Conversation
|
Tagging @xypron, @brianredbeard and @jmontleon since some of the changes come from them. |
eae94d7 to
3482501
Compare
|
Rebased on top of #777. Additionally, I have dropped all definitions related to @xypron perhaps you could revisit your patch, integrating the follow-up fixes and improvements from me, @brianredbeard and @jmontleon? And of course any feedback from @vathpela would be much appreciated too :) Thanks in advance! |
3482501 to
54f16b5
Compare
|
Rebased to match the current contents of the |
Apparently strrchr() became a macro in glibc recently, breaking our test case that tries to #define it. This patch undefines it before we try to define it. Signed-off-by: Peter Jones <pjones@redhat.com>
Replace the shim-specific fork with the upstream version, specifically the most recent release. Some adjustment to shim's code are necessary to adapt to this change. Signed-off-by: Andrea Bolognani <abologna@redhat.com>
shim is a standalone EFI application so it shouldn't be necessary to look at the glibc headers when building it, and in fact attempting to do so results in a build failure. Signed-off-by: Andrea Bolognani <abologna@redhat.com>
We could theoretically set GNU_EFI_USE_REALLOCATEPOOL_ABI=0 to keep using the legacy ABI, but since gnu-efi uses the modern ABI internally and we call into its build systemd directly, doing that messes things up. Switching to the new ABI is just a matter of changing the order of arguments. Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Signed-off-by: Jason Montleon <jmontleo@redhat.com>
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Signed-off-by: Callum Farmer <gmbr3@opensuse.org>
Add what is needed to build on riscv64. Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
They don't seem to be necessary when building against the latest upstream version of gnu-efi. Signed-off-by: Andrea Bolognani <abologna@redhat.com>
@davidlt and @xypron pointed out prior changed to binutils 2.42 which added support for RISC-V EFI objects. This reflects the upstream preference to avoid adding additional architectures which are emitting flat binary files via `objcopy` (i.e. `-O binary` architectures). Signed-off-by: Brian 'redbeard' Harrington <redbeard@dead-city.org>
The SBAT-related definitions are missing from upstream gnu-efi, so those are retained. The rest of the file is completely identical to the upstream counterpart. Signed-off-by: Jason Montleon <jmontleo@redhat.com>
Signed-off-by: Jason Montleon <jmontleo@redhat.com>
de1f86c to
224883c
Compare
|
Rebased once again on top of the updated #777. Squashed some patches (with agreement from the authors) and fixed VPATH builds. |
This is part of an attempt to bring riscv64 support into shim.
See #420 and #641 for previous discussion, as well as #777 which is a prerequisite of this PR.
Using this branch I was able to successfully build shim on both x86_64 and riscv64, and I was able to use the resulting binaries to boot Fedora 42 cloud images for both architectures, replacing the stock ones.