kernel/drivers/of/fdt.c, branch linux-6.9.y

of: Move all FDT reserved-memory handling into of_reserved_mem.c

2024-03-12T15:23:26Z

The split of /reserved-memory handling between fdt.c and of_reserved_mem.c makes for reading and restructuring the code difficult. As of_reserved_mem.c is only built for CONFIG_OF_EARLY_FLATTREE already, move all the code to one spot. Acked-by: Saravana Kannan Reviewed-by: Oreoluwa Babatunde Link: https://lore.kernel.org/r/20240311181303.1516514-2-robh@kernel.org Signed-off-by: Rob Herring

of: Create of_root if no dtb provided by firmware

2024-03-08T18:50:39Z

When enabling CONFIG_OF on a platform where 'of_root' is not populated by firmware, we end up without a root node. In order to apply overlays and create subnodes of the root node, we need one. Create this root node by unflattening an empty builtin dtb. If firmware provides a flattened device tree (FDT) then the FDT is unflattened via setup_arch(). Otherwise, the call to unflatten(_and_copy)?_device_tree() will create an empty root node. We make of_have_populated_dt() return true only if the DTB was loaded by firmware so that existing callers don't change behavior after this patch. The call in the of platform code is removed because it prevents overlays from creating platform devices when the empty root node is used. [sboyd@kernel.org: Update of_have_populated_dt() to treat this empty dtb as not populated. Drop setup_of() initcall] Signed-off-by: Frank Rowand Link: https://lore.kernel.org/r/20230317053415.2254616-2-frowand.list@gmail.com Cc: Rob Herring Signed-off-by: Stephen Boyd Link: https://lore.kernel.org/r/20240217010557.2381548-3-sboyd@kernel.org Signed-off-by: Rob Herring

of: Always unflatten in unflatten_and_copy_device_tree()

2024-03-08T18:50:39Z

We want to populate an empty DT whenever CONFIG_OF is enabled so that overlays can be applied and the DT unit tests can be run. Make unflatten_and_copy_device_tree() stop printing a warning if the 'initial_boot_params' pointer is NULL. Instead, simply copy the dtb if there is one and then unflatten it. If there isn't a DT to copy, then the call to unflatten_device_tree() is largely a no-op, so nothing really changes here. Cc: Rob Herring Cc: Frank Rowand Signed-off-by: Stephen Boyd Link: https://lore.kernel.org/r/20240217010557.2381548-2-sboyd@kernel.org Signed-off-by: Rob Herring

Merge tag 'devicetree-fixes-for-6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux

2023-05-05T20:27:59Z

Pull devicetree fixes from Rob Herring: - Add Conor Dooley as a DT binding maintainer - Swap the order of parsing /memreserve/ and /reserved-memory nodes so that the /reserved-memory nodes which have more information are handled first - Fix some property dependencies in riscv,pmu binding - Update maintainers entries on a couple of bindings * tag 'devicetree-fixes-for-6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux: MAINTAINERS: add Conor as a dt-bindings maintainer dt-bindings: perf: riscv,pmu: fix property dependencies dt-bindings: xilinx: Remove Naga from memory and mtd bindings of: fdt: Scan /memreserve/ last dt-bindings: clock: r9a06g032-sysctrl: Change maintainer to Fabrizio Castro dt-bindings: pinctrl: renesas,rzv2m: Change maintainer to Fabrizio Castro dt-bindings: pinctrl: renesas,rzn1: Change maintainer to Fabrizio Castro dt-bindings: i2c: renesas,rzv2m: Change maintainer to Fabrizio Castro

of: fdt: Scan /memreserve/ last

2023-04-27T20:52:16Z

Change the scanning /memreserve/ and /reserved-memory node order to fix Kernel panic on Khadas Vim3 Board. If /memreserve/ goes first, the memory is reserved, but nomap can't be applied to the region. So the memory won't be used by Linux, but it is still present in the linear map as normal memory, which allows speculation. Legitimate access to adjacent pages will cause the CPU to end up prefetching into them leading to Kernel panic. So /reserved-memory node should go first, as it has a more updated description of the memory regions and can apply flags, like nomap. Link: https://lore.kernel.org/all/CAJX_Q+1Tjc+-TjZ6JW9X0NxEdFe=82a9626yL63j7uVD4LpxEA@mail.gmail.com/ Signed-off-by: Lucas Tanure Link: https://lore.kernel.org/r/20230424113846.46382-1-tanure@linux.com Signed-off-by: Rob Herring

