[00/44] Update RTEMS-specific header files to latest FreeBSD versions

Message ID 20180809060936.9726-1-sebastian.huber@embedded-brains.de
Headers show
Series
  • Update RTEMS-specific header files to latest FreeBSD versions
Related show

Message

Sebastian Huber Aug. 9, 2018, 6:08 a.m.
This is an RTEMS-specific patch set. It is a preparation to update the libbsd
(a port of FreeBSD code to RTEMS, e.g. the network stack) to new FreeBSD
baseline.

The patch set does two things:

1. It updates several header files imported from FreeBSD using the commits from
   FreeBSD.

2. It removes the kernel space content from some files and includes a
   <machine/_kernel_*.h> file instead.  This allows to use the Newlib provided
   header files with different FreeBSD baselines.

Sebastian Huber (7):
  RTEMS: Update FreeBSD version tags
  RTEMS: Add __BSD_VISIBLE in <sys/_termios.h>
  RTEMS: Introduce <machine/_kernel_uio.h>
  RTEMS: Introduce <machine/_kernel_in6.h>
  RTEMS: Introduce <machine/_kernel_in.h>
  RTEMS: Introduce <machine/_kernel_if.h>
  RTEMS: Introduce <machine/_kernel_socket.h>

ae@FreeBSD.org (1):
  Follow the RFC6980 and silently ignore following IPv6 NDP messages
    that had the IPv6 fragmentation header: o Neighbor Solicitation o
    Neighbor Advertisement o Router Solicitation o Router Advertisement
    o Redirect

brooks (5):
  Add _IOC_NEWLEN() and _IOC_NEWTYPE() macros.
  Move uio enums to sys/_uio.h.
  Fix access to ifru_buffer on freebsd32.
  Use an accessor function to access ifr_data.
  Add 32-bit compat for ioctls that take struct ifgroupreq.

delphij (1):
  Implement INHERIT_ZERO for minherit(2).

des (1):
  Correct sysctl names.

ed@FreeBSD.org (1):
  mprotect(): Change prototype to comply to POSIX.

glebius (2):
  Listening sockets improvements.
  Garbage collect IFCAP_POLLING_NOCOUNT. It wasn't used since very
    beginning of polling(4). The module always ignored return value from
    driver polling handler.

imp (1):
  Renumber copyright clause 4

jeff (1):
  Implement several enhancements to NUMA policies.

jhb (4):
  Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a
    mapping use an address in the first 2GB of the process's address
    space. This flag should have the same semantics as the same flag on
    Linux.
  Add a new fo_fill_kinfo fileops method to add type-specific
    information to struct kinfo_file. - Move the various fill_*_info()
    methods out of kern_descrip.c and into the various file type
    implementations. - Rework the support for kinfo_ofile to generate a
    suitable kinfo_file object for each file and then convert that to a
    kinfo_ofile structure rather than keeping a second, different set of
    code that directly manipulates type-specific file information. -
    Remove the shm_path() and ksem_info() layering violations.
  Retire the unimplemented MAP_RENAME and MAP_NORESERVE flags to
    mmap(2). Older binaries are still permitted to use these flags.
  Add a new file operations hook for mmap operations. File type-specific
    logic is now placed in the mmap hook implementation rather than
    requiring it to be placed in sys/vm/vm_mmap.c. This hook allows new
    file types to support mmap() as well as potentially allowing mmap()
    for existing file types that do not currently support any mapping.

jtl (1):
  Add the "TCP Blackbox Recorder" which we discussed at the developer
    summits at BSDCan and BSDCam in 2017.

kib (8):
  Implement read(2)/write(2) and neccessary lseek(2) for posix shmfd.
    Add MAC framework entries for posix shm read and write.
  Add MAP_EXCL flag for mmap(2). It should be combined with MAP_FIXED,
    and prevents the request from deleting existing mappings in the
    region, failing instead.
  Implement process-shared locks support for libthr.so.3, without
    breaking the ABI. Special value is stored in the lock pointer to
    indicate shared lock, and offline page in the shared memory is
    allocated to store the actual lock.
  Implement address space guards.
  Relax visibility for some termios symbols.
  Add a place for a driver to report rx timestamps in nanoseconds from
    boot for the received packets.
  Use hardware timestamps to report packet timestamps for SO_TIMESTAMP
    and other similar socket options.
  Allow to specify PCP on packets not belonging to any VLAN.

