| Message ID | 20200625234516.31406-1-atish.patra@wdc.com |
|---|---|
| Headers | show |
| Series | Add UEFI support for RISC-V | expand |
On Thu, Jun 25, 2020 at 7:43 PM Heinrich Schuchardt <xypron.glpk@gmx.de> wrote: > > On 6/26/20 1:45 AM, Atish Patra wrote: > > 32bit gcc doesn't support modulo operation on 64 bit data. It results in > > a __umoddi3 error while building EFI for 32 bit. > > > > Use bitwise operations instead of modulo operations to fix the issue. > > > > Signed-off-by: Atish Patra <atish.patra@wdc.com> > > --- > > drivers/firmware/efi/libstub/alignedmem.c | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > diff --git a/drivers/firmware/efi/libstub/alignedmem.c b/drivers/firmware/efi/libstub/alignedmem.c > > index cc89c4d6196f..1de9878ddd3a 100644 > > --- a/drivers/firmware/efi/libstub/alignedmem.c > > +++ b/drivers/firmware/efi/libstub/alignedmem.c > > @@ -44,7 +44,7 @@ efi_status_t efi_allocate_pages_aligned(unsigned long size, unsigned long *addr, > > *addr = ALIGN((unsigned long)alloc_addr, align); > > > > if (slack > 0) { > > - int l = (alloc_addr % align) / EFI_PAGE_SIZE; > > + int l = (alloc_addr & (align - 1)) / EFI_PAGE_SIZE; > > Here you rely on the compiler to silently convert the division by > EFI_PAGE_SIZE into a right shift. Wouldn't it be cleaner to use > EFI_PAGE_SHIFT to explicitly avoid a dependency on __udivdi3()? > > slack = (align >> EFI_PAGE_SHIFT) - 1; > ... > int l = (alloc_addr & (align - 1)) >> EFI_PAGE_SHIFT; > Sure. I will update it in the next version. Thanks for the suggestion. > Best regards > > Heinrich > > > > > if (l) { > > efi_bs_call(free_pages, alloc_addr, slack - l + 1); > > >
On Fri, 26 Jun 2020 at 23:56, Atish Patra <atishp@atishpatra.org> wrote: > > On Thu, Jun 25, 2020 at 7:43 PM Heinrich Schuchardt <xypron.glpk@gmx.de> wrote: > > > > On 6/26/20 1:45 AM, Atish Patra wrote: > > > 32bit gcc doesn't support modulo operation on 64 bit data. It results in > > > a __umoddi3 error while building EFI for 32 bit. > > > > > > Use bitwise operations instead of modulo operations to fix the issue. > > > > > > Signed-off-by: Atish Patra <atish.patra@wdc.com> > > > --- > > > drivers/firmware/efi/libstub/alignedmem.c | 2 +- > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > diff --git a/drivers/firmware/efi/libstub/alignedmem.c b/drivers/firmware/efi/libstub/alignedmem.c > > > index cc89c4d6196f..1de9878ddd3a 100644 > > > --- a/drivers/firmware/efi/libstub/alignedmem.c > > > +++ b/drivers/firmware/efi/libstub/alignedmem.c > > > @@ -44,7 +44,7 @@ efi_status_t efi_allocate_pages_aligned(unsigned long size, unsigned long *addr, > > > *addr = ALIGN((unsigned long)alloc_addr, align); > > > > > > if (slack > 0) { > > > - int l = (alloc_addr % align) / EFI_PAGE_SIZE; > > > + int l = (alloc_addr & (align - 1)) / EFI_PAGE_SIZE; > > > > Here you rely on the compiler to silently convert the division by > > EFI_PAGE_SIZE into a right shift. Wouldn't it be cleaner to use > > EFI_PAGE_SHIFT to explicitly avoid a dependency on __udivdi3()? > > > > slack = (align >> EFI_PAGE_SHIFT) - 1; > > ... > > int l = (alloc_addr & (align - 1)) >> EFI_PAGE_SHIFT; > > > > Sure. I will update it in the next version. Thanks for the suggestion. Please don't. Dividing by EFI_PAGE_SIZE is perfectly fine, and is more readable.
On Sat, 27 Jun 2020 at 00:03, Ard Biesheuvel <ardb@kernel.org> wrote: > > On Fri, 26 Jun 2020 at 23:56, Atish Patra <atishp@atishpatra.org> wrote: > > > > On Thu, Jun 25, 2020 at 7:43 PM Heinrich Schuchardt <xypron.glpk@gmx.de> wrote: > > > > > > On 6/26/20 1:45 AM, Atish Patra wrote: > > > > 32bit gcc doesn't support modulo operation on 64 bit data. It results in > > > > a __umoddi3 error while building EFI for 32 bit. > > > > > > > > Use bitwise operations instead of modulo operations to fix the issue. > > > > > > > > Signed-off-by: Atish Patra <atish.patra@wdc.com> > > > > --- > > > > drivers/firmware/efi/libstub/alignedmem.c | 2 +- > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > diff --git a/drivers/firmware/efi/libstub/alignedmem.c b/drivers/firmware/efi/libstub/alignedmem.c > > > > index cc89c4d6196f..1de9878ddd3a 100644 > > > > --- a/drivers/firmware/efi/libstub/alignedmem.c > > > > +++ b/drivers/firmware/efi/libstub/alignedmem.c > > > > @@ -44,7 +44,7 @@ efi_status_t efi_allocate_pages_aligned(unsigned long size, unsigned long *addr, > > > > *addr = ALIGN((unsigned long)alloc_addr, align); > > > > > > > > if (slack > 0) { > > > > - int l = (alloc_addr % align) / EFI_PAGE_SIZE; > > > > + int l = (alloc_addr & (align - 1)) / EFI_PAGE_SIZE; > > > > > > Here you rely on the compiler to silently convert the division by > > > EFI_PAGE_SIZE into a right shift. Wouldn't it be cleaner to use > > > EFI_PAGE_SHIFT to explicitly avoid a dependency on __udivdi3()? > > > > > > slack = (align >> EFI_PAGE_SHIFT) - 1; > > > ... > > > int l = (alloc_addr & (align - 1)) >> EFI_PAGE_SHIFT; > > > > > > > Sure. I will update it in the next version. Thanks for the suggestion. > > Please don't. Dividing by EFI_PAGE_SIZE is perfectly fine, and is more readable. Actually, I will take this patch as a fix right away - 32-bit ARM has the same issue (although it does not actually incorporate this object file) In the meantime, please check how the stub gets pulled into your kernel: the idea of a static library is that only the objects that are used are included in the final build, but this turned out to be broken for arm64 [0]. IOW, simply applying a similar change might fix your build issue as well (assuming RISC-V does not actually call efi_allocate_pages_aligned() anywhere) [0] https://lore.kernel.org/linux-arm-kernel/20200604022031.164207-1-masahiroy@kernel.org/
On Sat, Jun 27, 2020 at 2:22 AM Ard Biesheuvel <ardb@kernel.org> wrote: > > On Sat, 27 Jun 2020 at 00:03, Ard Biesheuvel <ardb@kernel.org> wrote: > > > > On Fri, 26 Jun 2020 at 23:56, Atish Patra <atishp@atishpatra.org> wrote: > > > > > > On Thu, Jun 25, 2020 at 7:43 PM Heinrich Schuchardt <xypron.glpk@gmx.de> wrote: > > > > > > > > On 6/26/20 1:45 AM, Atish Patra wrote: > > > > > 32bit gcc doesn't support modulo operation on 64 bit data. It results in > > > > > a __umoddi3 error while building EFI for 32 bit. > > > > > > > > > > Use bitwise operations instead of modulo operations to fix the issue. > > > > > > > > > > Signed-off-by: Atish Patra <atish.patra@wdc.com> > > > > > --- > > > > > drivers/firmware/efi/libstub/alignedmem.c | 2 +- > > > > > 1 file changed, 1 insertion(+), 1 deletion(-) > > > > > > > > > > diff --git a/drivers/firmware/efi/libstub/alignedmem.c b/drivers/firmware/efi/libstub/alignedmem.c > > > > > index cc89c4d6196f..1de9878ddd3a 100644 > > > > > --- a/drivers/firmware/efi/libstub/alignedmem.c > > > > > +++ b/drivers/firmware/efi/libstub/alignedmem.c > > > > > @@ -44,7 +44,7 @@ efi_status_t efi_allocate_pages_aligned(unsigned long size, unsigned long *addr, > > > > > *addr = ALIGN((unsigned long)alloc_addr, align); > > > > > > > > > > if (slack > 0) { > > > > > - int l = (alloc_addr % align) / EFI_PAGE_SIZE; > > > > > + int l = (alloc_addr & (align - 1)) / EFI_PAGE_SIZE; > > > > > > > > Here you rely on the compiler to silently convert the division by > > > > EFI_PAGE_SIZE into a right shift. Wouldn't it be cleaner to use > > > > EFI_PAGE_SHIFT to explicitly avoid a dependency on __udivdi3()? > > > > > > > > slack = (align >> EFI_PAGE_SHIFT) - 1; > > > > ... > > > > int l = (alloc_addr & (align - 1)) >> EFI_PAGE_SHIFT; > > > > > > > > > > Sure. I will update it in the next version. Thanks for the suggestion. > > > > Please don't. Dividing by EFI_PAGE_SIZE is perfectly fine, and is more readable. > > Actually, I will take this patch as a fix right away - 32-bit ARM has > the same issue (although it does not actually incorporate this object > file) > Thanks. I will drop it from my next version in that case. > In the meantime, please check how the stub gets pulled into your > kernel: the idea of a static library is that only the objects that are > used are included in the final build, but this turned out to be broken > for arm64 [0]. IOW, simply applying a similar change might fix your > build issue as well (assuming RISC-V does not actually call > efi_allocate_pages_aligned() anywhere) > Sorry. I missed this thread earlier. Yes. Applying a similar change fixes the static library linking for RISC-V. I don't see efi_allocate_pages_aligned or efi_random_alloc in vmlinux anymore. I will update my patch. Thanks. I did not see any new version of the next patch in that series [1]. Thus, I am leaving libstub/Makefile changes as it is. Please let me know if I missed something. [1] https://lore.kernel.org/linux-efi/CAMj1kXGg-w6N7jyG0pBJmeRctAhG2wWGoU=ryWj+Qi2UH-_m9Q@mail.gmail.com/ > [0] https://lore.kernel.org/linux-arm-kernel/20200604022031.164207-1-masahiroy@kernel.org/
This series adds UEFI support for RISC-V. Linux kernel: 5.8-rc2 U-Boot: master OpenSBI: master Patch 1-6 are preparatory patches that fixes some of the geric efi and riscv issues. Patch 7-9 adds the efi stub support for RISC-V which was reviewed few months back. http://lists.infradead.org/pipermail/linux-riscv/2020-April/009586.html Patch 10 just renames arm-init code so that it can be used across different architectures. Patch 11 adds the runtime services for RISC-V. The patches can also be found in following git repo. https://github.com/atishp04/linux/tree/uefi_riscv_5.9_v1 The patches have been verified on Qemu using bootefi command in U-Boot for both RV32 and RV64. For RV32, maximum allocated memory should be 1G as RISC-V kernel can not map beyond 1G of physical memory for RV32. EDK2 can boot quite far into Linux with current series. Currently, we are seeing some traps from drivers (spi/network). At first glance, they don't seem to be caused by efi. I thought it is better to get some early feedback on the series while EDK2 issue is being debugged. That's why uefi runtime services are not actually well tested in RISC-V. Any suggestions to test the efi run time services are appreciated. Changes from previous version: 1. Added full ioremap support. 2. Added efi runtime services support. 3. Fixes mm issues Anup Patel (1): RISC-V: Move DT mapping outof fixmap Atish Patra (10): efi: Fix gcc error around __umoddi3 for 32 bit builds RISC-V: Setup exception vector early RISC-V: Add early ioremap support RISC-V: Set maximum number of mapped pages correctly riscv: Parse all memory blocks to remove unusable memory include: pe.h: Add RISC-V related PE definition RISC-V: Add PE/COFF header for EFI stub RISC-V: Add EFI stub support. efi: Rename arm-init to efi-init common for all arch RISC-V: Add EFI runtime services arch/riscv/Kconfig | 25 ++++ arch/riscv/Makefile | 1 + arch/riscv/configs/defconfig | 1 + arch/riscv/include/asm/Kbuild | 1 + arch/riscv/include/asm/efi.h | 56 +++++++ arch/riscv/include/asm/fixmap.h | 16 +- arch/riscv/include/asm/io.h | 1 + arch/riscv/include/asm/mmu.h | 2 + arch/riscv/include/asm/pgalloc.h | 12 ++ arch/riscv/include/asm/pgtable.h | 4 + arch/riscv/include/asm/sections.h | 13 ++ arch/riscv/kernel/Makefile | 5 + arch/riscv/kernel/efi-header.S | 104 +++++++++++++ arch/riscv/kernel/efi.c | 106 +++++++++++++ arch/riscv/kernel/head.S | 27 +++- arch/riscv/kernel/head.h | 2 - arch/riscv/kernel/image-vars.h | 51 +++++++ arch/riscv/kernel/setup.c | 16 +- arch/riscv/kernel/smpboot.c | 1 - arch/riscv/kernel/traps.c | 8 +- arch/riscv/kernel/vmlinux.lds.S | 22 ++- arch/riscv/mm/init.c | 104 ++++++++----- drivers/firmware/efi/Kconfig | 3 +- drivers/firmware/efi/Makefile | 4 +- .../firmware/efi/{arm-init.c => efi-init.c} | 0 drivers/firmware/efi/libstub/Makefile | 10 ++ drivers/firmware/efi/libstub/alignedmem.c | 2 +- drivers/firmware/efi/libstub/efi-stub.c | 11 +- drivers/firmware/efi/libstub/riscv-stub.c | 110 ++++++++++++++ drivers/firmware/efi/riscv-runtime.c | 141 ++++++++++++++++++ include/linux/pe.h | 3 + 31 files changed, 796 insertions(+), 66 deletions(-) create mode 100644 arch/riscv/include/asm/efi.h create mode 100644 arch/riscv/include/asm/sections.h create mode 100644 arch/riscv/kernel/efi-header.S create mode 100644 arch/riscv/kernel/efi.c create mode 100644 arch/riscv/kernel/image-vars.h rename drivers/firmware/efi/{arm-init.c => efi-init.c} (100%) create mode 100644 drivers/firmware/efi/libstub/riscv-stub.c create mode 100644 drivers/firmware/efi/riscv-runtime.c -- 2.24.0