| Message ID | 20230329081932.79831-4-alexghiti@rivosinc.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | 1b50f956c8fe9082bdee4a9cfd798149c52f7043 |
| Headers | show |
| Series | Fixes for dtb mapping | expand |
| Context | Check | Description |
|---|---|---|
| conchuod/cover_letter | success | Series has a cover letter |
| conchuod/tree_selection | success | Guessed tree name to be fixes at HEAD e89c2e815e76 |
| conchuod/fixes_present | success | Fixes tag present in non-next series |
| conchuod/maintainers_pattern | success | MAINTAINERS pattern errors before the patch: 1 and now 1 |
| conchuod/verify_signedoff | success | Signed-off-by tag matches author and committer |
| conchuod/kdoc | success | Errors and warnings before: 0 this patch: 0 |
| conchuod/build_rv64_clang_allmodconfig | success | Errors and warnings before: 18 this patch: 18 |
| conchuod/module_param | success | Was 0 now: 0 |
| conchuod/build_rv64_gcc_allmodconfig | success | Errors and warnings before: 19 this patch: 19 |
| conchuod/build_rv32_defconfig | success | Build OK |
| conchuod/dtb_warn_rv64 | success | Errors and warnings before: 3 this patch: 3 |
| conchuod/header_inline | success | No static functions without inline keyword in header files |
| conchuod/checkpatch | success | total: 0 errors, 0 warnings, 0 checks, 27 lines checked |
| conchuod/source_inline | success | Was 0 now: 0 |
| conchuod/build_rv64_nommu_k210_defconfig | success | Build OK |
| conchuod/verify_fixes | success | No Fixes tag |
| conchuod/build_rv64_nommu_virt_defconfig | success | Build OK |
On Wed, Mar 29, 2023 at 10:19:32AM +0200, Alexandre Ghiti wrote: > We used to access the dtb via its linear mapping address but now that the > dtb early mapping was moved in the fixmap region, we can keep using this > address since it is present in swapper_pg_dir, and remove the dtb > relocation. > > Note that the relocation was wrong anyway since early_memremap() is > restricted to 256K whereas the maximum fdt size is 2MB. So, should this be marked as a fix, and backported along with 1/3? Either way, Reviewed-by: Conor Dooley <conor.dooley@microchip.com> And from v1 (although I didn't actually send it for idk what reason): Tested-by: Conor Dooley <conor.dooley@microchip.com> Thanks, Conor. > > Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com> > --- > arch/riscv/mm/init.c | 21 ++------------------- > 1 file changed, 2 insertions(+), 19 deletions(-) > > diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c > index fb78d6bbabae..0f14f4a8d179 100644 > --- a/arch/riscv/mm/init.c > +++ b/arch/riscv/mm/init.c > @@ -249,25 +249,8 @@ static void __init setup_bootmem(void) > * early_init_fdt_reserve_self() since __pa() does > * not work for DTB pointers that are fixmap addresses > */ > - if (!IS_ENABLED(CONFIG_BUILTIN_DTB)) { > - /* > - * In case the DTB is not located in a memory region we won't > - * be able to locate it later on via the linear mapping and > - * get a segfault when accessing it via __va(dtb_early_pa). > - * To avoid this situation copy DTB to a memory region. > - * Note that memblock_phys_alloc will also reserve DTB region. > - */ > - if (!memblock_is_memory(dtb_early_pa)) { > - size_t fdt_size = fdt_totalsize(dtb_early_va); > - phys_addr_t new_dtb_early_pa = memblock_phys_alloc(fdt_size, PAGE_SIZE); > - void *new_dtb_early_va = early_memremap(new_dtb_early_pa, fdt_size); > - > - memcpy(new_dtb_early_va, dtb_early_va, fdt_size); > - early_memunmap(new_dtb_early_va, fdt_size); > - _dtb_early_pa = new_dtb_early_pa; > - } else > - memblock_reserve(dtb_early_pa, fdt_totalsize(dtb_early_va)); > - } > + if (!IS_ENABLED(CONFIG_BUILTIN_DTB)) > + memblock_reserve(dtb_early_pa, fdt_totalsize(dtb_early_va)); > > dma_contiguous_reserve(dma32_phys_limit); > if (IS_ENABLED(CONFIG_64BIT)) > -- > 2.37.2 >