mmacy (1):
  iflib(9): Add support for cloning pseudo interfaces

pfg (4):
  sys: further adoption of SPDX licensing ID tags.
  include: further adoption of SPDX licensing ID tags.
  sys/sys: further adoption of SPDX licensing ID tags.
  SPDX: license IDs for some ISC-related files.

pkelsey (1):
  This is an implementation of the client side of TCP Fast Open (TFO)
    [RFC7413]. It also includes a pre-shared key mode of operation in
    which the server requires the client to be in possession of a shared
    secret in order to successfully open TFO connections with that
    server.

rrs (1):
  This commit brings in a new refactored TCP stack called Rack. Rack
    includes the following features: - A different SACK processing
    scheme (the old sack structures are not used). - RACK (Recent
    acknowledgment) where counting dup-acks is no longer done instead
    time is used to knwo when to retransmit. (see the I-D) - TLP (Tail
    Loss Probe) where we will probe for tail-losses to attempt to try
    not to take a retransmit time-out. (see the I-D) - Burst mitigation
    using TCPHTPS - PRR (partial rate reduction) see the RFC.

sbruno (3):
  Load balance sockets with new SO_REUSEPORT_LB option
  Revert r332894 at the request of the submitter.
  Load balance sockets with new SO_REUSEPORT_LB option.

sephe (1):
  if: Add ioctls to get RSS key and hash type/function.

 newlib/libc/sys/rtems/include/arpa/inet.h          |   6 +-
 newlib/libc/sys/rtems/include/net/if.h             |  66 ++++++--
 newlib/libc/sys/rtems/include/netdb.h              |   4 +-
 newlib/libc/sys/rtems/include/netinet/in.h         |  36 +----
 newlib/libc/sys/rtems/include/netinet/tcp.h        |  85 +++++++++-
 newlib/libc/sys/rtems/include/netinet6/in6.h       | 175 ++-------------------
 newlib/libc/sys/rtems/include/sys/_bitset.h        |  10 +-
 newlib/libc/sys/rtems/include/sys/_cpuset.h        |   4 +-
 newlib/libc/sys/rtems/include/sys/_iovec.h         |   4 +-
 .../libc/sys/rtems/include/sys/_sockaddr_storage.h |   4 +-
 newlib/libc/sys/rtems/include/sys/_termios.h       |  14 +-
 newlib/libc/sys/rtems/include/sys/_uio.h           |  52 ++++++
 newlib/libc/sys/rtems/include/sys/bitset.h         |   4 +-
 newlib/libc/sys/rtems/include/sys/filio.h          |   4 +-
 newlib/libc/sys/rtems/include/sys/ioccom.h         |   6 +
 newlib/libc/sys/rtems/include/sys/ioctl.h          |   4 +-
 newlib/libc/sys/rtems/include/sys/mman.h           |  35 ++++-
 newlib/libc/sys/rtems/include/sys/socket.h         | 101 ++++--------
 newlib/libc/sys/rtems/include/sys/sockio.h         |  11 +-
 newlib/libc/sys/rtems/include/sys/syslog.h         |   4 +-
 newlib/libc/sys/rtems/include/sys/ttycom.h         |   4 +-
 newlib/libc/sys/rtems/include/sys/ttydefaults.h    |   4 +-
 newlib/libc/sys/rtems/include/sys/uio.h            |  76 ++-------
 newlib/libc/sys/rtems/include/sys/un.h             |   4 +-
 newlib/libc/sys/rtems/include/termios.h            |   4 +-
 25 files changed, 357 insertions(+), 364 deletions(-)
 create mode 100644 newlib/libc/sys/rtems/include/sys/_uio.h

-- 
2.13.7

Comments

Eric Blake Aug. 9, 2018, 1:12 p.m. | #1
On 08/09/2018 01:08 AM, Sebastian Huber wrote:
> This is an RTEMS-specific patch set. It is a preparation to update the libbsd

> (a port of FreeBSD code to RTEMS, e.g. the network stack) to new FreeBSD

> baseline.

> 

> The patch set does two things:

> 

> 1. It updates several header files imported from FreeBSD using the commits from

>     FreeBSD.

> 

> 2. It removes the kernel space content from some files and includes a

>     <machine/_kernel_*.h> file instead.  This allows to use the Newlib provided

>     header files with different FreeBSD baselines.


