Patchwork xhci: Use ffs() to find page size in xhci_mem_init()

login
register
mail settings
Submitter Andrey Smirnov
Date Feb. 7, 2019, 12:03 a.m.
Message ID <20190207000349.7816-1-andrew.smirnov@gmail.com>
Download mbox | patch
Permalink /patch/719955/
State New
Headers show

Comments

Andrey Smirnov - Feb. 7, 2019, 12:03 a.m.
Get page size order using ffs() instead of open coding it with a loop.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Cc: Mathias Nyman <mathias.nyman@intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: linux-usb@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
---
 drivers/usb/host/xhci-mem.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)
Mathias Nyman - Feb. 7, 2019, 9:04 a.m.
On 07.02.2019 02:03, Andrey Smirnov wrote:
> Get page size order using ffs() instead of open coding it with a loop.
> 
> Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
> Cc: Mathias Nyman <mathias.nyman@intel.com>
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: linux-usb@vger.kernel.org
> Cc: linux-kernel@vger.kernel.org
> ---
>   drivers/usb/host/xhci-mem.c | 6 +-----
>   1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
> index 36a3eb8849f1..44b43c3d819f 100644
> --- a/drivers/usb/host/xhci-mem.c
> +++ b/drivers/usb/host/xhci-mem.c
> @@ -2362,11 +2362,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
>   	page_size = readl(&xhci->op_regs->page_size);
>   	xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>   			"Supported page size register = 0x%x", page_size);
> -	for (i = 0; i < 16; i++) {
> -		if ((0x1 & page_size) != 0)
> -			break;
> -		page_size = page_size >> 1;
> -	}
> +	i = ffs(page_size);
>   	if (i < 16)
>   		xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>   			"Supported page size of %iK", (1 << (i+12)) / 1024);

Hi

using ffs() is a welcome change, but it will give different a result than the loop.

*old loop
   valid page_size value if i < 16
*ffs()
   valid page_size value if i >= 1 and i < 17
  
-Mathias
Felipe Balbi - Feb. 7, 2019, 9:06 a.m.
Hi,

Mathias Nyman <mathias.nyman@linux.intel.com> writes:
> On 07.02.2019 02:03, Andrey Smirnov wrote:
>> Get page size order using ffs() instead of open coding it with a loop.
>> 
>> Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
>> Cc: Mathias Nyman <mathias.nyman@intel.com>
>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>> Cc: linux-usb@vger.kernel.org
>> Cc: linux-kernel@vger.kernel.org
>> ---
>>   drivers/usb/host/xhci-mem.c | 6 +-----
>>   1 file changed, 1 insertion(+), 5 deletions(-)
>> 
>> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
>> index 36a3eb8849f1..44b43c3d819f 100644
>> --- a/drivers/usb/host/xhci-mem.c
>> +++ b/drivers/usb/host/xhci-mem.c
>> @@ -2362,11 +2362,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
>>   	page_size = readl(&xhci->op_regs->page_size);
>>   	xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>   			"Supported page size register = 0x%x", page_size);
>> -	for (i = 0; i < 16; i++) {
>> -		if ((0x1 & page_size) != 0)
>> -			break;
>> -		page_size = page_size >> 1;
>> -	}
>> +	i = ffs(page_size);
>>   	if (i < 16)
>>   		xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>   			"Supported page size of %iK", (1 << (i+12)) / 1024);
>
> Hi
>
> using ffs() is a welcome change, but it will give different a result than the loop.
>
> *old loop
>    valid page_size value if i < 16
> *ffs()
>    valid page_size value if i >= 1 and i < 17

