On Fri, 2 Aug 2019 at 17:07, Ortiz, Baltazar <[email protected]> wrote: > > Quick update - I had to build the toolchain with the "--with-arch=rv32gc > --with-abi=ilp32" flags. seL4test now runs and passes all but the FPU tests. > The seL4 RISC-V port doesn't support FPU yet. You will need "--with-arch=rv32imac --with-abi=ilp32" or --enable-multilib for the FPU tests not to generate FPU instructions.
> -----Original Message----- > From: Ortiz, Baltazar > Sent: Wednesday, July 31, 2019 2:01 PM > To: 'Zhuang, Siwei (Data61, Kensington NSW)' <[email protected]>; > [email protected] > Subject: RE: [seL4] SeL4 on 32 bit RISC-V > > Thanks for the response! I tried using the latest sel4test and got the same > assertion failure. I once again had to remove the two CONST labels from > thread.h to fix the below error. > > > [196/214] Building C object > >kernel/CMakeFiles/kernel.elf.dir/kernel_all.i.obj > > FAILED: kernel/CMakeFiles/kernel.elf.dir/kernel_all.i.obj > > ccache /home/baltazar/riscv/bin/riscv32-unknown-linux-gnu-gcc > > --sysroot=/host/sel4test-upstream-riscv/build32 -I../kernel/include > > -I../kernel/include/32 -I../kernel/include/arch/riscv > > -I../kernel/include/arch/riscv/arch/32 -I../kernel/include/plat/spike > > -I../kernel/include/plat/spike/plat/32 -Ikernel/gen_config > > -Ikernel/autoconf -Ikernel/gen_headers -Ikernel/generated -march=rv32ima > > -mabi=ilp32 -D__KERNEL_32__ -nostdinc -nostdlib -O2 -DHAVE_AUTOCONF > > -DDEBUG -g -ggdb -mcmodel=medany -fno-pic -fno-pie -fno-stack-protector > > -fno-asynchronous-unwind-tables -std=c99 -Wall -Werror -Wstrict-prototypes > > -> Wmissing-prototypes -Wnested-externs -Wmissing-declarations -Wundef > > -Wpointer-arith -Wno-nonnull -ffreestanding -MD -MT > > kernel/CMakeFiles/kernel.elf.dir/kernel_all.i.obj -MF > > kernel/CMakeFiles/kernel.elf.dir/kernel_all.i.obj.d -o > > kernel/CMakeFiles/kernel.elf.dir/kernel_all.i.obj -c kernel/kernel_all.i > > In file included from > > /host/sel4test-upstream-riscv/kernel/src/api/faults.c:18: > >../kernel/include/arch/riscv/arch/kernel/thread.h:27:1: error: 'const' > >attribute on function returning 'void' [-Werror=attributes] void CONST > >Arch_activateIdleThread(tcb_t *tcb); > > ^~~~ > > I'm going to build the 64 bit version of the toolchain and see if that works > any better than the 32 bit. Worst case, I'll give the obsolete version of the > toolchain another shot, though I had trouble getting it set up in the past. > > Thanks, > > Baltazar > > -----Original Message----- > From: Devel <[email protected]> On Behalf Of Zhuang, Siwei (Data61, > Kensington NSW) > Sent: Monday, July 29, 2019 10:28 PM > To: [email protected] > Subject: Re: [seL4] SeL4 on 32 bit RISC-V > > Hi Baltazar, > > To get that version of seL4 working, you should use an obsolete version of > the RISC-V toolchain. Follow the instructions here, > > https://docs.sel4.systems/Hardware/RISCV.html > > If you prefer to use the upstream riscv-gnu-toolchain, you need to checkout > the latest sel4test-manifest. > > In either case, modifying the seL4 source code shouldn't be required. > > Regards, > Siwei > > On Mon, 2019-07-29 at 21:36 +0000, Ortiz, Baltazar wrote: > > Hi all, > > > > I'm looking into getting seL4 booting on 32 bit RISC-V. Currently > > working with seL4test 10.1.1, QEMU 4.0.0, and an upstream copy of > > riscv-gnu- toolchain (is there a specific release I should try?). > > > > I'm able to get seL4 to build after deleting a few CONST labels that > > the compiler was complaining about, but when I try to run the > > simulation script, I get the following output: > > > > > ./simulate -b > > > /home/baltazar/qemu/build/riscv32-softmmu/qemu-system- > > > riscv32 > > > /home/baltazar/qemu/build/riscv32-softmmu/qemu-system-riscv32 > > > -machine > > > spike_v1.10 -nographic -s -serial mon:stdio -m size=2000M -kernel > > > images/sel4test-driver-image-riscv-spike > > > bbl loader > > > /host/sel4test-upstream-riscv/tools/riscv-pk/machine/minit.c:74: > > > assertion > > > failed: !(read_csr(misa) & fd_mask) > > > Power off > > > > Are there any tricks for getting this working right now? > > > > Thanks, > > > > Baltazar > > ________________________________ > > Notice: This email and any attachments may contain proprietary (Draper > > non- > > public) and/or export-controlled information of Draper. If you are not > > the intended recipient of this email, please immediately notify the > > sender by replying to this email and immediately destroy all copies of this > > email. > > ________________________________ > > _______________________________________________ > > Devel mailing list > > [email protected] > > https://sel4.systems/lists/listinfo/devel > _______________________________________________ > Devel mailing list > [email protected] > https://sel4.systems/lists/listinfo/devel > ________________________________ > Notice: This email and any attachments may contain proprietary (Draper > non-public) and/or export-controlled information of Draper. If you are not > the intended recipient of this email, please immediately notify the sender by > replying to this email and immediately destroy all copies of this email. > ________________________________ > _______________________________________________ > Devel mailing list > [email protected] > https://sel4.systems/lists/listinfo/devel -- Hesham _______________________________________________ Devel mailing list [email protected] https://sel4.systems/lists/listinfo/devel