This series is full of poorly-formatted commit messages:

> ae@FreeBSD.org (1):

>    Follow the RFC6980 and silently ignore following IPv6 NDP messages

>      that had the IPv6 fragmentation header: o Neighbor Solicitation o

>      Neighbor Advertisement o Router Solicitation o Router Advertisement

>      o Redirect


> glebius (2):

...
>    Garbage collect IFCAP_POLLING_NOCOUNT. It wasn't used since very

>      beginning of polling(4). The module always ignored return value from

>      driver polling handler.

> 


> jhb (4):

>    Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a

>      mapping use an address in the first 2GB of the process's address

>      space. This flag should have the same semantics as the same flag on

>      Linux.

>    Add a new fo_fill_kinfo fileops method to add type-specific

>      information to struct kinfo_file. - Move the various fill_*_info()

>      methods out of kern_descrip.c and into the various file type

>      implementations. - Rework the support for kinfo_ofile to generate a

>      suitable kinfo_file object for each file and then convert that to a

>      kinfo_ofile structure rather than keeping a second, different set of

>      code that directly manipulates type-specific file information. -

>      Remove the shm_path() and ksem_info() layering violations.

>    Retire the unimplemented MAP_RENAME and MAP_NORESERVE flags to

>      mmap(2). Older binaries are still permitted to use these flags.

>    Add a new file operations hook for mmap operations. File type-specific

>      logic is now placed in the mmap hook implementation rather than

>      requiring it to be placed in sys/vm/vm_mmap.c. This hook allows new

>      file types to support mmap() as well as potentially allowing mmap()

>      for existing file types that do not currently support any mapping.

> 

> jtl (1):

>    Add the "TCP Blackbox Recorder" which we discussed at the developer

>      summits at BSDCan and BSDCam in 2017.

> 

> kib (8):

>    Implement read(2)/write(2) and neccessary lseek(2) for posix shmfd.

>      Add MAC framework entries for posix shm read and write.

>    Add MAP_EXCL flag for mmap(2). It should be combined with MAP_FIXED,

>      and prevents the request from deleting existing mappings in the

>      region, failing instead.

>    Implement process-shared locks support for libthr.so.3, without

>      breaking the ABI. Special value is stored in the lock pointer to

>      indicate shared lock, and offline page in the shared memory is

>      allocated to store the actual lock.

...
>    Add a place for a driver to report rx timestamps in nanoseconds from

>      boot for the received packets.

>    Use hardware timestamps to report packet timestamps for SO_TIMESTAMP

>      and other similar socket options.


> 

> pkelsey (1):

>    This is an implementation of the client side of TCP Fast Open (TFO)

>      [RFC7413]. It also includes a pre-shared key mode of operation in

>      which the server requires the client to be in possession of a shared

>      secret in order to successfully open TFO connections with that

>      server.

> 

> rrs (1):

>    This commit brings in a new refactored TCP stack called Rack. Rack

>      includes the following features: - A different SACK processing

>      scheme (the old sack structures are not used). - RACK (Recent

>      acknowledgment) where counting dup-acks is no longer done instead

>      time is used to knwo when to retransmit. (see the I-D) - TLP (Tail

>      Loss Probe) where we will probe for tail-losses to attempt to try

>      not to take a retransmit time-out. (see the I-D) - Burst mitigation

>      using TCPHTPS - PRR (partial rate reduction) see the RFC.


Please reformat the commit messages to have a SHORT summary (60-70 
characters), then a blank line, then the rest of the description, rather 
than diving right in to description that git then uses as the overlong 
subject line.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3266
Virtualization:  qemu.org | libvirt.org
Sebastian Huber Aug. 9, 2018, 1:13 p.m. | #2
On 09/08/18 15:12, Eric Blake wrote:
> On 08/09/2018 01:08 AM, Sebastian Huber wrote:

>> This is an RTEMS-specific patch set. It is a preparation to update 

>> the libbsd

>> (a port of FreeBSD code to RTEMS, e.g. the network stack) to new FreeBSD

>> baseline.

>>

>> The patch set does two things:

>>

>> 1. It updates several header files imported from FreeBSD using the 

>> commits from

>>     FreeBSD.

>>

>> 2. It removes the kernel space content from some files and includes a

>>     <machine/_kernel_*.h> file instead.  This allows to use the 

>> Newlib provided

>>     header files with different FreeBSD baselines.

>

> This series is full of poorly-formatted commit messages:

>

>> ae@FreeBSD.org (1):

>>    Follow the RFC6980 and silently ignore following IPv6 NDP messages

>>      that had the IPv6 fragmentation header: o Neighbor Solicitation o

>>      Neighbor Advertisement o Router Solicitation o Router Advertisement

>>      o Redirect

>

>> glebius (2):

> ...