off-by-one, just use i = ffs() - 1. Or use __ffs().
Mathias Nyman - Feb. 7, 2019, 10:46 a.m.
On 07.02.2019 11:06, Felipe Balbi wrote:
> 
> Hi,
> 
> Mathias Nyman <mathias.nyman@linux.intel.com> writes:
>> On 07.02.2019 02:03, Andrey Smirnov wrote:
>>> Get page size order using ffs() instead of open coding it with a loop.
>>>
>>> Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
>>> Cc: Mathias Nyman <mathias.nyman@intel.com>
>>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>> Cc: linux-usb@vger.kernel.org
>>> Cc: linux-kernel@vger.kernel.org
>>> ---
>>>    drivers/usb/host/xhci-mem.c | 6 +-----
>>>    1 file changed, 1 insertion(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
>>> index 36a3eb8849f1..44b43c3d819f 100644
>>> --- a/drivers/usb/host/xhci-mem.c
>>> +++ b/drivers/usb/host/xhci-mem.c
>>> @@ -2362,11 +2362,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
>>>    	page_size = readl(&xhci->op_regs->page_size);
>>>    	xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>>    			"Supported page size register = 0x%x", page_size);
>>> -	for (i = 0; i < 16; i++) {
>>> -		if ((0x1 & page_size) != 0)
>>> -			break;
>>> -		page_size = page_size >> 1;
>>> -	}
>>> +	i = ffs(page_size);
>>>    	if (i < 16)
>>>    		xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>>    			"Supported page size of %iK", (1 << (i+12)) / 1024);
>>
>> Hi
>>
>> using ffs() is a welcome change, but it will give different a result than the loop.
>>
>> *old loop
>>     valid page_size value if i < 16
>> *ffs()
>>     valid page_size value if i >= 1 and i < 17
> 
> off-by-one, just use i = ffs() - 1. Or use __ffs().
> 

and handle the page_size == 0 case.

-Mathias
Felipe Balbi - Feb. 7, 2019, 10:58 a.m.
Hi,

Mathias Nyman <mathias.nyman@intel.com> writes:
>>>> Get page size order using ffs() instead of open coding it with a loop.
>>>>
>>>> Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
>>>> Cc: Mathias Nyman <mathias.nyman@intel.com>
>>>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>>> Cc: linux-usb@vger.kernel.org
>>>> Cc: linux-kernel@vger.kernel.org
>>>> ---
>>>>    drivers/usb/host/xhci-mem.c | 6 +-----
>>>>    1 file changed, 1 insertion(+), 5 deletions(-)
>>>>
>>>> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
>>>> index 36a3eb8849f1..44b43c3d819f 100644
>>>> --- a/drivers/usb/host/xhci-mem.c
>>>> +++ b/drivers/usb/host/xhci-mem.c
>>>> @@ -2362,11 +2362,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
>>>>    	page_size = readl(&xhci->op_regs->page_size);
>>>>    	xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>>>    			"Supported page size register = 0x%x", page_size);
>>>> -	for (i = 0; i < 16; i++) {
>>>> -		if ((0x1 & page_size) != 0)
>>>> -			break;
>>>> -		page_size = page_size >> 1;
>>>> -	}
>>>> +	i = ffs(page_size);
>>>>    	if (i < 16)
>>>>    		xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>>>    			"Supported page size of %iK", (1 << (i+12)) / 1024);
>>>
>>> Hi
>>>
>>> using ffs() is a welcome change, but it will give different a result than the loop.
>>>
>>> *old loop
>>>     valid page_size value if i < 16
>>> *ffs()
>>>     valid page_size value if i >= 1 and i < 17
>> 
>> off-by-one, just use i = ffs() - 1. Or use __ffs().
>
> and handle the page_size == 0 case.

