The commit cefc8be82403cf ("Consolidate bust_spinlocks()") kept the s390-specific implementation because of the absence of CONFIG_VT. In fact, the only difference was calling console_unblank() instead of unblank_screen().

The common implementation in lib/bust_spinlocks.c started to call unblank_screen() explicitly since the commit b61312d353da187 ("oops handling: ensure that any oops is flushed to the mtdoops console").

As a result, the custom implementation is not longer necessary. And we could get all the other improvements of the common implementation for free.

Quoted 8 lines
Quoted 5 lines

I believe I sent a similar patch several weeks ago and it's in s390 patch queue as of now, waiting for the next merge window.

lkml.kernel.org/r/20181025081108.GB26561@osiris

Yes, it will be added soon to the features branch of the s390/linux.git repository on kernel.org and then hit linux-next.

The patch is now queued for the next merge window.

Signature · 4 lines
blue skies,
   Martin.

"Reality continues to ruin my life." - Calvin.