fs/dax.c uses copy_user_page() but ARC does not provide that interface,
resulting in a build error.
Provide copy_user_page() in <asm/page.h> (beside copy_page()) and
add <asm/page.h> to fs/dax.c to fix the build error.
../fs/dax.c: In function 'copy_cow_page_dax':
../fs/dax.c:702:2: error: implicit declaration of function 'copy_user_page'; did you mean 'copy_to_user_page'? [-Werror=implicit-function-declaration]
Fixes: cccbce671582 ("filesystem-dax: convert to dax_direct_access()")
Reported-by: kernel test robot <[email protected]>
Signed-off-by: Randy Dunlap <[email protected]>
Cc: Vineet Gupta <[email protected]>
Cc: [email protected]
Cc: Dan Williams <[email protected]>
Acked-by: Vineet Gupta <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Matthew Wilcox <[email protected]>
Cc: Jan Kara <[email protected]>
Cc: [email protected]
Cc: [email protected]
---
v2: rebase, add more Cc:
arch/arc/include/asm/page.h | 1 +
fs/dax.c | 1 +
2 files changed, 2 insertions(+)
--- lnx-511-rc1.orig/fs/dax.c
+++ lnx-511-rc1/fs/dax.c
@@ -25,6 +25,7 @@
#include <linux/sizes.h>
#include <linux/mmu_notifier.h>
#include <linux/iomap.h>
+#include <asm/page.h>
#include <asm/pgalloc.h>
#define CREATE_TRACE_POINTS
--- lnx-511-rc1.orig/arch/arc/include/asm/page.h
+++ lnx-511-rc1/arch/arc/include/asm/page.h
@@ -10,6 +10,7 @@
#ifndef __ASSEMBLY__
#define clear_page(paddr) memset((paddr), 0, PAGE_SIZE)
+#define copy_user_page(to, from, vaddr, pg) copy_page(to, from)
#define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
struct vm_area_struct;
On Thu, Dec 31, 2020 at 08:29:14PM -0800, Randy Dunlap wrote:
> fs/dax.c uses copy_user_page() but ARC does not provide that interface,
> resulting in a build error.
>
> Provide copy_user_page() in <asm/page.h> (beside copy_page()) and
> add <asm/page.h> to fs/dax.c to fix the build error.
>
> ../fs/dax.c: In function 'copy_cow_page_dax':
> ../fs/dax.c:702:2: error: implicit declaration of function 'copy_user_page'; did you mean 'copy_to_user_page'? [-Werror=implicit-function-declaration]
>
> Fixes: cccbce671582 ("filesystem-dax: convert to dax_direct_access()")
> Reported-by: kernel test robot <[email protected]>
> Signed-off-by: Randy Dunlap <[email protected]>
Looks reasonable
Reviewed-by: Ira Weiny <[email protected]>
> Cc: Vineet Gupta <[email protected]>
> Cc: [email protected]
> Cc: Dan Williams <[email protected]>
> Acked-by: Vineet Gupta <[email protected]>
> Cc: Andrew Morton <[email protected]>
> Cc: Matthew Wilcox <[email protected]>
> Cc: Jan Kara <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> ---
> v2: rebase, add more Cc:
>
> arch/arc/include/asm/page.h | 1 +
> fs/dax.c | 1 +
> 2 files changed, 2 insertions(+)
>
> --- lnx-511-rc1.orig/fs/dax.c
> +++ lnx-511-rc1/fs/dax.c
> @@ -25,6 +25,7 @@
> #include <linux/sizes.h>
> #include <linux/mmu_notifier.h>
> #include <linux/iomap.h>
> +#include <asm/page.h>
> #include <asm/pgalloc.h>
>
> #define CREATE_TRACE_POINTS
> --- lnx-511-rc1.orig/arch/arc/include/asm/page.h
> +++ lnx-511-rc1/arch/arc/include/asm/page.h
> @@ -10,6 +10,7 @@
> #ifndef __ASSEMBLY__
>
> #define clear_page(paddr) memset((paddr), 0, PAGE_SIZE)
> +#define copy_user_page(to, from, vaddr, pg) copy_page(to, from)
> #define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
>
> struct vm_area_struct;
> _______________________________________________
> Linux-nvdimm mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
On Thu, Dec 31, 2020 at 8:29 PM Randy Dunlap <[email protected]> wrote:
>
> fs/dax.c uses copy_user_page() but ARC does not provide that interface,
> resulting in a build error.
>
> Provide copy_user_page() in <asm/page.h> (beside copy_page()) and
> add <asm/page.h> to fs/dax.c to fix the build error.
