[librm] Prepare for long-mode memory map

The bulk of the iPXE binary (the .textdata section) is physically
relocated at runtime to the top of the 32-bit address space in order
to allow space for an OS to be loaded.  The relocation is achieved
with the assistance of segmentation: we adjust the code and data
segment bases so that the link-time addresses remain valid.

Segmentation is not available (for normal code and data segments) in
long mode.  We choose to compile the C code with -mcmodel=kernel and
use a link-time address of 0xffffffffeb000000.  This choice allows us
to identity-map the entirety of the 32-bit address space, and to alias
our chosen link-time address to the physical location of our .textdata
section.  (This requires the .textdata section to always be aligned to
a page boundary.)

We simultaneously choose to set the 32-bit virtual address segment
bases such that the link-time addresses may simply be truncated to 32
bits in order to generate a valid 32-bit virtual address.  This allows
symbols in .textdata to be trivially accessed by both 32-bit and
64-bit code.

There is no (sensible) way in 32-bit assembly code to generate the
required R_X86_64_32S relocation records for these truncated symbols.
However, subtracting the fixed constant 0xffffffff00000000 has the
same effect as truncation, and can be represented in a standard
R_X86_64_32 relocation record.  We define the VIRTUAL() macro to
abstract away this truncation operation, and apply it to all
references by 32-bit (or 16-bit) assembly code to any symbols within
the .textdata section.

We define "virt_offset" for a 64-bit build as "the value to be added
to an address within .textdata in order to obtain its physical
address".  With this definition, the low 32 bits of "virt_offset" can
be treated by 32-bit code as functionally equivalent to "virt_offset"
in a 32-bit build.

We define "text16" and "data16" for a 64-bit build as the physical
addresses of the .text16 and .data16 sections.  Since a physical
address within the 32-bit address space may be used directly as a
64-bit virtual address (thanks to the identity map), this definition
provides the most natural access to variables in .text16 and .data16.
Note that this requires a minor adjustment in prot_to_real(), which
accesses .text16 using 32-bit virtual addresses.

Signed-off-by: Michael Brown <mcb30@ipxe.org>

1 files changed, 36 insertions, 3 deletions

diff --git a/src/arch/x86/include/librm.h b/src/arch/x86/include/librm.h
index 97461640..fc31c503 100644
--- a/src/arch/x86/include/librm.h
+++ b/src/arch/x86/include/librm.h
@@ -7,7 +7,6 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
  *
  * Don't change these unless you really know what you're doing.
  */
-
 #define VIRTUAL_CS 0x08
 #define VIRTUAL_DS 0x10
 #define PHYSICAL_CS 0x18
@@ -16,6 +15,40 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
 #define REAL_DS 0x30
 #define P2R_DS 0x38
 
+/* Calculate symbol address within VIRTUAL_CS or VIRTUAL_DS
+ *
+ * In a 64-bit build, we set the bases of VIRTUAL_CS and VIRTUAL_DS
+ * such that truncating a .textdata symbol value to 32 bits gives a
+ * valid 32-bit virtual address.
+ *
+ * The C code is compiled with -mcmodel=kernel and so we must place
+ * all .textdata symbols within the negative 2GB of the 64-bit address
+ * space.  Consequently, all .textdata symbols will have the MSB set
+ * after truncation to 32 bits.  This means that a straightforward
+ * R_X86_64_32 relocation record for the symbol will fail, since the
+ * truncated symbol value will not correctly zero-extend to the
+ * original 64-bit value.
+ *
+ * Using an R_X86_64_32S relocation record would work, but there is no
+ * (sensible) way to generate these relocation records within 32-bit
+ * or 16-bit code.
+ *
+ * The simplest solution is to generate an R_X86_64_32 relocation
+ * record with an addend of (-0xffffffff00000000).  Since all
+ * .textdata symbols are within the negative 2GB of the 64-bit address
+ * space, this addend acts to effectively truncate the symbol to 32
+ * bits, thereby matching the semantics of the R_X86_64_32 relocation
+ * records generated for 32-bit and 16-bit code.
+ *
+ * In a 32-bit build, this problem does not exist, and we can just use
+ * the .textdata symbol values directly.
+ */
+#ifdef __x86_64__
+#define VIRTUAL(address) ( (address) - 0xffffffff00000000 )
+#else
+#define VIRTUAL(address) (address)
+#endif
+
 #ifdef ASSEMBLY
 
 /**
@@ -24,7 +57,7 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
  * @v function		C function
  */
 .macro virtcall function
-	pushl	$\function
+	pushl	$VIRTUAL(\function)
 	call	prot_call
 .endm
 
@@ -42,7 +75,7 @@ FILE_LICENCE ( GPL2_OR_LATER_OR_UBDL );
  * @v function		C function
  */
 #define VIRT_CALL( function )						\
-	"pushl $( " #function " )\n\t"					\
+	"pushl $( " _S2 ( VIRTUAL ( function ) ) " )\n\t"		\
 	"call prot_call\n\t"
 
 /* Variables in librm.S */