>>    Garbage collect IFCAP_POLLING_NOCOUNT. It wasn't used since very

>>      beginning of polling(4). The module always ignored return value 

>> from

>>      driver polling handler.

>>

>

>> jhb (4):

>>    Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a

>>      mapping use an address in the first 2GB of the process's address

>>      space. This flag should have the same semantics as the same flag on

>>      Linux.

>>    Add a new fo_fill_kinfo fileops method to add type-specific

>>      information to struct kinfo_file. - Move the various fill_*_info()

>>      methods out of kern_descrip.c and into the various file type

>>      implementations. - Rework the support for kinfo_ofile to generate a

>>      suitable kinfo_file object for each file and then convert that to a

>>      kinfo_ofile structure rather than keeping a second, different 

>> set of

>>      code that directly manipulates type-specific file information. -

>>      Remove the shm_path() and ksem_info() layering violations.

>>    Retire the unimplemented MAP_RENAME and MAP_NORESERVE flags to

>>      mmap(2). Older binaries are still permitted to use these flags.

>>    Add a new file operations hook for mmap operations. File 

>> type-specific

>>      logic is now placed in the mmap hook implementation rather than

>>      requiring it to be placed in sys/vm/vm_mmap.c. This hook allows new

>>      file types to support mmap() as well as potentially allowing mmap()

>>      for existing file types that do not currently support any mapping.

>>

>> jtl (1):

>>    Add the "TCP Blackbox Recorder" which we discussed at the developer

>>      summits at BSDCan and BSDCam in 2017.

>>

>> kib (8):

>>    Implement read(2)/write(2) and neccessary lseek(2) for posix shmfd.

>>      Add MAC framework entries for posix shm read and write.

>>    Add MAP_EXCL flag for mmap(2). It should be combined with MAP_FIXED,

>>      and prevents the request from deleting existing mappings in the

>>      region, failing instead.

>>    Implement process-shared locks support for libthr.so.3, without

>>      breaking the ABI. Special value is stored in the lock pointer to

>>      indicate shared lock, and offline page in the shared memory is

>>      allocated to store the actual lock.

> ...

>>    Add a place for a driver to report rx timestamps in nanoseconds from

>>      boot for the received packets.

>>    Use hardware timestamps to report packet timestamps for SO_TIMESTAMP

>>      and other similar socket options.

>

>>

>> pkelsey (1):

>>    This is an implementation of the client side of TCP Fast Open (TFO)

>>      [RFC7413]. It also includes a pre-shared key mode of operation in

>>      which the server requires the client to be in possession of a 

>> shared

>>      secret in order to successfully open TFO connections with that

>>      server.

>>

>> rrs (1):

>>    This commit brings in a new refactored TCP stack called Rack. Rack

>>      includes the following features: - A different SACK processing

>>      scheme (the old sack structures are not used). - RACK (Recent

>>      acknowledgment) where counting dup-acks is no longer done instead

>>      time is used to knwo when to retransmit. (see the I-D) - TLP (Tail

>>      Loss Probe) where we will probe for tail-losses to attempt to try

>>      not to take a retransmit time-out. (see the I-D) - Burst mitigation

>>      using TCPHTPS - PRR (partial rate reduction) see the RFC.

>

> Please reformat the commit messages to have a SHORT summary (60-70 

> characters), then a blank line, then the rest of the description, 

> rather than diving right in to description that git then uses as the 

> overlong subject line.

>


I would like to keep the messages as is. This makes it easier to compare 
upstream FreeBSD changes with changes imported to Newlib.

-- 
Sebastian Huber, embedded brains GmbH

Address : Dornierstr. 4, D-82178 Puchheim, Germany
Phone   : +49 89 189 47 41-16
Fax     : +49 89 189 47 41-09
E-Mail  : sebastian.huber@embedded-brains.de
PGP     : Public key available on request.

Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.
Gedare Bloom Aug. 9, 2018, 4:46 p.m. | #3
On Thu, Aug 9, 2018 at 9:13 AM, Sebastian Huber
<sebastian.huber@embedded-brains.de> wrote:
> On 09/08/18 15:12, Eric Blake wrote:

>>

>> On 08/09/2018 01:08 AM, Sebastian Huber wrote:

>>>

>>> This is an RTEMS-specific patch set. It is a preparation to update the

>>> libbsd

>>> (a port of FreeBSD code to RTEMS, e.g. the network stack) to new FreeBSD

>>> baseline.