riscv: Use PUD/P4D/PGD pages for the linear mapping

2023-04-19T03:43:04Z

During the early page table creation, we used to set the mapping for PAGE_OFFSET to the kernel load address: but the kernel load address is always offseted by PMD_SIZE which makes it impossible to use PUD/P4D/PGD pages as this physical address is not aligned on PUD/P4D/PGD size (whereas PAGE_OFFSET is). But actually we don't have to establish this mapping (ie set va_pa_offset) that early in the boot process because: - first, setup_vm installs a temporary kernel mapping and among other things, discovers the system memory, - then, setup_vm_final creates the final kernel mapping and takes advantage of the discovered system memory to create the linear mapping. During the first phase, we don't know the start of the system memory and then until the second phase is finished, we can't use the linear mapping at all and phys_to_virt/virt_to_phys translations must not be used because it would result in a different translation from the 'real' one once the final mapping is installed. So here we simply delay the initialization of va_pa_offset to after the system memory discovery. But to make sure noone uses the linear mapping before, we add some guard in the DEBUG_VIRTUAL config. Finally we can use PUD/P4D/PGD hugepages when possible, which will result in a better TLB utilization. Note that: - this does not apply to rv32 as the kernel mapping lies in the linear mapping. - we rely on the firmware to protect itself using PMP. Signed-off-by: Alexandre Ghiti Acked-by: Rob Herring # DT bits Reviewed-by: Andrew Jones Reviewed-by: Anup Patel Tested-by: Anup Patel Link: https://lore.kernel.org/r/20230324155421.271544-4-alexghiti@rivosinc.com Signed-off-by: Palmer Dabbelt

Revert "mm: kmemleak: alloc gray object for reserved region with direct map"

2023-02-01T00:44:09Z

This reverts commit 972fa3a7c17c9d60212e32ecc0205dc585b1e769. Kmemleak operates by periodically scanning memory regions for pointers to allocated memory blocks to determine if they are leaked or not. However, reserved memory regions can be used for DMA transactions between a device and a CPU, and thus, wouldn't contain pointers to allocated memory blocks, making them inappropriate for kmemleak to scan. Thus, revert this commit. Link: https://lkml.kernel.org/r/20230124230254.295589-1-isaacmanjarres@google.com Fixes: 972fa3a7c17c9 ("mm: kmemleak: alloc gray object for reserved region with direct map") Signed-off-by: Isaac J. Manjarres Acked-by: Catalin Marinas Cc: Calvin Zhang Cc: Frank Rowand Cc: Rob Herring Cc: Saravana Kannan Cc: [5.17+] Signed-off-by: Andrew Morton

of: fdt: Honor CONFIG_CMDLINE* even without /chosen node, take 2

2023-01-05T03:31:59Z

I do not read a strict requirement on /chosen node in either ePAPR or in Documentation/devicetree. Help text for CONFIG_CMDLINE and CONFIG_CMDLINE_EXTEND doesn't make their behavior explicitly dependent on the presence of /chosen or the presense of /chosen/bootargs. However the early check for /chosen and bailing out in early_init_dt_scan_chosen() skips CONFIG_CMDLINE handling which is not really related to /chosen node or the particular method of passing cmdline from bootloader. This leads to counterintuitive combinations (assuming CONFIG_CMDLINE_EXTEND=y): a) bootargs="foo", CONFIG_CMDLINE="bar" => cmdline=="foo bar" b) /chosen missing, CONFIG_CMDLINE="bar" => cmdline=="" c) bootargs="", CONFIG_CMDLINE="bar" => cmdline==" bar" Rework early_init_dt_scan_chosen() so that the cmdline config options are always handled. [commit msg written by Alexander Sverdlin] Cc: Alexander Sverdlin Cc: Linus Walleij Cc: Arnd Bergmann Tested-by: Geoff Levand Reviewed-by: Alexander Sverdlin Link: https://lore.kernel.org/r/20230103-dt-cmdline-fix-v1-2-7038e88b18b6@kernel.org Signed-off-by: Rob Herring

Revert "of: fdt: Honor CONFIG_CMDLINE* even without /chosen node"

2023-01-05T03:31:59Z

This reverts commit a7d550f82b445cf218b47a2c1a9c56e97ecb8c7a. Some arches (PPC at least) don't call early_init_dt_scan_nodes(), so moving the cmdline processing there breaks them. Reported-by: Geoff Levand Cc: Alexander Sverdlin Tested-by: Geoff Levand Reviewed-by: Alexander Sverdlin Link: https://lore.kernel.org/r/20230103-dt-cmdline-fix-v1-1-7038e88b18b6@kernel.org Signed-off-by: Rob Herring

of/fdt: run soc memory setup when early_init_dt_scan_memory fails

2023-01-04T00:41:01Z

If memory has been found early_init_dt_scan_memory now returns 1. If it hasn't found any memory it will return 0, allowing other memory setup mechanisms to carry on. Previously early_init_dt_scan_memory always returned 0 without distinguishing between any kind of memory setup being done or not. Any code path after the early_init_dt_scan memory call in the ramips plat_mem_setup code wouldn't be executed anymore. Making early_init_dt_scan_memory the only way to initialize the memory. Some boards, including my mt7621 based Cudy X6 board, depend on memory initialization being done via the soc_info.mem_detect function pointer. Those wouldn't be able to obtain memory and panic the kernel during early bootup with the message "early_init_dt_alloc_memory_arch: Failed to allocate 12416 bytes align=0x40". Fixes: 1f012283e936 ("of/fdt: Rework early_init_dt_scan_memory() to call directly") Cc: stable@vger.kernel.org Signed-off-by: Andreas Rammhold Link: https://lore.kernel.org/r/20221223112748.2935235-1-andreas@rammhold.de Signed-off-by: Rob Herring