>
> ../fs/dax.c: In function 'copy_cow_page_dax':
> ../fs/dax.c:702:2: error: implicit declaration of function 'copy_user_page'; did you mean 'copy_to_user_page'? [-Werror=implicit-function-declaration]
>
> Fixes: cccbce671582 ("filesystem-dax: convert to dax_direct_access()")
> Reported-by: kernel test robot <[email protected]>
> Signed-off-by: Randy Dunlap <[email protected]>
> Cc: Vineet Gupta <[email protected]>
> Cc: [email protected]
> Cc: Dan Williams <[email protected]>
> Acked-by: Vineet Gupta <[email protected]>
> Cc: Andrew Morton <[email protected]>
> Cc: Matthew Wilcox <[email protected]>
> Cc: Jan Kara <[email protected]>
> Cc: [email protected]
> Cc: [email protected]
> ---
> v2: rebase, add more Cc:
>
> arch/arc/include/asm/page.h | 1 +
> fs/dax.c | 1 +
> 2 files changed, 2 insertions(+)
>
> --- lnx-511-rc1.orig/fs/dax.c
> +++ lnx-511-rc1/fs/dax.c
> @@ -25,6 +25,7 @@
> #include <linux/sizes.h>
> #include <linux/mmu_notifier.h>
> #include <linux/iomap.h>
> +#include <asm/page.h>
I would expect this to come from one of the linux/ includes like
linux/mm.h. asm/ headers are implementation linux/ headers are api.
Once you drop that then the subject of this patch can just be "arc:
add a copy_user_page() implementation", and handled by the arc
maintainer (or I can take it with Vineet's ack).
> #include <asm/pgalloc.h>
Yes, this one should have a linux/ api header to front it, but that's
a cleanup for another day.
>
> #define CREATE_TRACE_POINTS
> --- lnx-511-rc1.orig/arch/arc/include/asm/page.h
> +++ lnx-511-rc1/arch/arc/include/asm/page.h
> @@ -10,6 +10,7 @@
> #ifndef __ASSEMBLY__
>
> #define clear_page(paddr) memset((paddr), 0, PAGE_SIZE)
> +#define copy_user_page(to, from, vaddr, pg) copy_page(to, from)
> #define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
>
> struct vm_area_struct;
On Mon, Jan 04, 2021 at 12:13:02PM -0800, Dan Williams wrote:
> On Thu, Dec 31, 2020 at 8:29 PM Randy Dunlap <[email protected]> wrote:
> > +++ lnx-511-rc1/fs/dax.c
> > @@ -25,6 +25,7 @@
> > #include <linux/sizes.h>
> > #include <linux/mmu_notifier.h>
> > #include <linux/iomap.h>
> > +#include <asm/page.h>
>
> I would expect this to come from one of the linux/ includes like
> linux/mm.h. asm/ headers are implementation linux/ headers are api.
It does indeed come from linux/mm.h already. And a number of
other random places, including linux/serial.h. Our headers are a mess,
but they shouldn't be made worse by adding _this_ include. So I
second Dan's objection here.
> Once you drop that then the subject of this patch can just be "arc:
> add a copy_user_page() implementation", and handled by the arc
> maintainer (or I can take it with Vineet's ack).
>
> > #include <asm/pgalloc.h>
>
> Yes, this one should have a linux/ api header to front it, but that's
> a cleanup for another day.
Definitely more involved.
On 1/4/21 12:13 PM, Dan Williams wrote:
> On Thu, Dec 31, 2020 at 8:29 PM Randy Dunlap <[email protected]> wrote:
>>
>> fs/dax.c uses copy_user_page() but ARC does not provide that interface,
>> resulting in a build error.
>>
>> Provide copy_user_page() in <asm/page.h> (beside copy_page()) and
>> add <asm/page.h> to fs/dax.c to fix the build error.
>>
>> ../fs/dax.c: In function 'copy_cow_page_dax':
>> ../fs/dax.c:702:2: error: implicit declaration of function 'copy_user_page'; did you mean 'copy_to_user_page'? [-Werror=implicit-function-declaration]
>>
>> Fixes: cccbce671582 ("filesystem-dax: convert to dax_direct_access()")
>> Reported-by: kernel test robot <[email protected]>
>> Signed-off-by: Randy Dunlap <[email protected]>
>> Cc: Vineet Gupta <[email protected]>
>> Cc: [email protected]
>> Cc: Dan Williams <[email protected]>
>> Acked-by: Vineet Gupta <[email protected]>
>> Cc: Andrew Morton <[email protected]>
>> Cc: Matthew Wilcox <[email protected]>
>> Cc: Jan Kara <[email protected]>
>> Cc: [email protected]
>> Cc: [email protected]
>> ---
>> v2: rebase, add more Cc:
>>
>> arch/arc/include/asm/page.h | 1 +
>> fs/dax.c | 1 +
>> 2 files changed, 2 insertions(+)
>>
>> --- lnx-511-rc1.orig/fs/dax.c
>> +++ lnx-511-rc1/fs/dax.c
>> @@ -25,6 +25,7 @@
>> #include <linux/sizes.h>
>> #include <linux/mmu_notifier.h>
>> #include <linux/iomap.h>
>> +#include <asm/page.h>
>
> I would expect this to come from one of the linux/ includes like
> linux/mm.h. asm/ headers are implementation linux/ headers are api.