>>>

>>> The patch set does two things:

>>>

>>> 1. It updates several header files imported from FreeBSD using the

>>> commits from

>>>     FreeBSD.

>>>

>>> 2. It removes the kernel space content from some files and includes a

>>>     <machine/_kernel_*.h> file instead.  This allows to use the Newlib

>>> provided

>>>     header files with different FreeBSD baselines.

>>

>>

>> This series is full of poorly-formatted commit messages:

>>

>>> ae@FreeBSD.org (1):

>>>    Follow the RFC6980 and silently ignore following IPv6 NDP messages

>>>      that had the IPv6 fragmentation header: o Neighbor Solicitation o

>>>      Neighbor Advertisement o Router Solicitation o Router Advertisement

>>>      o Redirect

>>

>>

>>> glebius (2):

>>

>> ...

>>>

>>>    Garbage collect IFCAP_POLLING_NOCOUNT. It wasn't used since very

>>>      beginning of polling(4). The module always ignored return value from

>>>      driver polling handler.

>>>

>>

>>> jhb (4):

>>>    Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a

>>>      mapping use an address in the first 2GB of the process's address

>>>      space. This flag should have the same semantics as the same flag on

>>>      Linux.

>>>    Add a new fo_fill_kinfo fileops method to add type-specific

>>>      information to struct kinfo_file. - Move the various fill_*_info()

>>>      methods out of kern_descrip.c and into the various file type

>>>      implementations. - Rework the support for kinfo_ofile to generate a

>>>      suitable kinfo_file object for each file and then convert that to a

>>>      kinfo_ofile structure rather than keeping a second, different set of

>>>      code that directly manipulates type-specific file information. -

>>>      Remove the shm_path() and ksem_info() layering violations.

>>>    Retire the unimplemented MAP_RENAME and MAP_NORESERVE flags to

>>>      mmap(2). Older binaries are still permitted to use these flags.

>>>    Add a new file operations hook for mmap operations. File type-specific

>>>      logic is now placed in the mmap hook implementation rather than

>>>      requiring it to be placed in sys/vm/vm_mmap.c. This hook allows new

>>>      file types to support mmap() as well as potentially allowing mmap()

>>>      for existing file types that do not currently support any mapping.

>>>

>>> jtl (1):

>>>    Add the "TCP Blackbox Recorder" which we discussed at the developer

>>>      summits at BSDCan and BSDCam in 2017.

>>>

>>> kib (8):

>>>    Implement read(2)/write(2) and neccessary lseek(2) for posix shmfd.

>>>      Add MAC framework entries for posix shm read and write.

>>>    Add MAP_EXCL flag for mmap(2). It should be combined with MAP_FIXED,

>>>      and prevents the request from deleting existing mappings in the

>>>      region, failing instead.

>>>    Implement process-shared locks support for libthr.so.3, without

>>>      breaking the ABI. Special value is stored in the lock pointer to

>>>      indicate shared lock, and offline page in the shared memory is

>>>      allocated to store the actual lock.

>>

>> ...

>>>

>>>    Add a place for a driver to report rx timestamps in nanoseconds from

>>>      boot for the received packets.

>>>    Use hardware timestamps to report packet timestamps for SO_TIMESTAMP

>>>      and other similar socket options.

>>

>>

>>>

>>> pkelsey (1):

>>>    This is an implementation of the client side of TCP Fast Open (TFO)

>>>      [RFC7413]. It also includes a pre-shared key mode of operation in

>>>      which the server requires the client to be in possession of a shared

>>>      secret in order to successfully open TFO connections with that

>>>      server.

>>>

>>> rrs (1):

>>>    This commit brings in a new refactored TCP stack called Rack. Rack

>>>      includes the following features: - A different SACK processing

>>>      scheme (the old sack structures are not used). - RACK (Recent

>>>      acknowledgment) where counting dup-acks is no longer done instead

>>>      time is used to knwo when to retransmit. (see the I-D) - TLP (Tail

>>>      Loss Probe) where we will probe for tail-losses to attempt to try

>>>      not to take a retransmit time-out. (see the I-D) - Burst mitigation

>>>      using TCPHTPS - PRR (partial rate reduction) see the RFC.

>>

>>

>> Please reformat the commit messages to have a SHORT summary (60-70

>> characters), then a blank line, then the rest of the description, rather

>> than diving right in to description that git then uses as the overlong

>> subject line.

>>

>

> I would like to keep the messages as is. This makes it easier to compare

> upstream FreeBSD changes with changes imported to Newlib.

