[RFC] arm64:use set_fixmap_offset to make it more clear

[RFC] arm64:use set_fixmap_offset to make it more clear

[yalin wang]

A little change to patch_map() function,
use set_fixmap_offset() to make code more clear.

Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
---
 arch/arm64/kernel/insn.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
index dd9671c..7dafd5a 100644
--- a/arch/arm64/kernel/insn.c
+++ b/arch/arm64/kernel/insn.c
@@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
 		return addr;
 
 	BUG_ON(!page);
-	set_fixmap(fixmap, page_to_phys(page));
-
-	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
+	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
+			(addr & ~PAGE_MASK));
 }
 
 static void __kprobes patch_unmap(int fixmap)
-- 
1.9.1

Re: [RFC] arm64:use set_fixmap_offset to make it more clear

[Catalin Marinas]

On Thu, Jul 23, 2015 at 07:45:53PM +0800, yalin wang wrote:
> A little change to patch_map() function,
> use set_fixmap_offset() to make code more clear.
> 
> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
> ---
>  arch/arm64/kernel/insn.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
> index dd9671c..7dafd5a 100644
> --- a/arch/arm64/kernel/insn.c
> +++ b/arch/arm64/kernel/insn.c
> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
>  		return addr;
>  
>  	BUG_ON(!page);
> -	set_fixmap(fixmap, page_to_phys(page));
> -
> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
> +			(addr & ~PAGE_MASK));

It looks fine. Do you get any compiler warning for the automatic pointer
to long conversion? You may want to add some explicit casts, otherwise:

Acked-by: Catalin Marinas <catalin.marinas@arm.com>

Re: [RFC] arm64:use set_fixmap_offset to make it more clear

[Laura Abbott]

On 07/23/2015 04:45 AM, yalin wang wrote:
> A little change to patch_map() function,
> use set_fixmap_offset() to make code more clear.
>
> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
> ---
>   arch/arm64/kernel/insn.c | 5 ++---
>   1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
> index dd9671c..7dafd5a 100644
> --- a/arch/arm64/kernel/insn.c
> +++ b/arch/arm64/kernel/insn.c
> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
>   		return addr;
>
>   	BUG_ON(!page);
> -	set_fixmap(fixmap, page_to_phys(page));
> -
> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
> +			(addr & ~PAGE_MASK));
>   }
>
>   static void __kprobes patch_unmap(int fixmap)
>

Reviewed-by: Laura Abbott <labbott@redhat.com>

Re: [RFC] arm64:use set_fixmap_offset to make it more clear

[yalin wang]

> On Jul 23, 2015, at 21:03, Catalin Marinas <catalin.marinas@arm.com> wrote:
> 
> On Thu, Jul 23, 2015 at 07:45:53PM +0800, yalin wang wrote:
>> A little change to patch_map() function,
>> use set_fixmap_offset() to make code more clear.
>> 
>> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
>> ---
>> arch/arm64/kernel/insn.c | 5 ++---
>> 1 file changed, 2 insertions(+), 3 deletions(-)
>> 
>> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
>> index dd9671c..7dafd5a 100644
>> --- a/arch/arm64/kernel/insn.c
>> +++ b/arch/arm64/kernel/insn.c
>> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
>> 		return addr;
>> 
>> 	BUG_ON(!page);
>> -	set_fixmap(fixmap, page_to_phys(page));
>> -
>> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
>> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
>> +			(addr & ~PAGE_MASK));
> 
> It looks fine. Do you get any compiler warning for the automatic pointer
> to long conversion? You may want to add some explicit casts, otherwise:
> 
> Acked-by: Catalin Marinas <catalin.marinas@arm.com>
i have build it, there is no warning about this change. :)
Thanks

Re: [RFC] arm64:use set_fixmap_offset to make it more clear

[Mark Rutland]

On Fri, Jul 24, 2015 at 04:56:59AM +0100, yalin wang wrote:
> 
> > On Jul 23, 2015, at 21:03, Catalin Marinas <catalin.marinas@arm.com> wrote:
> > 
> > On Thu, Jul 23, 2015 at 07:45:53PM +0800, yalin wang wrote:
> >> A little change to patch_map() function,
> >> use set_fixmap_offset() to make code more clear.
> >> 
> >> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
> >> ---
> >> arch/arm64/kernel/insn.c | 5 ++---
> >> 1 file changed, 2 insertions(+), 3 deletions(-)
> >> 
> >> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
> >> index dd9671c..7dafd5a 100644
> >> --- a/arch/arm64/kernel/insn.c
> >> +++ b/arch/arm64/kernel/insn.c
> >> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
> >> 		return addr;
> >> 
> >> 	BUG_ON(!page);
> >> -	set_fixmap(fixmap, page_to_phys(page));
> >> -
> >> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
> >> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
> >> +			(addr & ~PAGE_MASK));
> > 
> > It looks fine. Do you get any compiler warning for the automatic pointer
> > to long conversion? You may want to add some explicit casts, otherwise:
> > 
> > Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> i have build it, there is no warning about this change. :)

