1. 15 Dec, 2018 15 commits
    • mckusick's avatar
      Reorder ffs_verify_dinode_ckhash() so that it checks the inode check-hash · b27d4425
      mckusick authored
      before copying in the inode so that the mode and link-count are not set
      if the check-hash fails. This change ensures that the vnode will be properly
      unwound and recycled rather than being held in the cache.
      
      Initialize the file mode is zero so that if the loading of the inode
      fails (for example because of a check-hash failure), the vnode will be
      properly unwound and recycled.
      
      Reported by:  Gary Jennejohn (gj)
      Sponsored by: Netflix
      b27d4425
    • mckusick's avatar
      Must set ip->i_effnlink = ip->i_nlink to avoid a soft updates · c376c836
      mckusick authored
      "panic: softdep_update_inodeblock: bad link count" when releasing
      a partially initialized vnode after an inode check-hash failure.
      
      Reported by:  Gary Jennejohn <gljennjohn@gmail.com>
      Reported by:  Peter Holm (pho)
      Sponsored by: Netflix
      c376c836
    • mckusick's avatar
      Fsck would find, report, and offer to fix inode check-hash failures. · 94367dc4
      mckusick authored
      If requested to fix the inode check-hash it would confirm having done
      it, but then fail to make the fix. The same code is used in fsdb which,
      unlike fsck, would actually fix the inode check-hash.
      
      The discrepancy occurred because fsck has two ways to fetch inodes.
      The inode by number function ginode() and the streaming inode
      function getnextinode() used during pass1. Fsdb uses the ginode()
      function which correctly does the fix, while fsck first encounters
      the bad inode check-hash in pass1 where it is using the getnextinode()
      function that failed to make the correction. This patch corrects
      the getnextinode() function so that fsck now correctly fixes inodes
      with incorrect inode check-hashs.
      
      Reported by:  Gary Jennejohn <gljennjohn@gmail.com>
      Sponsored by: Netflix
      94367dc4
    • hiren's avatar
      Revert r331567 CC Cubic: fix underflow for cubic_cwnd() · 3dc78ca6
      hiren authored
      This change is causing TCP connections using cubic to hang. Need to dig more to
      find exact cause and fix it.
      
      Reported by:	tj at mrsk dot me, Matt Garber (via twitter)
      Discussed with:	sbruno (previously), allanjude, cperciva
      MFC after:	3 days
      3dc78ca6
    • se's avatar
      Remove CTM from 13-CURRENT after the release of FreeBSD-12.0. · 6126d79b
      se authored
      The removal (and creation of a port) has been pre-announced in UPDATING
      1 month ago. Packages are available for all supported FreeBSD vesions.
      
      I did not think that another entry in UPDATING is required to note the
      actual removal.
      
      No MFC is planned - CTM shall be kept in base for all releases up to 12.x.
      
      Reviewed by:	rgrimes
      Approved by:	imp, bcr (manpages)
      Relnotes:	yes
      Differential Revision:	https://reviews.freebsd.org/D17935
      6126d79b
    • brooks's avatar
      Fix bugs in plugable CC algorithm and siftr sysctls. · acaa8abf
      brooks authored
      Use the sysctl_handle_int() handler to write out the old value and read
      the new value into a temporary variable. Use the temporary variable
      for any checks of values rather than using the CAST_PTR_INT() macro on
      req->newptr. The prior usage read directly from userspace memory if the
      sysctl() was called correctly. This is unsafe and doesn't work at all on
      some architectures (at least i386.)
      
      In some cases, the code could also be tricked into reading from kernel
      memory and leaking limited information about the contents or crashing
      the system. This was true for CDG, newreno, and siftr on all platforms
      and true for i386 in all cases. The impact of this bug is largest in
      VIMAGE jails which have been configured to allow writing to these
      sysctls.
      
      Per discussion with the security officer, we will not be issuing an
      advisory for this issue as root access and a non-default config are
      required to be impacted.
      
      Reviewed by:	markj, bz
      Discussed with:	gordon (security officer)
      MFC after:	3 days
      Security:	kernel information leak, local DoS (both require root)
      Differential Revision:	https://reviews.freebsd.org/D18443
      acaa8abf
    • avos's avatar
      Add new USB id in rtwn_usb(4) (RTL8812AU) · 1df45307
      avos authored
      PR:		234029
      Submitted by:	<hakotani000@gmail.com>
      MFC after:	4 days
      1df45307
    • dim's avatar
      Update clang, llvm, lld, lldb, compiler-rt and libc++ version number to · d3f31084
      dim authored
      7.0.1 release r349250.  There were no functional changes since the 7.0.1
      rc3 import.
      
      PR:		230240, 230355
      Relnotes:	yes
      MFC after:	2 months
      X-MFC-With:	r341825
      d3f31084
    • trasz's avatar
      Make fsck(8) use pread(2). This cuts the number of syscalls by half. · 4b9b1e24
      trasz authored
      Reviewed by:	kib, mckusick
      MFC after:	2 weeks
      Sponsored by:	DARPA, AFRL
      Differential Revision:	https://reviews.freebsd.org/D17586
      4b9b1e24
    • trasz's avatar
      Add kern.rpc.gss.client_max, to make it possible to bump it easily. · 2ee1fcb2
      trasz authored
      This can drastically lower the load on gssd(8) on large NFS servers.
      
      Submitted by:	Per Andersson <pa at chalmers dot se>
      Reviewed by:	rmacklem@
      MFC after:	2 weeks
      Sponsored by:	Chalmers University of Technology
      Differential Revision:	https://reviews.freebsd.org/D18393
      2ee1fcb2
    • mmel's avatar
      Improve R_AARCH64_TLSDESC relocation. · 4a03477e
      mmel authored
      The original code did not support dynamically loaded libraries and used
      suboptimal access to TLS variables.
      New implementation removes lazy resolving of TLS relocation - due to flaw
      in TLSDESC design is impossible to switch resolver function at runtime
      without expensive locking.
      
      Due to this, 3 specialized resolvers are implemented:
       - load time resolver for TLS relocation from libraries loaded with main
         executable (thus with known TLS offset).
       - resolver for undefined thread weak symbols.
       - slower lazy resolver for dynamically loaded libraries with fast path for
         already resolved symbols.
      
      PR:		228892, 232149, 233204, 232311
      MFC after:	2 weeks
      Differential Revision:	https://reviews.freebsd.org/D18417
      4a03477e
    • cem's avatar
      Revert accidentally included changes in r342108 · b838d670
      cem authored
      If you're curious, please follow along in https://reviews.freebsd.org/D18537 .
      
      Sorry for the noise.
      b838d670
    • cem's avatar
      efirt: When present, attempt to use EFI runtime services to shutdown · a4ea2c92
      cem authored
      PR:		maybe related to 233998 (inconclusive at this time)
      Submitted by:	byuu <byuu AT tutanota.com> (previous version)
      Reviewed by:	imp
      Differential Revision:	https://reviews.freebsd.org/D18506
      a4ea2c92
    • jhibbits's avatar
      powerpcspe: Don't require FPU_EMU for powerpcspe IEEE emulation · 30f48ca3
      jhibbits authored
      Build only the necessary fpu_emu files for supporting the SPE IEEE-754
      emulation exception handler.
      
      MFC after:	1 week
      30f48ca3
    • gonzo's avatar
      [mv_pci] Do not attempt to attach disabled PCI ports · 8a26a904
      gonzo authored
      Fail probe for PCI port if the respective FDT node is not enabled
      
      Differential Revision:	https://reviews.freebsd.org/D18385
      8a26a904
  2. 14 Dec, 2018 25 commits