Hi Andrew On Fri, Jun 14, 2024 at 10:31 AM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote: > > On Fri, 14 Jun 2024 07:32:07 -0700 Jeff Xu <jeffxu@xxxxxxxxxxxx> wrote: > > > Hi Andrew, > > > > By fixing the document, the usage of MFD_NOEXEC_SEAL shall be clear. > > Was the previous patch (which changed ABI) removed from unstable ? > > > > I pulled the mm-unstable this morning, it seems that patch is still here. > > I'm not sure about the process of reverting it, hence asking. > > > > c1e11be9abae86ebe5cecc42abc412f61ae563c6 > > memfd: `MFD_NOEXEC_SEAL` should not imply `MFD_ALLOW_SEALING` > > > > I'm not completely sure what you're asking here. > > Please try not to use terms like "the previous patch". There are a lot > of patches! > > I *think* you're saying that the patch "mm/memfd: add documentation for > MFD_NOEXEC_SEAL MFD_EXEC" (and its two fixups) are tied to the patch > "memfd: `MFD_NOEXEC_SEAL` should not imply `MFD_ALLOW_SEALING`". > > If so then this escaped me and I shall rearrange the grouping to be > more logical. > > The base patch "memfd: `MFD_NOEXEC_SEAL` should not imply `MFD_ALLOW_SEALING`" is in mm-unstable but I need to revisit the discussion to see whether we should be proceeding with it. Any guidance you can offer here will be helpful. > > Apology for the confusion. The patch memfd: `MFD_NOEXEC_SEAL` should not imply `MFD_ALLOW_SEALING`" changes ABI, and I think it is risky and should be removed. The main discussion is in: [1] There is an additional discussion when I sent v1 of the current patch [2] In short, current patch obsoletes the patch: memfd: `MFD_NOEXEC_SEAL` should not imply `MFD_ALLOW_SEALING` [1] https://lore.kernel.org/lkml/CABi2SkXyHegwpZKF7ZVDAOsqzaHE4BdDQOMmbuJ+51HkQ3CmJg@xxxxxxxxxxxxxx/ [2] https://lore.kernel.org/lkml/CABi2SkV9fnVjscLzKwa6FxNzfYU64ibGQxi7VWVEaGr-3gXfOg@xxxxxxxxxxxxxx/ Thanks Best regards, -Jeff
