On Thu, Jul 24, 2025 at 02:10:05PM +0200, Michal Simek wrote: > Hi, > > I am sending you another series of patches for rc1. Gitlab CI is not showing > any issue. The biggest is FPGA series and there is on patch revert in spi > which should get to tree. > I have some other patches pending but I want to give others more time for > review that's why will apply them in future. > > Thanks, > Michal > > > The following changes since commit 3532f1f5edfc97c9dcea723cdeb732eda44bc669: > > block: Remove blk_find_first/next (2025-07-23 17:37:14 -0600) > > are available in the Git repository at: > > https://source.denx.de/u-boot/custodians/u-boot-microblaze.git > tags/xilinx-for-v2025.10-rc1-v2 > > for you to fetch changes up to 27fa1e2c1794cc589d72d485488c0bbb3b877f09: > > firmware: zynqmp: Simplify power-domain driver bind (2025-07-24 09:06:52 > +0200) > > ---------------------------------------------------------------- > AMD/Xilinx/FPGA changes for v2025.10-rc1 v2 > > zynqmp: > - Generate fit-dtb.blob all the time > - Simplify power-domain driver bind > > zynqmp_mini: > - Remove PSCI_RESET > > fpga: > - Improve user feedback in case of FPGA bitstream load failure > > misc: > - Fix kernel-doc in gpio zynq and axi_mrmac > > spi: > - Revert fix in STIG mode > > ---------------------------------------------------------------- > Michal Simek (3): > xilinx: zynqmp: Remove PSCI_RESET from mini configurations > arm64: zynqmp: Let fit-dtb.blob generated all the time > firmware: zynqmp: Simplify power-domain driver bind > > Pieter Van Trappen (6): > cmd: fpga: improve user feedback in case of bitstream load failure > drivers: fpga: add Kconfig dependency > drivers: fpga: correct compiler errors and warnings > drivers: fpga: cleanup printf usage > drivers: fpga: fix function declaration without a prototype > configs: enable FPGA commands and drivers at sandbox_defconfig
So here's a real fun problem that for now I've just worked around by disabling CONFIG_FPGA_VERSALPL=y on sandbox. One of the tests that's in Azure but not Gitlab is building sandbox with clang with ASAN (AddressSanitizer) support enabled. And only in that configuration do we see this build failure: /usr/bin/ld: /tmp/lto-llvm-3b170f.o: in function `versal_load': drivers/fpga/versalpl.c:(.text.versal_load+0x11d): undefined reference to `xilinx_pm_request' clang-18: error: linker command failed with exit code 1 (use -v to see invocation) This function is defined in drivers/firmware/firmware-zynqmp.c (requires CONFIG_ZYNQMP_FIRMWARE, which can't build on sandbox) and a weak do-nothing in drivers/mmc/zynq_sdhci.c (can't build on sandbox). I don't know why clang trips up on this, but gcc does not. But in the end, now that this series has been applied I'll do a further "Tighten up..." patch on drivers/fpga/Kconfig and update drivers/fpga/versalpl.c and drivers/fpga/zynqmppl.c to depend on ZYNQMP_FIRMWARE. I'm testing the rest of this PR still, I just wanted to note this part now while I've dug it out. -- Tom
signature.asc
Description: PGP signature