>
> Once you drop that then the subject of this patch can just be "arc:
> add a copy_user_page() implementation", and handled by the arc
> maintainer (or I can take it with Vineet's ack).
Got it. Thanks.
Vineet is copied. I expect that he will take the v3 patch.
>> #include <asm/pgalloc.h>
>
> Yes, this one should have a linux/ api header to front it, but that's
> a cleanup for another day.
That line is only part of the contextual diff in this patch.
I guess you are just commenting in general, along with your earlier
paragraph.
>>
>> #define CREATE_TRACE_POINTS
>> --- lnx-511-rc1.orig/arch/arc/include/asm/page.h
>> +++ lnx-511-rc1/arch/arc/include/asm/page.h
>> @@ -10,6 +10,7 @@
>> #ifndef __ASSEMBLY__
>>
>> #define clear_page(paddr) memset((paddr), 0, PAGE_SIZE)
>> +#define copy_user_page(to, from, vaddr, pg) copy_page(to, from)
>> #define copy_page(to, from) memcpy((to), (from), PAGE_SIZE)
>>
>> struct vm_area_struct;
--
~Randy
On Mon, Jan 4, 2021 at 7:47 PM Randy Dunlap <[email protected]> wrote:
>
> On 1/4/21 12:13 PM, Dan Williams wrote:
> > On Thu, Dec 31, 2020 at 8:29 PM Randy Dunlap <[email protected]> wrote:
> >>
> >> fs/dax.c uses copy_user_page() but ARC does not provide that interface,
> >> resulting in a build error.
> >>
> >> Provide copy_user_page() in <asm/page.h> (beside copy_page()) and
> >> add <asm/page.h> to fs/dax.c to fix the build error.
> >>
> >> ../fs/dax.c: In function 'copy_cow_page_dax':
> >> ../fs/dax.c:702:2: error: implicit declaration of function 'copy_user_page'; did you mean 'copy_to_user_page'? [-Werror=implicit-function-declaration]
> >>
> >> Fixes: cccbce671582 ("filesystem-dax: convert to dax_direct_access()")
> >> Reported-by: kernel test robot <[email protected]>
> >> Signed-off-by: Randy Dunlap <[email protected]>
> >> Cc: Vineet Gupta <[email protected]>
> >> Cc: [email protected]
> >> Cc: Dan Williams <[email protected]>
> >> Acked-by: Vineet Gupta <[email protected]>
> >> Cc: Andrew Morton <[email protected]>
> >> Cc: Matthew Wilcox <[email protected]>
> >> Cc: Jan Kara <[email protected]>
> >> Cc: [email protected]
> >> Cc: [email protected]
> >> ---
> >> v2: rebase, add more Cc:
> >>
> >> arch/arc/include/asm/page.h | 1 +
> >> fs/dax.c | 1 +
> >> 2 files changed, 2 insertions(+)
> >>
> >> --- lnx-511-rc1.orig/fs/dax.c
> >> +++ lnx-511-rc1/fs/dax.c
> >> @@ -25,6 +25,7 @@
> >> #include <linux/sizes.h>
> >> #include <linux/mmu_notifier.h>
> >> #include <linux/iomap.h>
> >> +#include <asm/page.h>
> >
> > I would expect this to come from one of the linux/ includes like
> > linux/mm.h. asm/ headers are implementation linux/ headers are api.
> >
> > Once you drop that then the subject of this patch can just be "arc:
> > add a copy_user_page() implementation", and handled by the arc
> > maintainer (or I can take it with Vineet's ack).
>
> Got it. Thanks.
> Vineet is copied. I expect that he will take the v3 patch.
>
> >> #include <asm/pgalloc.h>
> >
> > Yes, this one should have a linux/ api header to front it, but that's
> > a cleanup for another day.
>
> That line is only part of the contextual diff in this patch.
> I guess you are just commenting in general, along with your earlier
> paragraph.
Yes, I was prefetching a "...hey, but, there's an asm/ include right
below this one?"
On Thu, Dec 31, 2020 at 08:29:14PM -0800, Randy Dunlap wrote:
> fs/dax.c uses copy_user_page() but ARC does not provide that interface,
> resulting in a build error.
>
> Provide copy_user_page() in <asm/page.h> (beside copy_page()) and
> add <asm/page.h> to fs/dax.c to fix the build error.
>
> ../fs/dax.c: In function 'copy_cow_page_dax':
> ../fs/dax.c:702:2: error: implicit declaration of function 'copy_user_page'; did you mean 'copy_to_user_page'? [-Werror=implicit-function-declaration]
Could somebody explain what the force-cast is doing in there?
I mean, the call is
copy_user_page(vto, (void __force *)kaddr, vaddr, to);
kaddr is a local variable there, declared as void *; AFAICS, that
had been pure cargo-cult since
commit 7a9eb20666317794d0279843fbd091af93907780
Author: Dan Williams <[email protected]>
Date: Fri Jun 3 18:06:47 2016 -0700
pmem: kill __pmem address space
I mean, it's been more than 4 years, time to bury that body...