Ticket 4503

[Gedare Bloom]

On Thu, Nov 18, 2021 at 7:22 PM zack leung <zakthertemsdev at gmail.com> wrote:
>
> bump
>
> On Mon, 18 Oct 2021 at 23:58, zack leung <zakthertemsdev at gmail.com> wrote:
>
> > bump
> >
> >
> > On Sat, 25 Sept 2021 at 00:26, zack leung <zakthertemsdev at gmail.com>
> > wrote:
> >
> >> bump
> >>
> >> On Thu, 9 Sept 2021 at 02:17, zack leung <zakthertemsdev at gmail.com>
> >> wrote:
> >>
> >>> >Thanks! I guess i'm really unsure about how the pointer relates to the
> >>> amount of memory that you can use. I assume the Malloc keeps track of the
> >>> sections being used in the heap. Does this function Allocate 2 blocks since
> >>> the pointer is set to
> >>>
> >>> (uintptr_t) next_block + HEAP_ALLOC_BONUS - alloc_begin;
> >>>

No, this is calculating the size of 'this block' considering that it's
allocated memory starts at "alloc_begin" and it ends at the
"next_block" heap block metadata or header. You should read about how
malloc implementations generally work. The allocated memory has some
metadata before it to implement the linked data structures necessary
to support dynamic allocation. So a block is the header + allocated
range. Only the allocated range plus any excess before the next
block's header is considered usable.

> >>> >(I know it's a bit naive to assume this given my first question)
> >>>
> >>> think the method is _Heap_Size_of_alloc_area. It may need a wrapper
> >>> added to the protected heap wrapper for safety. I don't know if it has one.
> >>> It should be the equivalent functionality.
> >>> > How would i implement a wrapper and what safety measures need to be
> >>> implemented. So far it's checking if :  the block is not in the heap and if
> >>> it's been previously used. Is the usable memory in the heap calculated by:
> >>> the size of the block * # of blocks?
> >>>

You would have to see if the Protected Heap implementation has some
additional metadata between the end of an allocated range in one block
and the block header in the next. I don't think it matters too much
because most of the 'protected' aspects have to do with free block
management. But that should be confirmed.

The check is not "if it's been previously used" I suggest that you
look at what the invoked function does and how it is documented. Avoid
making assumptions based on names.

The usable memory is NOT the size of the block * # of blocks, because
the size of the block includes the header.

> >>> Thanks, Zack
> >>>
> >>> ---------- Forwarded message ---------
> >>> From: Joel Sherrill <joel at rtems.org>
> >>> Date: Wed, 8 Sept 2021 at 20:30
> >>> Subject: Re: Ticket 4503
> >>> To: Gedare Bloom <gedare at rtems.org>
> >>> Cc: zack leung <zakthertemsdev at gmail.com>, rtems-devel at rtems.org <
> >>> devel at rtems.org>
> >>>
> >>>
> >>>
> >>>
> >>> On Wed, Sep 8, 2021 at 11:02 AM Gedare Bloom <gedare at rtems.org> wrote:
> >>>
> >>>> Hi Zack,
> >>>>
> >>>> https://devel.rtems.org/ticket/4503
> >>>>
> >>>> The malloc implementation exists in the score as the Heap Manager.
> >>>> search for "Heap_" within cpukit to get some ideas where to look.
> >>>>
> >>>
> >>> I think the method is _Heap_Size_of_alloc_area. It may need a wrapper
> >>> added to the protected heap wrapper for safety. I don't know if it has one.
> >>> It should be the equivalent functionality.
> >>>
> >>> newlib's malloc.h will need looking at to make sure we provide the right
> >>> methods.
> >>>
> >>>
> >>>
> >>>>
> >>>> @Joel are these two tickets duplicates?
> >>>> https://devel.rtems.org/ticket/4503
> >>>> https://devel.rtems.org/ticket/4271
> >>>>
> >>>
> >>> Yes. Closed #4271 and merged comments.
> >>>
> >>> --joel
> >>>
> >>>
> >>>>
> >>>>
> >>>> -Gedare
> >>>>
> >>>> On Tue, Sep 7, 2021 at 7:14 PM zack leung <zakthertemsdev at gmail.com>
> >>>> wrote:
> >>>>
> >>>>> I was wondering about ticket 4503. I was wondering where The required
> >>>>> functionality should be in the underlying score/ capability used can be
> >>>>> found to write this function. Also how does this relate  to
> >>>>>
> >>>>> Add common malloc family extension method malloc_usable_size()
> >>>>> <https://devel.rtems.org/ticket/4271>?
> >>>>> Thanks
> >>>>> Zack
> >>>>> _______________________________________________
> >>>>> devel mailing list
> >>>>> devel at rtems.org
> >>>>> http://lists.rtems.org/mailman/listinfo/devel
> >>>>
> >>>>
> _______________________________________________
> devel mailing list
> devel at rtems.org
> http://lists.rtems.org/mailman/listinfo/devel