| Message ID | 20260831054331.625505-1-anshuman.khandual@arm.com |
|---|---|
| Headers | show |
| Series | mm: Drop pxd_ERROR() | expand |
Hi Anshuman, On Mon, 2026-08-31 at 11:13 +0530, Anshuman Khandual wrote: > pxd_ERROR() macros have been provided by all platforms, which are very much > identical and can be dropped off completely if these pgtable printing could > be moved to callers in generic MM aka all pxd_clear_bad(). If pxd_ERROR() macros are really all identical, why not provide a common definiton of that macro? I'm always a bit concerned about such changes as there is always a risk to oversee arch-specific peculiarities. Adrian
On 31/08/26 11:38 AM, John Paul Adrian Glaubitz wrote: > Hi Anshuman, > > On Mon, 2026-08-31 at 11:13 +0530, Anshuman Khandual wrote: >> pxd_ERROR() macros have been provided by all platforms, which are very much >> identical and can be dropped off completely if these pgtable printing could >> be moved to callers in generic MM aka all pxd_clear_bad(). > > If pxd_ERROR() macros are really all identical, why not provide a common definiton > of that macro? I'm always a bit concerned about such changes as there is always > a risk to oversee arch-specific peculiarities. We don't those platform macros any more after converting generic MM into using ptval_bytes_to_hex_str() which now handles different pgtable entry types. > > Adrian >
On Mon, 31 Aug 2026 11:13:22 +0530 Anshuman Khandual <anshuman.khandual@arm.com> wrote: > pxd_ERROR() macros have been provided by all platforms, which are very much > identical and can be dropped off completely if these pgtable printing could > be moved to callers in generic MM aka all pxd_clear_bad(). > > But first cleanups and re-organizations are required in some platforms that > are using these macros internally. Afterwards [pte|pmd|pud|p4d|pgd]_ERROR() > macros have been completely dropped from the entire tree. All was going nicely until... > This series is based on v7.3-rc1 after the following patch and has also been > built tested on multiple platforms. *what* following patch??? Oh well, I'll queue it all up and shall hope for the arch maintainer acks to roll in. Please.
On Mon, Aug 31, 2026 at 06:55:19PM -0700, Andrew Morton wrote: > On Mon, 31 Aug 2026 11:13:22 +0530 Anshuman Khandual <anshuman.khandual@arm.com> wrote: > > > pxd_ERROR() macros have been provided by all platforms, which are very much > > identical and can be dropped off completely if these pgtable printing could > > be moved to callers in generic MM aka all pxd_clear_bad(). > > > > But first cleanups and re-organizations are required in some platforms that > > are using these macros internally. Afterwards [pte|pmd|pud|p4d|pgd]_ERROR() > > macros have been completely dropped from the entire tree. > > All was going nicely until... > > > This series is based on v7.3-rc1 after the following patch and has also been > > built tested on multiple platforms. > > *what* following patch??? The dependency patch has already been merged upstream. 092836fedd82 (mm: standardize printing for pgtable entries) Please ignore this. > > Oh well, I'll queue it all up and shall hope for the arch maintainer > acks to roll in. Please. Thanks !