On Thu, Jul 02, 2026 at 02:55:22PM -0600, Tom Rini wrote: > On Thu, Jul 02, 2026 at 08:39:44PM +0000, Yao Zi wrote: > > On Wed, Jul 01, 2026 at 11:39:40AM -0600, Tom Rini wrote: > > > On Wed, Jul 01, 2026 at 11:17:55AM +0000, Yao Zi wrote: > > > > > > > From: Jiaxun Yang <[email protected]> > > > > > > > > For LoongArch the start of the image is not the entry > > > > point to the image. > > > > > > > > We refactor the code base to allow entry point to be > > > > supplied by setup_booti. > > > > > > > > Signed-off-by: Jiaxun Yang <[email protected]> > > > > Signed-off-by: Yao Zi <[email protected]> > > > > --- > > > > > > > > Changed from v1 > > > > - Correct type of "entry" argument for RISC-V and sandbox > > > > > > Can you please also provide a link to > > > https://www.kernel.org/doc/html/latest/arch/loongarch/booting.html which > > > probably should cover this, but I'm not entirely sure does? > > > > It does in a non-obvious way, quoting part of the image header, > > > > u32 MZ_MAGIC /* "MZ", MS-DOS header */ > > u32 res0 = 0 /* Reserved */ > > u64 kernel_entry /* Kernel entry point */ > > > > comparing it to the RISC-V version[1], > > > > u32 code0; /* Executable code */ > > u32 code1; /* Executable code */ > > u64 text_offset; /* Image load offset, little endian */ > > u64 image_size; /* Effective Image size, little endian */ > > > > The start of LoongArch image isn't marked as executable. > > Perhaps it's worth clarifying in the kernel documentation then? Thanks.
Sure, Jiaxun and I plan to improve the kernel side documentation, and additionally improve LoongArch image header to make the image executable from its start, aligning with aarch64 and riscv64. However, there isn't a promise when it would land, and even with the improvement to image header, we still have to parse the text_offset field to keep compatibility with older images. > > Additionally, should I include the URL of "latest" kernel documentation, > > or using a link with explicit kernel version (e.g. [2]) to ensure the > > content doesn't change or become 404 over time? > > Ah, true, yes, lets point at the v7.1 tag, or whatever you're testing > this against, to cover future changes if any. > > -- > Tom Best regards, Yao Zi
