Re: [PATCH v2 3/6] mm: replace vma->vm_flags direct modifications with modifier calls
From: Suren Baghdasaryan <surenb@google.com>
Date: 2023-01-25 16:56:03
Also in:
alsa-devel, amd-gfx, bpf, dmaengine, dri-devel, intel-gfx, kexec, kvm, linux-acpi, linux-arch, linux-arm-msm, linux-crypto, linux-ext4, linux-fbdev, linux-fsdevel, linux-media, linux-mediatek, linux-mm, linux-perf-users, linux-rdma, linux-s390, linux-samsung-soc, linux-scsi, linux-staging, linux-tegra, linux-um, linux-usb, linux-xfs, lkml, loongarch, netdev, nvdimm, selinux, sparclinux, target-devel, xen-devel
On Wed, Jan 25, 2023 at 1:30 AM 'Michal Hocko' via kernel-team [off-list ref] wrote:
On Wed 25-01-23 00:38:48, Suren Baghdasaryan wrote:quoted
Replace direct modifications to vma->vm_flags with calls to modifier functions to be able to track flag changes and to keep vma locking correctness.Is this a manual (git grep) based work or have you used Coccinele for the patch generation?
It was a manual "search and replace" and in the process I temporarily renamed vm_flags to ensure I did not miss any usage.
My potentially incomplete check $ git grep ">[[:space:]]*vm_flags[[:space:]]*[&|^]=" shows that nothing should be left after this. There is still quite a lot of direct checks of the flags (more than 600). Maybe it would be good to make flags accessible only via accessors which would also prevent any future direct setting of those flags in uncontrolled way as well.
Yes, I think Peter's suggestion in the first patch would also require that. Much more churn but probably worth it for the future maintenance. I'll add a patch which converts all readers as well.
Anyway Acked-by: Michal Hocko <mhocko@suse.com>
Thanks for all the reviews!
-- Michal Hocko SUSE Labs -- To unsubscribe from this group and stop receiving emails from it, send an email to kernel-team+unsubscribe@android.com.