Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 11 Nov 2000 15:12:29 -0800 (PST)
From:      Bosko Milekic <bmilekic@FreeBSD.org>
To:        cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   cvs commit: src/sys/dev/en midway.c src/sys/kern uipc_mbuf2.c uipc_syscalls.c src/sys/netinet ip_fil.c src/sys/netinet6 ipsec.c src/sys/pci if_sk.c if_ti.c if_wb.c src/sys/sys mbuf.h
Message-ID:  <200011112312.PAA35598@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help
bmilekic    2000/11/11 15:12:29 PST

  Modified files:
    sys/dev/en           midway.c 
  Log:
  Change check for existence of mbuf->m_ext.ext_free to check of new ext_type
  in order to determine whether the ext_buf is a cluster, or some other type
  of storage.
  
  Reviewed by: jlemon
  
  Revision  Changes    Path
  1.26      +2 -2      src/sys/dev/en/midway.c


  Modified files:
    sys/kern             uipc_mbuf2.c uipc_syscalls.c 
  Log:
  * Have m_pulldown() use the new M_WRITABLE() macro in order to determine
  whether the given ext_buf is shared.
  
  * Have the sf_bufs be setup with the mbuf subsystem using MEXTADD() with the
  two new arguments.
  
  Note: m_pulldown() is somewhat crotchy; the added comment explains the
  situation.
  
  Reviewed by: jlemon
  
  Revision  Changes    Path
  1.5       +32 -15    src/sys/kern/uipc_mbuf2.c
  1.75      +3 -2      src/sys/kern/uipc_syscalls.c


  Modified files:
    sys/netinet          ip_fil.c 
  Log:
  While I'm here, get rid of (now useless) MCLISREFERENCED and use MEXT_IS_REF
  instead.
  Also, fix a small set of "avail." If we're setting `avail,' we shouldn't
  be re-checking whether m_flags is M_EXT, because we know that it is, as if
  it wasn't, we would have already returned several lines above.
  
  Reviewed by: jlemon
  
  Revision  Changes    Path
  1.25      +3 -7      src/sys/netinet/ip_fil.c


  Modified files:
    sys/netinet6         ipsec.c 
  Log:
  Change check from mbuf->m_ext.ext_free to use the new ext_type in order
  to determine whether the given mbuf has a cluster (or some other type of
  external storage) attached to it.
  
  Note: This code should eventually be made to use M_WRITABLE() to determine
  whether or not a copy should be made.
  
  Reviewed by: jlemon
  
  Revision  Changes    Path
  1.9       +4 -7      src/sys/netinet6/ipsec.c


  Modified files:
    sys/pci              if_sk.c if_ti.c if_wb.c 
  Log:
  Change MEXTADD usage to pass the two new arguments.
  
  Reviewed by: jlemon
  
  Revision  Changes    Path
  1.35      +3 -3      src/sys/pci/if_sk.c
  1.40      +3 -3      src/sys/pci/if_ti.c
  1.35      +4 -3      src/sys/pci/if_wb.c


  Modified files:
    sys/sys              mbuf.h 
  Log:
  Add new M_RDONLY flag, which can be set during setup (MEXTADD, for example),
  and which indicates that the ext_buf is to be marked READ-ONLY.
  
  Also introduce M_WRITABLE() macro, which will return true if:
  * M_RDONLY is _not_ set AND:
  * either mbuf has no ext_buf attached OR it has an ext_buf attached but the ref
    count is NOT above 1.
  
  sf_bufs (sendfile) are marked M_RDONLY.
  
  Now code determining whether an mbuf's data area is "shared" or not can be
  made to use this instead of doing its own (often incorrect) thing.
  
  Submitted by: bmilekic, dwmalone, and Ian Dowse
  Reviewed by: dwmalone, jlemon
  
  Revision  Changes    Path
  1.63      +33 -15    src/sys/sys/mbuf.h



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200011112312.PAA35598>