I see no warnings with defconfig, but there's an (unrelated) set of warnings if
CONFIG_DEBUG_SET_MODULE_RONX or CONFIG_DEBUG_RODATA are enabled:

----
In file included from ./arch/arm64/include/asm/fixmap.h:85:0,
                 from arch/arm64/kernel/insn.c:32:
arch/arm64/kernel/insn.c: In function ‘__aarch64_insn_write’:
include/asm-generic/fixmap.h:73:2: warning: ‘addr’ may be used uninitialized in this function [-Wmaybe-uninitialized]
  __set_fixmap(idx, phys, flags);         \
  ^
include/asm-generic/fixmap.h:72:16: note: ‘addr’ was declared here
  unsigned long addr;          \
                ^
include/asm-generic/fixmap.h:79:2: note: in expansion of macro ‘__set_fixmap_offset’
  __set_fixmap_offset(idx, phys, FIXMAP_PAGE_NORMAL)
  ^
arch/arm64/kernel/insn.c:104:17: note: in expansion of macro ‘set_fixmap_offset’
  return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
                 ^
----

That seems to be due to the definition of __set_fixmap_offset in
asm-generic/fixmap.h:

/* Return a pointer with offset calculated */
#define __set_fixmap_offset(idx, phys, flags)		      \
({							      \
	unsigned long addr;				      \
	__set_fixmap(idx, phys, flags);			      \
	addr = fix_to_virt(idx) + ((phys) & (PAGE_SIZE - 1)); \
	addr;						      \
})

Where that new addr variable shadows patch_map's addr argument when the call to
__set_fixmap is expanded.

Which means that this patch currently breaks CONFIG_DEBUG_SET_MODULE_RONX and
CONFIG_DEBUG_RODATA.

Thanks,
Mark.

Re: [RFC] arm64:use set_fixmap_offset to make it more clear

[Mark Rutland]

On Fri, Jul 24, 2015 at 10:56:18AM +0100, Mark Rutland wrote:
> On Fri, Jul 24, 2015 at 04:56:59AM +0100, yalin wang wrote:
> > 
> > > On Jul 23, 2015, at 21:03, Catalin Marinas <catalin.marinas@arm.com> wrote:
> > > 
> > > On Thu, Jul 23, 2015 at 07:45:53PM +0800, yalin wang wrote:
> > >> A little change to patch_map() function,
> > >> use set_fixmap_offset() to make code more clear.
> > >> 
> > >> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
> > >> ---
> > >> arch/arm64/kernel/insn.c | 5 ++---
> > >> 1 file changed, 2 insertions(+), 3 deletions(-)
> > >> 
> > >> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
> > >> index dd9671c..7dafd5a 100644
> > >> --- a/arch/arm64/kernel/insn.c
> > >> +++ b/arch/arm64/kernel/insn.c
> > >> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
> > >> 		return addr;
> > >> 
> > >> 	BUG_ON(!page);
> > >> -	set_fixmap(fixmap, page_to_phys(page));
> > >> -
> > >> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
> > >> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
> > >> +			(addr & ~PAGE_MASK));
> > > 
> > > It looks fine. Do you get any compiler warning for the automatic pointer
> > > to long conversion? You may want to add some explicit casts, otherwise:
> > > 
> > > Acked-by: Catalin Marinas <catalin.marinas@arm.com>
> > i have build it, there is no warning about this change. :)
> 
> I see no warnings with defconfig, but there's an (unrelated) set of warnings if
> CONFIG_DEBUG_SET_MODULE_RONX or CONFIG_DEBUG_RODATA are enabled:

Actually they aren't unrelated after all. the unsigned long addr in
__set_fixmap_offset shadows the void* addr from patch_map making the
types match.