Can it be zero in real life, or are you protecting against academic
possibility that's never going to happen in HW?
Mathias Nyman - Feb. 7, 2019, 11:54 a.m.
On 07.02.2019 12:58, Felipe Balbi wrote:
> 
> Hi,
> 
> Mathias Nyman <mathias.nyman@intel.com> writes:
>>>>> Get page size order using ffs() instead of open coding it with a loop.
>>>>>
>>>>> Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
>>>>> Cc: Mathias Nyman <mathias.nyman@intel.com>
>>>>> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
>>>>> Cc: linux-usb@vger.kernel.org
>>>>> Cc: linux-kernel@vger.kernel.org
>>>>> ---
>>>>>     drivers/usb/host/xhci-mem.c | 6 +-----
>>>>>     1 file changed, 1 insertion(+), 5 deletions(-)
>>>>>
>>>>> diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
>>>>> index 36a3eb8849f1..44b43c3d819f 100644
>>>>> --- a/drivers/usb/host/xhci-mem.c
>>>>> +++ b/drivers/usb/host/xhci-mem.c
>>>>> @@ -2362,11 +2362,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
>>>>>     	page_size = readl(&xhci->op_regs->page_size);
>>>>>     	xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>>>>     			"Supported page size register = 0x%x", page_size);
>>>>> -	for (i = 0; i < 16; i++) {
>>>>> -		if ((0x1 & page_size) != 0)
>>>>> -			break;
>>>>> -		page_size = page_size >> 1;
>>>>> -	}
>>>>> +	i = ffs(page_size);
>>>>>     	if (i < 16)
>>>>>     		xhci_dbg_trace(xhci, trace_xhci_dbg_init,
>>>>>     			"Supported page size of %iK", (1 << (i+12)) / 1024);
>>>>
>>>> Hi
>>>>
>>>> using ffs() is a welcome change, but it will give different a result than the loop.
>>>>
>>>> *old loop
>>>>      valid page_size value if i < 16
>>>> *ffs()
>>>>      valid page_size value if i >= 1 and i < 17
>>>
>>> off-by-one, just use i = ffs() - 1. Or use __ffs().
>>
>> and handle the page_size == 0 case.
> 
> Can it be zero in real life, or are you protecting against academic
> possibility that's never going to happen in HW?
> 

whole page_size check is not really doing much, just printing out
different debug messages.

-Mathias
Andrey Smirnov - Feb. 7, 2019, 9:06 p.m.
On Thu, Feb 7, 2019 at 1:00 AM Mathias Nyman
<mathias.nyman@linux.intel.com> wrote:
>
> On 07.02.2019 02:03, Andrey Smirnov wrote:
> > Get page size order using ffs() instead of open coding it with a loop.
> >
> > Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
> > Cc: Mathias Nyman <mathias.nyman@intel.com>
> > Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> > Cc: linux-usb@vger.kernel.org
> > Cc: linux-kernel@vger.kernel.org
> > ---
> >   drivers/usb/host/xhci-mem.c | 6 +-----
> >   1 file changed, 1 insertion(+), 5 deletions(-)
> >
> > diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
> > index 36a3eb8849f1..44b43c3d819f 100644
> > --- a/drivers/usb/host/xhci-mem.c
> > +++ b/drivers/usb/host/xhci-mem.c
> > @@ -2362,11 +2362,7 @@ int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
> >       page_size = readl(&xhci->op_regs->page_size);
> >       xhci_dbg_trace(xhci, trace_xhci_dbg_init,
> >                       "Supported page size register = 0x%x", page_size);
> > -     for (i = 0; i < 16; i++) {
> > -             if ((0x1 & page_size) != 0)
> > -                     break;
> > -             page_size = page_size >> 1;
> > -     }
> > +     i = ffs(page_size);
> >       if (i < 16)
> >               xhci_dbg_trace(xhci, trace_xhci_dbg_init,
> >                       "Supported page size of %iK", (1 << (i+12)) / 1024);
>
> Hi
>
> using ffs() is a welcome change, but it will give different a result than the loop.
>
> *old loop
>    valid page_size value if i < 16
> *ffs()
>    valid page_size value if i >= 1 and i < 17
>

My bad, sorry about that, was using __fls() in the version I tested,
but then switched it to ffs() at the last minute and forgot to double
check. Will update in v2.

Thanks,
Andrey Smirnov

Patch

diff --git a/drivers/usb/host/xhci-mem.c b/drivers/usb/host/xhci-mem.c
index 36a3eb8849f1..44b43c3d819f 100644
--- a/drivers/usb/host/xhci-mem.c
+++ b/drivers/usb/host/xhci-mem.c
@@ -2362,11 +2362,7 @@  int xhci_mem_init(struct xhci_hcd *xhci, gfp_t flags)
 	page_size = readl(&xhci->op_regs->page_size);
 	xhci_dbg_trace(xhci, trace_xhci_dbg_init,
 			"Supported page size register = 0x%x", page_size);
-	for (i = 0; i < 16; i++) {
-		if ((0x1 & page_size) != 0)
-			break;
-		page_size = page_size >> 1;
-	}
+	i = ffs(page_size);
 	if (i < 16)
 		xhci_dbg_trace(xhci, trace_xhci_dbg_init,
 			"Supported page size of %iK", (1 << (i+12)) / 1024);