>


In this case, since the commits come from a third party repository, we
would like to request inclusion of the commits in newlib without
modification. This allows us to more easily update to future versions
of the third party code, and provides a cleaner history/provenance of
the code.

Gedare Bloom
RTEMS Project

> --

> Sebastian Huber, embedded brains GmbH

>

> Address : Dornierstr. 4, D-82178 Puchheim, Germany

> Phone   : +49 89 189 47 41-16

> Fax     : +49 89 189 47 41-09

> E-Mail  : sebastian.huber@embedded-brains.de

> PGP     : Public key available on request.

>

> Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.

>
Richard Earnshaw (lists) Aug. 9, 2018, 5:24 p.m. | #4
On 09/08/18 17:46, Gedare Bloom wrote:
> On Thu, Aug 9, 2018 at 9:13 AM, Sebastian Huber

> <sebastian.huber@embedded-brains.de> wrote:

>> On 09/08/18 15:12, Eric Blake wrote:

>>>

>>> On 08/09/2018 01:08 AM, Sebastian Huber wrote:

>>>>

>>>> This is an RTEMS-specific patch set. It is a preparation to update the

>>>> libbsd

>>>> (a port of FreeBSD code to RTEMS, e.g. the network stack) to new FreeBSD

>>>> baseline.

>>>>

>>>> The patch set does two things:

>>>>

>>>> 1. It updates several header files imported from FreeBSD using the

>>>> commits from

>>>>     FreeBSD.

>>>>

>>>> 2. It removes the kernel space content from some files and includes a

>>>>     <machine/_kernel_*.h> file instead.  This allows to use the Newlib

>>>> provided

>>>>     header files with different FreeBSD baselines.

>>>

>>>

>>> This series is full of poorly-formatted commit messages:

>>>

>>>> ae@FreeBSD.org (1):

>>>>    Follow the RFC6980 and silently ignore following IPv6 NDP messages

>>>>      that had the IPv6 fragmentation header: o Neighbor Solicitation o

>>>>      Neighbor Advertisement o Router Solicitation o Router Advertisement

>>>>      o Redirect

>>>

>>>

>>>> glebius (2):

>>>

>>> ...

>>>>

>>>>    Garbage collect IFCAP_POLLING_NOCOUNT. It wasn't used since very

>>>>      beginning of polling(4). The module always ignored return value from

>>>>      driver polling handler.

>>>>

>>>

>>>> jhb (4):

>>>>    Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a

>>>>      mapping use an address in the first 2GB of the process's address

>>>>      space. This flag should have the same semantics as the same flag on

>>>>      Linux.

>>>>    Add a new fo_fill_kinfo fileops method to add type-specific

>>>>      information to struct kinfo_file. - Move the various fill_*_info()

>>>>      methods out of kern_descrip.c and into the various file type

>>>>      implementations. - Rework the support for kinfo_ofile to generate a

>>>>      suitable kinfo_file object for each file and then convert that to a

>>>>      kinfo_ofile structure rather than keeping a second, different set of

>>>>      code that directly manipulates type-specific file information. -

>>>>      Remove the shm_path() and ksem_info() layering violations.

>>>>    Retire the unimplemented MAP_RENAME and MAP_NORESERVE flags to

>>>>      mmap(2). Older binaries are still permitted to use these flags.

>>>>    Add a new file operations hook for mmap operations. File type-specific

>>>>      logic is now placed in the mmap hook implementation rather than

>>>>      requiring it to be placed in sys/vm/vm_mmap.c. This hook allows new

>>>>      file types to support mmap() as well as potentially allowing mmap()

>>>>      for existing file types that do not currently support any mapping.

>>>>

>>>> jtl (1):

>>>>    Add the "TCP Blackbox Recorder" which we discussed at the developer

>>>>      summits at BSDCan and BSDCam in 2017.

>>>>

>>>> kib (8):

>>>>    Implement read(2)/write(2) and neccessary lseek(2) for posix shmfd.

>>>>      Add MAC framework entries for posix shm read and write.

>>>>    Add MAP_EXCL flag for mmap(2). It should be combined with MAP_FIXED,

>>>>      and prevents the request from deleting existing mappings in the

>>>>      region, failing instead.

>>>>    Implement process-shared locks support for libthr.so.3, without

>>>>      breaking the ABI. Special value is stored in the lock pointer to

>>>>      indicate shared lock, and offline page in the shared memory is

>>>>      allocated to store the actual lock.

>>>

>>> ...