With a few underscores addrd to __set_fixmap_offset's addr, I get the
following regarding patch_map:

----
In file included from ./arch/arm64/include/asm/fixmap.h:85:0,
                 from arch/arm64/kernel/insn.c:32:
arch/arm64/kernel/insn.c: In function ‘patch_map’:
arch/arm64/kernel/insn.c:105:10: error: invalid operands to binary & (have ‘void *’ and ‘long unsigned int’)
    (addr & ~PAGE_MASK));
          ^
include/asm-generic/fixmap.h:73:20: note: in definition of macro ‘__set_fixmap_offset’
  __set_fixmap(idx, phys, flags);         \
                    ^
arch/arm64/kernel/insn.c:104:17: note: in expansion of macro ‘set_fixmap_offset’
  return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
                 ^
arch/arm64/kernel/insn.c:105:10: error: invalid operands to binary & (have ‘void *’ and ‘long unsigned int’)
    (addr & ~PAGE_MASK));
          ^
include/asm-generic/fixmap.h:74:32: note: in definition of macro ‘__set_fixmap_offset’
  __addr = fix_to_virt(idx) + ((phys) & (PAGE_SIZE - 1)); \
                                ^
arch/arm64/kernel/insn.c:104:17: note: in expansion of macro ‘set_fixmap_offset’
  return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
                 ^
make[1]: *** [arch/arm64/kernel/insn.o] Error 1
make: *** [arch/arm64/kernel/insn.o] Error 2
----

we need both the cast in patch_map, and an update to __set_fixmap_offset to
prevent the name clash.

Thanks,
Mark.

[RFC V2] arm64:use set_fixmap_offset to make it more clear

[yalin wang]

A little change to patch_map() function,
use set_fixmap_offset() to make code more clear.

Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
---
 arch/arm64/kernel/insn.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
index dd9671c..f341866 100644
--- a/arch/arm64/kernel/insn.c
+++ b/arch/arm64/kernel/insn.c
@@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
 		return addr;
 
 	BUG_ON(!page);
-	set_fixmap(fixmap, page_to_phys(page));
-
-	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
+	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
+			(uintaddr & ~PAGE_MASK));
 }
 
 static void __kprobes patch_unmap(int fixmap)
-- 
1.9.1

Re: [RFC] arm64:use set_fixmap_offset to make it more clear

[yalin wang]

> 在 2015年7月24日，18:07，Mark Rutland <mark.rutland@arm.com> 写道：
> 
> On Fri, Jul 24, 2015 at 10:56:18AM +0100, Mark Rutland wrote:
>> On Fri, Jul 24, 2015 at 04:56:59AM +0100, yalin wang wrote:
>>> 
>>>> On Jul 23, 2015, at 21:03, Catalin Marinas <catalin.marinas@arm.com> wrote:
>>>> 
>>>> On Thu, Jul 23, 2015 at 07:45:53PM +0800, yalin wang wrote:
>>>>> A little change to patch_map() function,
>>>>> use set_fixmap_offset() to make code more clear.
>>>>> 
>>>>> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
>>>>> ---
>>>>> arch/arm64/kernel/insn.c | 5 ++---
>>>>> 1 file changed, 2 insertions(+), 3 deletions(-)
>>>>> 
>>>>> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
>>>>> index dd9671c..7dafd5a 100644
>>>>> --- a/arch/arm64/kernel/insn.c
>>>>> +++ b/arch/arm64/kernel/insn.c
>>>>> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
>>>>> 		return addr;
>>>>> 
>>>>> 	BUG_ON(!page);
>>>>> -	set_fixmap(fixmap, page_to_phys(page));
>>>>> -
>>>>> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
>>>>> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
>>>>> +			(addr & ~PAGE_MASK));
>>>> 
>>>> It looks fine. Do you get any compiler warning for the automatic pointer
>>>> to long conversion? You may want to add some explicit casts, otherwise:
>>>> 
>>>> Acked-by: Catalin Marinas <catalin.marinas@arm.com>
>>> i have build it, there is no warning about this change. :)
>> 
>> I see no warnings with defconfig, but there's an (unrelated) set of warnings if
>> CONFIG_DEBUG_SET_MODULE_RONX or CONFIG_DEBUG_RODATA are enabled:
> 
> Actually they aren't unrelated after all. the unsigned long addr in
> __set_fixmap_offset shadows the void* addr from patch_map making the
> types match.
> 
> With a few underscores addrd to __set_fixmap_offset's addr, I get the
> following regarding patch_map:
> 
> ----
> In file included from ./arch/arm64/include/asm/fixmap.h:85:0,
>                 from arch/arm64/kernel/insn.c:32:
> arch/arm64/kernel/insn.c: In function ‘patch_map’:
> arch/arm64/kernel/insn.c:105:10: error: invalid operands to binary & (have ‘void *’ and ‘long unsigned int’)
>    (addr & ~PAGE_MASK));
>          ^
> include/asm-generic/fixmap.h:73:20: note: in definition of macro ‘__set_fixmap_offset’
>  __set_fixmap(idx, phys, flags);         \
>                    ^
> arch/arm64/kernel/insn.c:104:17: note: in expansion of macro ‘set_fixmap_offset’
>  return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
>                 ^
> arch/arm64/kernel/insn.c:105:10: error: invalid operands to binary & (have ‘void *’ and ‘long unsigned int’)
>    (addr & ~PAGE_MASK));
>          ^
> include/asm-generic/fixmap.h:74:32: note: in definition of macro ‘__set_fixmap_offset’
>  __addr = fix_to_virt(idx) + ((phys) & (PAGE_SIZE - 1)); \
>                                ^
> arch/arm64/kernel/insn.c:104:17: note: in expansion of macro ‘set_fixmap_offset’
>  return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
>                 ^
> make[1]: *** [arch/arm64/kernel/insn.o] Error 1
> make: *** [arch/arm64/kernel/insn.o] Error 2
> ----
> 
> we need both the cast in patch_map, and an update to __set_fixmap_offset to
> prevent the name clash.
> 
> Thanks,
> Mark.
i see, i will send V2 to solve this.

