On Tue, May 20, 2025 at 3:08 PM Michael Tokarev <m...@tls.msk.ru> wrote: > > On 19.05.2025 07:04, alistai...@gmail.com wrote: > > From: Alistair Francis <alistair.fran...@wdc.com> > > > First RISC-V PR for 10.1 > > > > * Add support for RIMT to virt machine ACPI > > * Don't allow PMP RLB to bypass rule privileges > > * Fix checks on writes to pmpcfg in Smepmp MML mode > > * Generate strided vector loads/stores with tcg nodes > > * Improve Microchip Polarfire SoC customization > > * Use tcg ops generation to emulate whole reg rvv loads/stores > > * Expand the probe_pages helper function to handle probe flags > > * Fix type conflict of GLib function pointers > > * Fix endless translation loop on big endian systems > > * Use tail pseudoinstruction for calling tail > > * Fix some RISC-V vector instruction corner cases > > * MAINTAINERS: Add common-user/host/riscv to RISC-V section > > * Fix write_misa vs aligned next_pc > > * KVM CSR fixes > > * Virt machine memmap usage cleanup > > Hi! > > Alistar, there are multiple patches in this pull request which > are tagged for stable. Many of them (if not all) does not have > a Fixes: tag, or any other indication where they're applicable.
Yeah, I tried to Cc stable if it was a fix. > Is there some hint I can use to decide which active stable release > to apply to, for each patch? Maybe we should apply most of them My initial thinking was to apply all of the patches tagged for stable if they apply cleanly. If the patch doesn't apply cleanly then just skip it. Don't bother trying to get it to apply. > just to the latest releases (current: 9.2, to be end-of-line > after the next release, and 10.0), or some of them should be > applied for older release (current: 7.2)? That is also fine with me. Most of the fixes are for vector instructions, which is a fast moving target anyway, so just getting them in the latest release will be fine. Patches with a Fixes tag should be applied to branches with the relevant commits though. Alistair > > Thanks, > > /mjt
