Re: [PATCH] evl: fix the gdb accessing control buffer issue.

[Philippe Gerum <rpm@xenomai.org>]

Florian Bezdeka <florian.bezdeka@siemens.com> writes:

> On Mon, 2024-12-09 at 08:37 +0000, Qichen Qiu wrote:
>> Currently, EVL maps the control buffer using remap_pfn_range, tagging 
>> the memory with VM_IO. However, this prevents access by gdb.
>> 
>> This patch introduces the control_mmap_access function for the control 
>> VMA, enabling gdb access when CONFIG_HAVE_IOREMAP_PROT is supported on 
>> the target architecture.
>
> @Philippe: I haven't checked the details yet, but the memory mapping is
> different for Xenomai 3? Just wondering why this is necessary at all.
>

Cobalt gets the u-mmapped heaps from the vmalloc space
(cobalt_umm_init), so we can populate the vma using vm_insert_page()
there (mmap_kmem_helper), which should not set VM_IO. EVL currently
obtains the heap from kmalloc instead, hence remap_pfn_range() to map a
range from the logical memory space. IOW, "that should be fine already"
(famous last words).

> Does that mean that gdb did not work? On which platforms/architectures?
> Is there no test for gdb like the gdb test in smokey?
>

The existing gdb test only checks the synchronous stop feature I
believe, not accessing the memory behind those mappings.

-- 
Philippe.