On 1/4/25 08:39, Christophe Leroy wrote: > Le 03/01/2025 à 17:13, Petr Pavlu a écrit : >> On 12/5/24 20:46, Christophe Leroy wrote: >>> This series reworks module loading to avoid leaving the module in a >>> stale state when protecting ro_after_init section fails. >>> >>> Once module init has succeded it is too late to cancel loading. >>> If setting ro_after_init data section to read-only fails, all we can >>> do is to inform the user through a warning. This is what patch 2 does. >>> >>> Then patch 3 tries to go a bit further by testing the ability to write >>> protect ro-after-init section prior to initialising the module. >> >> I've been holding off on applying this series to modules-next because >> there was still some discussion on the previous RFC version [1], and >> I wanted to give people more time to potentially comment. >> >> Mike Rapoport also recently posted a series with a patch [2] that >> proposes restoring of large pages after fragmentation. Should the last >> patch here be then dropped? > > Indeed, if the large pages are restored when bringing back the > ro_after_init to RW, it defeats the purpose of patch 3. > > So I agree, let's first apply patches 1 and 2 in order to fix the actual > bug then see how we can improve as a second step. I've now queued the first two patches on modules-next. -- Thanks, Petr