>>>>

>>>>    Add a place for a driver to report rx timestamps in nanoseconds from

>>>>      boot for the received packets.

>>>>    Use hardware timestamps to report packet timestamps for SO_TIMESTAMP

>>>>      and other similar socket options.

>>>

>>>

>>>>

>>>> pkelsey (1):

>>>>    This is an implementation of the client side of TCP Fast Open (TFO)

>>>>      [RFC7413]. It also includes a pre-shared key mode of operation in

>>>>      which the server requires the client to be in possession of a shared

>>>>      secret in order to successfully open TFO connections with that

>>>>      server.

>>>>

>>>> rrs (1):

>>>>    This commit brings in a new refactored TCP stack called Rack. Rack

>>>>      includes the following features: - A different SACK processing

>>>>      scheme (the old sack structures are not used). - RACK (Recent

>>>>      acknowledgment) where counting dup-acks is no longer done instead

>>>>      time is used to knwo when to retransmit. (see the I-D) - TLP (Tail

>>>>      Loss Probe) where we will probe for tail-losses to attempt to try

>>>>      not to take a retransmit time-out. (see the I-D) - Burst mitigation

>>>>      using TCPHTPS - PRR (partial rate reduction) see the RFC.

>>>

>>>

>>> Please reformat the commit messages to have a SHORT summary (60-70

>>> characters), then a blank line, then the rest of the description, rather

>>> than diving right in to description that git then uses as the overlong

>>> subject line.

>>>

>>

>> I would like to keep the messages as is. This makes it easier to compare

>> upstream FreeBSD changes with changes imported to Newlib.

>>

> 

> In this case, since the commits come from a third party repository, we

> would like to request inclusion of the commits in newlib without

> modification. This allows us to more easily update to future versions

> of the third party code, and provides a cleaner history/provenance of

> the code.

> 


Surely *adding* one extra line (followed by one blank line) as a summary
at the head of the commit won't really interfere with such tracking.
We're not asking you to reformat any of the remaining text.

R.

> Gedare Bloom

> RTEMS Project

> 

>> --

>> Sebastian Huber, embedded brains GmbH

>>

>> Address : Dornierstr. 4, D-82178 Puchheim, Germany

>> Phone   : +49 89 189 47 41-16

>> Fax     : +49 89 189 47 41-09

>> E-Mail  : sebastian.huber@embedded-brains.de

>> PGP     : Public key available on request.

>>

>> Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG.

>>
Corinna Vinschen Aug. 9, 2018, 8:33 p.m. | #5
On Aug  9 18:24, Richard Earnshaw (lists) wrote:
> On 09/08/18 17:46, Gedare Bloom wrote:

> > On Thu, Aug 9, 2018 at 9:13 AM, Sebastian Huber

> > <sebastian.huber@embedded-brains.de> wrote:

> >> On 09/08/18 15:12, Eric Blake wrote:

> >>>

> >>> On 08/09/2018 01:08 AM, Sebastian Huber wrote:

> >>>>

> >>>> This is an RTEMS-specific patch set. It is a preparation to update the

> >>>> libbsd

> >>>> (a port of FreeBSD code to RTEMS, e.g. the network stack) to new FreeBSD

> >>>> baseline.

> >>>>

> >>>> The patch set does two things:

> >>>>

> >>>> 1. It updates several header files imported from FreeBSD using the

> >>>> commits from

> >>>>     FreeBSD.

> >>>>

> >>>> 2. It removes the kernel space content from some files and includes a

> >>>>     <machine/_kernel_*.h> file instead.  This allows to use the Newlib

> >>>> provided

> >>>>     header files with different FreeBSD baselines.

> >>>

> >>>

> >>> This series is full of poorly-formatted commit messages:

> >>>

> >>>> ae@FreeBSD.org (1):

> >>>>    Follow the RFC6980 and silently ignore following IPv6 NDP messages

> >>>>      that had the IPv6 fragmentation header: o Neighbor Solicitation o

> >>>>      Neighbor Advertisement o Router Solicitation o Router Advertisement

> >>>>      o Redirect

> >>>

> >>>

> >>>> glebius (2):

> >>>

> >>> ...

> >>>>

> >>>>    Garbage collect IFCAP_POLLING_NOCOUNT. It wasn't used since very

> >>>>      beginning of polling(4). The module always ignored return value from

> >>>>      driver polling handler.