On Wed, Mar 29, 2023 at 3:56 PM Conor Dooley <conor@kernel.org> wrote: > > On Wed, Mar 29, 2023 at 10:19:32AM +0200, Alexandre Ghiti wrote: > > We used to access the dtb via its linear mapping address but now that the > > dtb early mapping was moved in the fixmap region, we can keep using this > > address since it is present in swapper_pg_dir, and remove the dtb > > relocation. > > > > Note that the relocation was wrong anyway since early_memremap() is > > restricted to 256K whereas the maximum fdt size is 2MB. > > So, should this be marked as a fix, and backported along with 1/3? Hmmm the whole series should be backported, it does not make sense to move the dtb mapping and keep accessing it using its linear mapping address since it could fail for the exact reason the relocation was implemented in the first place and the relocation is wrong. Maybe we should simply add a "Cc: stable@vger.kernel.org" on all the patches? > Either way, > Reviewed-by: Conor Dooley <conor.dooley@microchip.com> > > And from v1 (although I didn't actually send it for idk what reason): > Tested-by: Conor Dooley <conor.dooley@microchip.com> Thanks! > > Thanks, > Conor. > > > > > Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com> > > --- > > arch/riscv/mm/init.c | 21 ++------------------- > > 1 file changed, 2 insertions(+), 19 deletions(-) > > > > diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c > > index fb78d6bbabae..0f14f4a8d179 100644 > > --- a/arch/riscv/mm/init.c > > +++ b/arch/riscv/mm/init.c > > @@ -249,25 +249,8 @@ static void __init setup_bootmem(void) > > * early_init_fdt_reserve_self() since __pa() does > > * not work for DTB pointers that are fixmap addresses > > */ > > - if (!IS_ENABLED(CONFIG_BUILTIN_DTB)) { > > - /* > > - * In case the DTB is not located in a memory region we won't > > - * be able to locate it later on via the linear mapping and > > - * get a segfault when accessing it via __va(dtb_early_pa). > > - * To avoid this situation copy DTB to a memory region. > > - * Note that memblock_phys_alloc will also reserve DTB region. > > - */ > > - if (!memblock_is_memory(dtb_early_pa)) { > > - size_t fdt_size = fdt_totalsize(dtb_early_va); > > - phys_addr_t new_dtb_early_pa = memblock_phys_alloc(fdt_size, PAGE_SIZE); > > - void *new_dtb_early_va = early_memremap(new_dtb_early_pa, fdt_size); > > - > > - memcpy(new_dtb_early_va, dtb_early_va, fdt_size); > > - early_memunmap(new_dtb_early_va, fdt_size); > > - _dtb_early_pa = new_dtb_early_pa; > > - } else > > - memblock_reserve(dtb_early_pa, fdt_totalsize(dtb_early_va)); > > - } > > + if (!IS_ENABLED(CONFIG_BUILTIN_DTB)) > > + memblock_reserve(dtb_early_pa, fdt_totalsize(dtb_early_va)); > > > > dma_contiguous_reserve(dma32_phys_limit); > > if (IS_ENABLED(CONFIG_64BIT)) > > -- > > 2.37.2 > >
On Wed, Mar 29, 2023 at 04:40:18PM +0200, Alexandre Ghiti wrote: > On Wed, Mar 29, 2023 at 3:56 PM Conor Dooley <conor@kernel.org> wrote: > > > > On Wed, Mar 29, 2023 at 10:19:32AM +0200, Alexandre Ghiti wrote: > > > We used to access the dtb via its linear mapping address but now that the > > > dtb early mapping was moved in the fixmap region, we can keep using this > > > address since it is present in swapper_pg_dir, and remove the dtb > > > relocation. > > > > > > Note that the relocation was wrong anyway since early_memremap() is > > > restricted to 256K whereas the maximum fdt size is 2MB. > > > > So, should this be marked as a fix, and backported along with 1/3? > > Hmmm the whole series should be backported, it does not make sense to > move the dtb mapping and keep accessing it using its linear mapping > address since it could fail for the exact reason the relocation was > implemented in the first place and the relocation is wrong. > > Maybe we should simply add a "Cc: stable@vger.kernel.org" on all the patches? Yup, although hopefully Palmer can handle that if nothing else needs changing.