Thanks

Re: [RFC V2] arm64:use set_fixmap_offset to make it more clear

[yalin wang]

> On Jul 24, 2015, at 19:52, yalin wang <yalin.wang2010@gmail.com> wrote:
> 
> A little change to patch_map() function,
> use set_fixmap_offset() to make code more clear.
> 
> Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
> ---
> arch/arm64/kernel/insn.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
> index dd9671c..f341866 100644
> --- a/arch/arm64/kernel/insn.c
> +++ b/arch/arm64/kernel/insn.c
> @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
> 		return addr;
> 
> 	BUG_ON(!page);
> -	set_fixmap(fixmap, page_to_phys(page));
> -
> -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
> +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
> +			(uintaddr & ~PAGE_MASK));
> }
> 
> static void __kprobes patch_unmap(int fixmap)
> -- 
> 1.9.1
> 
Marinas,
this V2 patch can build without warning even CONFIG_DEBUG_SET_MODULE_RONX enabled,
could you review it.
Thanks.

Re: [RFC V2] arm64:use set_fixmap_offset to make it more clear

[Will Deacon]

On Tue, Jul 28, 2015 at 07:31:20AM +0100, yalin wang wrote:
> 
> > On Jul 24, 2015, at 19:52, yalin wang <yalin.wang2010@gmail.com> wrote:
> > 
> > A little change to patch_map() function,
> > use set_fixmap_offset() to make code more clear.
> > 
> > Signed-off-by: yalin wang <yalin.wang2010@gmail.com>
> > ---
> > arch/arm64/kernel/insn.c | 5 ++---
> > 1 file changed, 2 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/arm64/kernel/insn.c b/arch/arm64/kernel/insn.c
> > index dd9671c..f341866 100644
> > --- a/arch/arm64/kernel/insn.c
> > +++ b/arch/arm64/kernel/insn.c
> > @@ -101,9 +101,8 @@ static void __kprobes *patch_map(void *addr, int fixmap)
> > 		return addr;
> > 
> > 	BUG_ON(!page);
> > -	set_fixmap(fixmap, page_to_phys(page));
> > -
> > -	return (void *) (__fix_to_virt(fixmap) + (uintaddr & ~PAGE_MASK));
> > +	return (void *)set_fixmap_offset(fixmap, page_to_phys(page) +
> > +			(uintaddr & ~PAGE_MASK));
> > }
> > 
> > static void __kprobes patch_unmap(int fixmap)
> > -- 
> > 1.9.1
> > 
> Marinas,
> this V2 patch can build without warning even CONFIG_DEBUG_SET_MODULE_RONX enabled,
> could you review it.
> Thanks.

I queued this already for 4.3. There's a (benign) sparse warning about the
addr parameter to patch_map being overridden by the local scope of
set_fixmap_offset which could be solved with a healthy portion of
underscores in the core code.

Will