[PATCH stable] x86/tsc: Prevent result truncation on 32 bit

[PATCH stable] x86/tsc: Prevent result truncation on 32 bit

[Thomas Gleixner]

Subject: x86/tsc: Prevent result truncation on 32 bit
From: Chuanhua Lei <chuanhua.lei@linux.intel.com>
Date: Thu Sep 6 18:03:23 2018 +0800

From: Chuanhua Lei <chuanhua.lei@linux.intel.com>

Commit 17f6bac2249356c795339e03a0742cd79be3cab8 upstream.

Loops per jiffy is calculated by multiplying tsc_khz with 1e3 and then
dividing it by HZ.
    
Both tsc_khz and the temporary variable holding the multiplication result
are of type unsigned long, so on 32bit the result is truncated to the lower
32bit.
    
Use u64 as type for the temporary variable and cast tsc_khz to it before
multiplying.
    
[ tglx: Backport to stable. Due to massive code changes is the upstream
  	commit not applicable anymore. The issue has gone unnoticed in
  	kernels pre 4.19 because the bogus LPJ value gets fixed up in a
  	later stage of early boot, but it still might cause subtle and hard
  	to debug issues between these two points. ]

Fixes: dd759d93f4dd ("x86/timers: Add simple udelay calibration")
Signed-off-by: Chuanhua Lei <chuanhua.lei@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: yixin.zhu@linux.intel.com
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Len Brown <len.brown@intel.com>
Cc: Pavel Tatashin <pavel.tatashin@microsoft.com>
Cc: Rajvi Jingar <rajvi.jingar@intel.com>
Cc: Dou Liyang <douly.fnst@cn.fujitsu.com>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Link: https://lkml.kernel.org/r/1536228203-18701-1-git-send-email-chuanhua.lei@linux.intel.com

---
 arch/x86/kernel/tsc.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

--- a/arch/x86/kernel/tsc.c
+++ b/arch/x86/kernel/tsc.c
@@ -1343,7 +1343,7 @@ device_initcall(init_tsc_clocksource);
 
 void __init tsc_early_delay_calibrate(void)
 {
-	unsigned long lpj;
+	u64 lpj;
 
 	if (!boot_cpu_has(X86_FEATURE_TSC))
 		return;
@@ -1355,7 +1355,7 @@ void __init tsc_early_delay_calibrate(vo
 	if (!tsc_khz)
 		return;
 
-	lpj = tsc_khz * 1000;
+	lpj = (u64)tsc_khz * 1000;
 	do_div(lpj, HZ);
 	loops_per_jiffy = lpj;
 }

Re: [PATCH stable] x86/tsc: Prevent result truncation on 32 bit

[Greg KH]

On Wed, Sep 12, 2018 at 02:42:08PM +0200, Thomas Gleixner wrote:
> Subject: x86/tsc: Prevent result truncation on 32 bit
> From: Chuanhua Lei <chuanhua.lei@linux.intel.com>
> Date: Thu Sep 6 18:03:23 2018 +0800
> 
> From: Chuanhua Lei <chuanhua.lei@linux.intel.com>
> 
> Commit 17f6bac2249356c795339e03a0742cd79be3cab8 upstream.
> 
> Loops per jiffy is calculated by multiplying tsc_khz with 1e3 and then
> dividing it by HZ.
>     
> Both tsc_khz and the temporary variable holding the multiplication result
> are of type unsigned long, so on 32bit the result is truncated to the lower
> 32bit.
>     
> Use u64 as type for the temporary variable and cast tsc_khz to it before
> multiplying.
>     
> [ tglx: Backport to stable. Due to massive code changes is the upstream
>   	commit not applicable anymore. The issue has gone unnoticed in
>   	kernels pre 4.19 because the bogus LPJ value gets fixed up in a
>   	later stage of early boot, but it still might cause subtle and hard
>   	to debug issues between these two points. ]

Thanks for the backport, now queued up.

greg k-h