Files
s390-tools/rust/pvimg
Jan Höppner e6ac87d7a2 common.mak: Remove NO_PIE_LDFLAGS
Commit c5a91199e3 ("zipl: Always build and link without PIE.")
introduced -fno-pie (for compilation) and -no-pie (for linking) for
zipl. At the time the linker was still used directly before commit
5e46632767 ("zipl: Use the compiler for linking instead of ld")
eventually switched to calling the compiler for the linking step.

During that adaption -static was introduced to the linker flags. -no-pie
was carried over as well. However -static implies -no-pie and it is
therefore not required. For GCC see also man 1 gcc (-static) [1]. Whilst
not explicitly documented, Clang shows the same behaviour. Clang also
complains when -static and -no-pie are specified in the linker step at
the same time with the following warning:

clang: warning: argument unused during compilation: '-no-pie' [-Wunused-command-line-argument]

Since -no-pie is not required, remove it and get rid of the warning.

[1] https://gcc.gnu.org/onlinedocs/gcc/Link-Options.html#index-static
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2025-11-07 14:31:47 +01:00
..
2025-11-07 14:31:47 +01:00
2025-09-15 11:47:32 +02:00
2024-12-05 15:09:03 +01:00

pvimg

pvimg create takes a kernel, key files, optionally an initrd image, optionally a file containing the kernel command line parameters, and generates a single, bootable image file. The generated image file consists of a concatenation of a plain text boot loader, the encrypted components for kernel, initrd, kernel command line, and the integrity-protected Secure Execution header, containing the metadata necessary for running the guest in protected mode. See Memory Layout for details about the internal structure of the created image.

It is possible to use the generated image as a kernel for zipl or for a direct kernel boot using QEMU.

Getting started

If all dependencies are met a simple make call in the source tree should be enough for building pvimg.

Details

The main idea of pvimg create is:

  1. Generate all keys, IVs, and other information needed for the encryption of the components and the generation of the PV header
  2. add stub stage3a (so we can calculate the memory addresses)
  3. add components: prepare the components (alignment and encryption) and add them to the memory layout
  4. build and add stage3b: generate the stage3b and add it to the memory layout
  5. generate the Secure Execution header: generate the hashes (pld, ald, and tld) of the components and create the header and IPIB
  6. parameterize the stub stage3a: uses the address of the IPIB and Secure Execution header
  7. write the final image to the specified output path and generate the boot image metadata at address 0xc000. The address 0xc000 is chosen as this is the BSS section of the stage3a loader and will therefore zeroed out as soon as the stage3a is executed and has therefore no leftovers in the memory.

Boot Loader

The boot loader consists of two parts:

  1. stage3a boot loader (cleartext), this loader is responsible for the transition into the protected mode by doing diag308 subcode 8 and 10 calls.
  2. stage3b boot loader (encrypted), this loader is very similar to the normal zipl stage3 boot loader. It will be loaded by the Ultravisor after the successful transition into protected mode. Like the zipl stage3 boot loader it moves the kernel and patches in the values for initrd and kernel command line.

The loaders have the following constraints:

  1. It must be possible to place stage3a and stage3b at a location greater than 0x10000 because the zipl stage3 loader zeroes out everything at addresses lower than 0x10000 of the image.
  2. As the stage3 loader of zipl assumes that the passed kernel image looks like a normal kernel image, the zipl stage3 loader modifies the content at the memory area 0x10400 - 0x10800, therefore we leave this area unused in our stage3a loader.
  3. The default entry address used by the zipl stage3 loader is 0x10000 so we add a simple branch to 0x11000 at 0x10000 so the zipl stage3 loader can modify the area 0x10400 - 0x10800 without affecting the stage3a loader.

Stage3b

The stage3b.bin is linked at address 0x9000, therefore it will not work at another address. The relocation support for the stage3b loader, so that it can be loaded at addresses != 0x9000, is added in the loader with the name stage3b_reloc.bin. By default, if we're talking about stage3b we refer to stage3b_reloc.bin.

Memory Layout

The memory layout of the bootable file looks like:

Start End Use
0 0x7 Short PSW, starting instruction at 0x11000
0x0c000 0x0cfff Image metadata, e.g. it includes the file offset of the SE-header
0x10000 0x10012 Branch to 0x11000
0x10013 0x10fff Left intentionally unused
0x11000 0x12fff Stage3a
0x14000 0x1[45]fff SE-header used for the diag308 call (size can be either 1 or 2 pages)
NEXT_PAGE_ALIGNED_ADDR Encrypted kernel
NEXT_PAGE_ALIGNED_ADDR Encrypted kernel parameters
NEXT_PAGE_ALIGNED_ADDR Encrypted initrd
NEXT_PAGE_ALIGNED_ADDR Encrypted stage3b_reloc
NEXT_PAGE_ALIGNED_ADDR IPIB used as argument for the diag308 call