Page MenuHomeFreeBSD

mbuf: Add M_WRITABLE_EXTPG
AcceptedPublic

Authored by jhb on Sep 25 2024, 3:13 PM.
Tags
None
Referenced Files
F170955208: D46787.diff
Mon, Sep 7, 8:30 PM
Unknown Object (File)
Sat, Sep 5, 5:06 AM
Unknown Object (File)
Sat, Sep 5, 4:18 AM
Unknown Object (File)
Sat, Sep 5, 3:05 AM
Unknown Object (File)
Thu, Sep 3, 6:38 PM
Unknown Object (File)
Thu, Sep 3, 4:45 PM
Unknown Object (File)
Thu, Sep 3, 5:20 AM
Unknown Object (File)
Thu, Sep 3, 1:30 AM
Subscribers

Details

Reviewers
gallatin
kp
Summary

This can be used instead of M_WRITABLE where the calling code is
prepared to handle writing to M_EXTPG mbufs.

Diff Detail

Repository
rG FreeBSD src repository
Lint
Lint Skipped
Unit
Tests Skipped
Build Status
Buildable 59597
Build 56484: arc lint + arc unit

Event Timeline

jhb requested review of this revision.Sep 25 2024, 3:13 PM

I'm confused.. if we are marking non-writable M_EXTPG mufs as M_RDONLY, why can't we simply remove the M_EXTPG check from M_WRITABLE? Why do we need a new macro?

I'm confused.. if we are marking non-writable M_EXTPG mufs as M_RDONLY, why can't we simply remove the M_EXTPG check from M_WRITABLE? Why do we need a new macro?

Because there are many places that call M_WRITABLE in the tree and if it succeeds expect to directly use mtod() to get a writable pointer to the data payload (possibly after using m_pullup which panics if called on an M_EXTPG mbuf). m_copyback() is explicitly aware of M_EXTPG mbufs and handles them correctly, so the new macro works fine the assertion there.

This revision is now accepted and ready to land.Sep 26 2024, 2:37 PM