> >>>>

> >>>

> >>>> jhb (4):

> >>>>    Add a mmap flag (MAP_32BIT) on 64-bit platforms to request that a

> >>>>      mapping use an address in the first 2GB of the process's address

> >>>>      space. This flag should have the same semantics as the same flag on

> >>>>      Linux.

> >>>>    Add a new fo_fill_kinfo fileops method to add type-specific

> >>>>      information to struct kinfo_file. - Move the various fill_*_info()

> >>>>      methods out of kern_descrip.c and into the various file type

> >>>>      implementations. - Rework the support for kinfo_ofile to generate a

> >>>>      suitable kinfo_file object for each file and then convert that to a

> >>>>      kinfo_ofile structure rather than keeping a second, different set of

> >>>>      code that directly manipulates type-specific file information. -

> >>>>      Remove the shm_path() and ksem_info() layering violations.

> >>>>    Retire the unimplemented MAP_RENAME and MAP_NORESERVE flags to

> >>>>      mmap(2). Older binaries are still permitted to use these flags.

> >>>>    Add a new file operations hook for mmap operations. File type-specific

> >>>>      logic is now placed in the mmap hook implementation rather than

> >>>>      requiring it to be placed in sys/vm/vm_mmap.c. This hook allows new

> >>>>      file types to support mmap() as well as potentially allowing mmap()

> >>>>      for existing file types that do not currently support any mapping.

> >>>>

> >>>> jtl (1):

> >>>>    Add the "TCP Blackbox Recorder" which we discussed at the developer

> >>>>      summits at BSDCan and BSDCam in 2017.

> >>>>

> >>>> kib (8):

> >>>>    Implement read(2)/write(2) and neccessary lseek(2) for posix shmfd.

> >>>>      Add MAC framework entries for posix shm read and write.

> >>>>    Add MAP_EXCL flag for mmap(2). It should be combined with MAP_FIXED,

> >>>>      and prevents the request from deleting existing mappings in the

> >>>>      region, failing instead.

> >>>>    Implement process-shared locks support for libthr.so.3, without

> >>>>      breaking the ABI. Special value is stored in the lock pointer to

> >>>>      indicate shared lock, and offline page in the shared memory is

> >>>>      allocated to store the actual lock.

> >>>

> >>> ...

> >>>>

> >>>>    Add a place for a driver to report rx timestamps in nanoseconds from

> >>>>      boot for the received packets.

> >>>>    Use hardware timestamps to report packet timestamps for SO_TIMESTAMP

> >>>>      and other similar socket options.

> >>>

> >>>

> >>>>

> >>>> pkelsey (1):

> >>>>    This is an implementation of the client side of TCP Fast Open (TFO)

> >>>>      [RFC7413]. It also includes a pre-shared key mode of operation in

> >>>>      which the server requires the client to be in possession of a shared

> >>>>      secret in order to successfully open TFO connections with that

> >>>>      server.

> >>>>

> >>>> rrs (1):

> >>>>    This commit brings in a new refactored TCP stack called Rack. Rack

> >>>>      includes the following features: - A different SACK processing

> >>>>      scheme (the old sack structures are not used). - RACK (Recent

> >>>>      acknowledgment) where counting dup-acks is no longer done instead

> >>>>      time is used to knwo when to retransmit. (see the I-D) - TLP (Tail

> >>>>      Loss Probe) where we will probe for tail-losses to attempt to try

> >>>>      not to take a retransmit time-out. (see the I-D) - Burst mitigation

> >>>>      using TCPHTPS - PRR (partial rate reduction) see the RFC.

> >>>

> >>>

> >>> Please reformat the commit messages to have a SHORT summary (60-70

> >>> characters), then a blank line, then the rest of the description, rather

> >>> than diving right in to description that git then uses as the overlong

> >>> subject line.

> >>>

> >>

> >> I would like to keep the messages as is. This makes it easier to compare

> >> upstream FreeBSD changes with changes imported to Newlib.

> >>

> > 

> > In this case, since the commits come from a third party repository, we

> > would like to request inclusion of the commits in newlib without

> > modification. This allows us to more easily update to future versions

> > of the third party code, and provides a cleaner history/provenance of

> > the code.

> > 

> 

> Surely *adding* one extra line (followed by one blank line) as a summary

> at the head of the commit won't really interfere with such tracking.

> We're not asking you to reformat any of the remaining text.

> 

> R.


Right.  I fully agree with Eric and Richard.

If you really need to keep track of the upstream patches, add a
"cherry-picked from" line and use the SHA-1 for tracking.  That's much
better and easier than tracking via full text.


Corinna

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat