On 24/10/2018 00:33, Vineet Gupta wrote:
> On 10/17/2018 04:30 AM, Alexey Brodkin wrote:
>> It turned out we used to use default implementation of sched_clock()
>> from kernel/sched/clock.c which was as precise as 1/HZ, i.e.
>> by default we had 10 msec granularity of time measurement.
>>
>> Now gi
On 24/10/2018 00:33, Vineet Gupta wrote:
> On 10/17/2018 04:30 AM, Alexey Brodkin wrote:
>> It turned out we used to use default implementation of sched_clock()
>> from kernel/sched/clock.c which was as precise as 1/HZ, i.e.
>> by default we had 10 msec granularity of time measurement.
>>
>> Now gi
On Wed, Oct 31, 2018 at 12:28:41PM -0700, Florian Fainelli wrote:
> ARM64 is the only architecture that re-defines
> __early_init_dt_declare_initrd() in order for that function to populate
> initrd_start/initrd_end with physical addresses instead of virtual
> addresses. Instead of having an overrid
On Wed, Oct 31, 2018 at 12:28:37PM -0700, Florian Fainelli wrote:
> Hi all,
>
> Changes in v3:
>
> - use C conditionals in drivers/of/fdt.c
> - added check on phys_initrd_size in arch/arm64/mm/init.c to determine
> whether initrd_start must be populated
> - fixed a build warning with ARC that w
Hi Florian,
On 31 October 2018 at 20:28, Florian Fainelli wrote:
> ARM64 is the only architecture that re-defines
> __early_init_dt_declare_initrd() in order for that function to populate
> initrd_start/initrd_end with physical addresses instead of virtual
> addresses. Instead of having an overri
On 11/5/18 12:39 PM, Ard Biesheuvel wrote:
> Hi Florian,
>
> On 31 October 2018 at 20:28, Florian Fainelli wrote:
>> ARM64 is the only architecture that re-defines
>> __early_init_dt_declare_initrd() in order for that function to populate
>> initrd_start/initrd_end with physical addresses instead
On 5 November 2018 at 21:41, Florian Fainelli wrote:
> On 11/5/18 12:39 PM, Ard Biesheuvel wrote:
>> Hi Florian,
>>
>> On 31 October 2018 at 20:28, Florian Fainelli wrote:
>>> ARM64 is the only architecture that re-defines
>>> __early_init_dt_declare_initrd() in order for that function to populat
On 11/5/18 12:44 PM, Ard Biesheuvel wrote:
> On 5 November 2018 at 21:41, Florian Fainelli wrote:
>> On 11/5/18 12:39 PM, Ard Biesheuvel wrote:
>>> Hi Florian,
>>>
>>> On 31 October 2018 at 20:28, Florian Fainelli wrote:
ARM64 is the only architecture that re-defines
__early_init_dt_dec
On 5 November 2018 at 21:51, Florian Fainelli wrote:
> On 11/5/18 12:44 PM, Ard Biesheuvel wrote:
>> On 5 November 2018 at 21:41, Florian Fainelli wrote:
>>> On 11/5/18 12:39 PM, Ard Biesheuvel wrote:
Hi Florian,
On 31 October 2018 at 20:28, Florian Fainelli wrote:
> ARM64 is
On 11/5/18 1:00 PM, Ard Biesheuvel wrote:
> On 5 November 2018 at 21:51, Florian Fainelli wrote:
>> On 11/5/18 12:44 PM, Ard Biesheuvel wrote:
>>> On 5 November 2018 at 21:41, Florian Fainelli wrote:
On 11/5/18 12:39 PM, Ard Biesheuvel wrote:
> Hi Florian,
>
> On 31 October 2018
On 5 November 2018 at 22:05, Florian Fainelli wrote:
> On 11/5/18 1:00 PM, Ard Biesheuvel wrote:
>> On 5 November 2018 at 21:51, Florian Fainelli wrote:
>>> On 11/5/18 12:44 PM, Ard Biesheuvel wrote:
On 5 November 2018 at 21:41, Florian Fainelli wrote:
> On 11/5/18 12:39 PM, Ard Biesheu
Make phys_initrd_start and phys_initrd_size global variables declared in
init/do_mounts_initrd.c such that we can later have generic code in
drivers/of/fdt.c populate those variables for us.
This requires both the ARM and unicore32 implementations to be properly
guarded against CONFIG_BLK_DEV_INIT
Now that we have central and global variables holding the physical
address and size of the initrd, we can have
early_init_dt_check_for_initrd() populate
phys_initrd_start/phys_initrd_size for us.
This allows us to remove a chunk of code from arch/arm/mm/init.c
introduced with commit 65939301acdb (
Hi all,
Changes in v4:
- dropped initrd_below_start_ok assignment in ARM64, not necessary at
all (Ard)
- replace #ifdef CONFIG_BLK_DEV_INITRD with if
(IS_ENABLED(CONFIG_BLK_DEV_INITRD) for consistency with other parts
of arm64_memblock_init() (Rob)
Changes in v3:
- use C conditionals in d
This will conflict with a subsequent change making phys_initrd_start and
phys_initrd_size global variables. nds32 does not make use of those nor
provides a suitable declarations so just get rid of them.
Signed-off-by: Florian Fainelli
---
arch/nds32/mm/init.c | 2 --
1 file changed, 2 deletions(
Now that ARM64 uses phys_initrd_start/phys_initrd_size, we can get rid
of its custom __early_init_dt_declare_initrd() which causes a fair
amount of objects rebuild when changing CONFIG_BLK_DEV_INITRD. In order
to make sure ARM64 does not produce a BUG() when VM debugging is turned
on though, we mus
ARM64 is the only architecture that re-defines
__early_init_dt_declare_initrd() in order for that function to populate
initrd_start/initrd_end with physical addresses instead of virtual
addresses. Instead of having an override we can leverage
drivers/of/fdt.c populating phys_initrd_start/phys_initr
ARC, ARM, ARM64 and Unicore32 are all capable of parsing the "initrd="
command line parameter to allow specifying the physical address and size
of an initrd. Move that parsing into init/do_mounts_initrd.c such that
we no longer duplicate that logic.
Signed-off-by: Florian Fainelli
---
arch/arc/m
On 10/31/18 8:24 PM, Peter Xu wrote:
> In do_page_fault() of ARC we have:
>
> ...
> fault = handle_mm_fault(vma, address, flags);
>
> /* If Pagefault was interrupted by SIGKILL, exit page fault "early" */
> if (unlikely(fatal_signal_pending(current))) {
>
On Tue, Nov 06, 2018 at 12:48:31AM +, Vineet Gupta wrote:
> On 10/31/18 8:24 PM, Peter Xu wrote:
> > In do_page_fault() of ARC we have:
> >
> > ...
> > fault = handle_mm_fault(vma, address, flags);
> >
> > /* If Pagefault was interrupted by SIGKILL, exit page fault "early"
On Sun, Nov 04, 2018 at 12:56:48AM -0600, William Kucharski wrote:
>
>
> > On Nov 3, 2018, at 12:32 PM, Joel Fernandes wrote:
> >
> > Looks like more architectures don't define set_pmd_at. I am thinking the
> > easiest way forward is to just do the following, instead of defining
> > set_pmd_at
21 matches
Mail list logo