On Wed, 29 Mar 2023 08:33:45 PDT (-0700), Conor Dooley wrote: > On Wed, Mar 29, 2023 at 04:40:18PM +0200, Alexandre Ghiti wrote: >> On Wed, Mar 29, 2023 at 3:56 PM Conor Dooley <conor@kernel.org> wrote: >> > >> > On Wed, Mar 29, 2023 at 10:19:32AM +0200, Alexandre Ghiti wrote: >> > > We used to access the dtb via its linear mapping address but now that the >> > > dtb early mapping was moved in the fixmap region, we can keep using this >> > > address since it is present in swapper_pg_dir, and remove the dtb >> > > relocation. >> > > >> > > Note that the relocation was wrong anyway since early_memremap() is >> > > restricted to 256K whereas the maximum fdt size is 2MB. >> > >> > So, should this be marked as a fix, and backported along with 1/3? >> >> Hmmm the whole series should be backported, it does not make sense to >> move the dtb mapping and keep accessing it using its linear mapping >> address since it could fail for the exact reason the relocation was >> implemented in the first place and the relocation is wrong. >> >> Maybe we should simply add a "Cc: stable@vger.kernel.org" on all the patches? > > Yup, although hopefully Palmer can handle that if nothing else needs > changing. The bots usually pick it up anyway, but in general I try and put the stable CC on there when I see stuff go by. I actually missed it this time until seeing the comment, though...
diff --git a/arch/riscv/mm/init.c b/arch/riscv/mm/init.c index fb78d6bbabae..0f14f4a8d179 100644 --- a/arch/riscv/mm/init.c +++ b/arch/riscv/mm/init.c @@ -249,25 +249,8 @@ static void __init setup_bootmem(void) * early_init_fdt_reserve_self() since __pa() does * not work for DTB pointers that are fixmap addresses */ - if (!IS_ENABLED(CONFIG_BUILTIN_DTB)) { - /* - * In case the DTB is not located in a memory region we won't - * be able to locate it later on via the linear mapping and - * get a segfault when accessing it via __va(dtb_early_pa). - * To avoid this situation copy DTB to a memory region. - * Note that memblock_phys_alloc will also reserve DTB region. - */ - if (!memblock_is_memory(dtb_early_pa)) { - size_t fdt_size = fdt_totalsize(dtb_early_va); - phys_addr_t new_dtb_early_pa = memblock_phys_alloc(fdt_size, PAGE_SIZE); - void *new_dtb_early_va = early_memremap(new_dtb_early_pa, fdt_size); - - memcpy(new_dtb_early_va, dtb_early_va, fdt_size); - early_memunmap(new_dtb_early_va, fdt_size); - _dtb_early_pa = new_dtb_early_pa; - } else - memblock_reserve(dtb_early_pa, fdt_totalsize(dtb_early_va)); - } + if (!IS_ENABLED(CONFIG_BUILTIN_DTB)) + memblock_reserve(dtb_early_pa, fdt_totalsize(dtb_early_va)); dma_contiguous_reserve(dma32_phys_limit); if (IS_ENABLED(CONFIG_64BIT))
We used to access the dtb via its linear mapping address but now that the dtb early mapping was moved in the fixmap region, we can keep using this address since it is present in swapper_pg_dir, and remove the dtb relocation. Note that the relocation was wrong anyway since early_memremap() is restricted to 256K whereas the maximum fdt size is 2MB. Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com> --- arch/riscv/mm/init.c | 21 ++------------------- 1 file changed, 2 insertions(+), 19 deletions(-)