u-boot-rpi* has the existing "make first page reserved" code (shown); this appears to be where any fix would go (it works without FreeBSD changes)

[Mark Millard]

[The sysutils/u-boot-rpi* change is sufficient to
be a work around allowing existing FreeBSD versions
to be used (up to other separate problems).]

On 2020-Feb-13, at 13:22, Mark Millard <marklmi at yahoo.com> wrote:



> On 2020-Feb-13, at 12:13, Kyle Evans <kevans at freebsd.org> wrote:
> 
>> On Thu, Feb 13, 2020 at 1:43 PM Mark Millard <marklmi at yahoo.com> wrote:
>>> [ ... trim ...]
>>> 
>>> int ft_board_setup(void *blob, bd_t *bd)
>>> {
>>>       /*
>>>        * For now, we simply always add the simplefb DT node. Later, we
>>>        * should be more intelligent, and e.g. only do this if no enabled DT
>>>        * node exists for the "real" graphics driver.
>>>        */
>>>       lcd_dt_simplefb_add_node(blob);
>>> 
>>> #ifdef CONFIG_EFI_LOADER
>>>       /* Reserve the spin table */
>>>       efi_add_memory_map(0, 1, EFI_RESERVED_MEMORY_TYPE, 0);
>>> #endif
>>>       return 0;
>>> }
>>> 
>>> [... more trim ...]
>> 
>> This is largely irrelevant unless we're consulting either or both of
>> /reserved-memory and the EFI memory map. memreserve and the memory map
>> from U-Boot should likely both be fixed to reflect reality.
>> 
> 
> That is the only code causing the first page to have
> its exsting Reserved status. Otherwise it too would
> be replaced by the kernel.
> 
> The Reeserved status is from the EFI interface,
> not the FDT/DTB one.

So I've reverted armstub8-gic.bin to the normal
one ffrom sysutils/rpi-firmware and reverted
the kernel investigative HACKs, including the
one that enabled booting and operating:

# svnlite diff /usr/src/sys/dev/fdt/fdt_common.c
Index: /usr/src/sys/dev/fdt/fdt_common.c
===================================================================
--- /usr/src/sys/dev/fdt/fdt_common.c	(revision 357529)
+++ /usr/src/sys/dev/fdt/fdt_common.c	(working copy)
@@ -485,7 +485,18 @@
 
 	tuples = res_len / tuple_size;
 	reservep = (pcell_t *)&reserve;
+#ifdef __aarch64__
+	//HACK!!!
+	// Reserve the first few pages, for example to
+	// preserve armstub8-gic.bin or armstub.bin
+	// content.
+	mr[0].mr_start= 0;
+	mr[0].mr_size= 2*4096;
+	tuples++;
+	for (i = 1; i < tuples; i++) {
+#else
 	for (i = 0; i < tuples; i++) {
+#endif
 
 		rv = fdt_data_to_res(reservep, addr_cells, size_cells,
 			(u_long *)&mr[i].mr_start, (u_long *)&mr[i].mr_size);
@@ -512,6 +523,11 @@
 
 	root = OF_finddevice("/reserved-memory");
 	if (root == -1) {
+		// Fail over to checking for and handling memreserve,
+		// such as for a RPi4B.
+		if (0 == fdt_get_reserved_regions(reserved,mreserved))
+			return (0);
+
 		return (ENXIO);
 	}
 
# svnlite revert /usr/src/sys/dev/fdt/fdt_common.c
Reverted 'sys/dev/fdt/fdt_common.c'



So sysutils/rpi-firmware/ and FreeBSD itself are no
longer tailored to the problem.

But I've also switched to a sysutils/u-boot-rpi4 based on:

# diff ./board/raspberrypi/rpi/rpi.c.orig ./board/raspberrypi/rpi/rpi.c
485c485
< 	efi_add_memory_map(0, 1, EFI_RESERVED_MEMORY_TYPE, 0);
---
> 	efi_add_memory_map(0, 2, EFI_RESERVED_MEMORY_TYPE, 0);

And the result is that head -r357529
boots and operates just fine.

So FreeBSD itself does not need to change just to have
a work-around.

(I make no claim to know the best direction for a
long term fix.)


FYI: boot -v on the RPi4B used for testing now shows:

                   Type     Physical      Virtual   #Pages Attr
               Reserved 000000000000            0 00000002 WB 
     ConventionalMemory 000000002000         2000 00007ef0 WB 
       BootServicesData 000007ef2000      7ef2000 0000001c WB 
     ConventionalMemory 000007f0e000      7f0e000 00029f93 WB 
       BootServicesData 000031ea1000     31ea1000 00000001 WB 
             LoaderData 000031ea2000     31ea2000 00008001 WB 
             LoaderCode 000039ea3000     39ea3000 000000a6 WB 
               Reserved 000039f49000     39f49000 00000007 WB 
       BootServicesData 000039f50000     39f50000 00000001 WB 
               Reserved 000039f51000     39f51000 00000002 WB 
    RuntimeServicesData 000039f53000     39f53000 00000001 WB RUNTIME
               Reserved 000039f54000     39f54000 00000001 WB 
       BootServicesData 000039f55000     39f55000 00000002 WB 
    RuntimeServicesData 000039f57000     39f57000 00000001 WB RUNTIME
             LoaderData 000039f58000     39f58000 00001408 WB 
    RuntimeServicesCode 00003b360000     3b360000 00000010 WB RUNTIME
             LoaderData 00003b370000     3b370000 00000090 WB 
       BootServicesData 000040000000     40000000 000bc000 WB 
         MemoryMappedIO 0000fe100000     fe100000 00000001 RUNTIME
Physical memory chunk(s):
  0x00002000 - 0x39f48fff,   927 MB ( 237383 pages)
  0x39f50000 - 0x39f50fff,     0 MB (      1 pages)
  0x39f53000 - 0x39f53fff,     0 MB (      1 pages)
  0x39f55000 - 0x3b35ffff,    20 MB (   5131 pages)
  0x3b370000 - 0x3b3fffff,     0 MB (    144 pages)
  0x40000000 - 0xfbffffff,  3008 MB ( 770048 pages)
Excluded memory regions:
  0x00000000 - 0x00001fff,     0 MB (      2 pages) NoAlloc 
  0x32000000 - 0x33795fff,    23 MB (   6038 pages) NoAlloc 
  0x39f49000 - 0x39f4ffff,     0 MB (      7 pages) NoAlloc 
  0x39f51000 - 0x39f54fff,     0 MB (      4 pages) NoAlloc 
  0x39f57000 - 0x39f57fff,     0 MB (      1 pages) NoAlloc 
  0x3b360000 - 0x3b36ffff,     0 MB (     16 pages) NoAlloc 
  0x3e513000 - 0x3ebebfff,     6 MB (   1753 pages) NoAlloc 
  0xfe100000 - 0xfe100fff,     0 MB (      1 pages) NoAlloc 




===
Mark Millard
marklmi at yahoo.com
( dsl-only.net went
away in early 2018-Mar)