On Sat, Feb 01, 2025 at 07:03:33AM -0800, Guenter Roeck wrote: > On 2/1/25 00:01, Greg Kroah-Hartman wrote: > ... > > Anyway, are you all really caring about riscv on a 5.4.y kernel? Last I > > checked, the riscv maintainers said not to even use that kernel for that > > architecture. Do you all have real boards that care about this kernel > > tree that you are insisting on keeping alive? Why not move them to a > > newer LTS kernel? > > > > Looking into the 5.4 release candidate, I see: > > $ git log --oneline v5.4.289.. arch/riscv/ > 98d26e0254ff RISC-V: Don't enable all interrupts in trap_init() > 574c5efceb70 riscv: prefix IRQ_ macro names with an RV_ namespace > c57ffe372502 riscv: Fix sleeping in invalid context in die() > 98c62ee8bc75 riscv: Avoid enabling interrupts in die() > 88cb873873ff RISC-V: Avoid dereferening NULL regs in die() > 2a83ad25311e riscv: remove unused handle_exception symbol > 8652d51931cc riscv: abstract out CSR names for supervisor vs machine mode I've dropped them all now, as that is what was causing the build problems. > Why do you backport riscv patches to 5.4.y if you think they should not be > tested ? Shouldn't your question imply that there won't be any further > backports into 5.4.y for architecture(s) which are no longer supported > in that branch ? I'm not implying they are not to be tested, it's just a real "is this something that people actually care about" question. Last time we had riscv problems in this branch the riscv maintainers said "don't worry about it". I didn't notice that Sasha had queued these up here, otherwise I would have probably just dropped them then like I did right now :) thanks, greg k-h
