NetBSD/src AcVJve2 — sys/dev/scsipi sd.c

   Re-add fallback to READ CAPACITY 16 if size was truncated.
   Should fix PR 60933.
VersionDeltaFile
1.348+18-3sys/dev/scsipi/sd.c
+18-31 files

NetBSD/src hJecom8 — sys/arch/m68k/include pmap_68k.h, sys/arch/m68k/m68k pmap_68k.c

   Add infrastructure for re-loading the %urp on remote CPUs when a
   pmap gets a new lev1map.
VersionDeltaFile
1.110+60-17sys/arch/m68k/m68k/pmap_68k.c
1.33+2-1sys/arch/m68k/include/pmap_68k.h
+62-182 files

NetBSD/src zpKULoX — sys/arch/arm/sunxi sun60i_a733_rtc_ccu.h sun60i_a733_r_ccu.h

   sunxi: Rename SOC_SUN60I_W2 -> SOC_SUN60I_A733

   Apparently I forgot my own naming convention for these files.
VersionDeltaFile
1.1+267-0sys/arch/arm/sunxi/sun60i_a733_r_ccu.c
1.1+264-0sys/arch/arm/sunxi/sun60i_a733_gpio.c
1.1+168-0sys/arch/arm/sunxi/sun60i_a733_ppu.c
1.1+142-0sys/arch/arm/sunxi/sun60i_a733_rtc_ccu.c
1.1+84-0sys/arch/arm/sunxi/sun60i_a733_r_ccu.h
1.1+46-0sys/arch/arm/sunxi/sun60i_a733_rtc_ccu.h
+971-010 files not shown
+1,010-3916 files

NetBSD/src RFtS7XX — usr.bin/cmp regular.c

   Account for bytes when the block matches when using cmp -l.

   Otherwize the byte offset displayed by cmp -l shows the wrong byte number.

   XXX Pullup-11.
VersionDeltaFile
1.27+3-2usr.bin/cmp/regular.c
+3-21 files

NetBSD/src 4JCGhKI — sys/arch/arm/sunxi files.sunxi sun60i_w2_r_ccu.h, sys/arch/evbarm/conf GENERIC64

   sunxi: Add support for Allwinner A733 (sun60i-w2) PRCM clocks.

   Compile tested only at this point.
VersionDeltaFile
1.1+267-0sys/arch/arm/sunxi/sun60i_w2_r_ccu.c
1.1+84-0sys/arch/arm/sunxi/sun60i_w2_r_ccu.h
1.78+6-1sys/arch/arm/sunxi/files.sunxi
1.236+2-1sys/arch/evbarm/conf/GENERIC64
+359-24 files

NetBSD/src OKmH76O — sys/arch/arm/sunxi sunxi_ccu.h sunxi_ccu_nm.c

   sunxi: Add support for "m" only clock types.
VersionDeltaFile
1.5+12-6sys/arch/arm/sunxi/sunxi_ccu_nm.c
1.27+7-1sys/arch/arm/sunxi/sunxi_ccu.h
+19-72 files

NetBSD/src FyhTs3D — doc CHANGES-10.3

   Ticket #1423
VersionDeltaFile
1.1.2.31+6-1doc/CHANGES-10.3
+6-11 files

NetBSD/src 8xL35Ic — external/bsd/tmux/usr.bin/tmux Makefile

   Pull up following revision(s) (requested by wiz in ticket #1423):

        external/bsd/tmux/usr.bin/tmux/Makefile: revision 1.47 (patch)

   tmux: fix TMUX_CONF
   Addresses PR 61034.
VersionDeltaFile
1.28.2.4+2-2external/bsd/tmux/usr.bin/tmux/Makefile
+2-21 files

NetBSD/src 1EABlsZ — doc CHANGES-11.2

   Ticket 584
VersionDeltaFile
1.1.2.4+6-1doc/CHANGES-11.2
+6-11 files

NetBSD/src YwapXbl — external/bsd/tmux/usr.bin/tmux Makefile

   Pull up following revision(s) (requested by wiz in ticket #584):

        external/bsd/tmux/usr.bin/tmux/Makefile: revision 1.47

   tmux: fix TMUX_CONF
   Addresses PR 61034.
VersionDeltaFile
1.38.2.3+2-2external/bsd/tmux/usr.bin/tmux/Makefile
+2-21 files

NetBSD/src 0iFTtPQ — sys/arch/arm/sunxi files.sunxi sun60i_w2_rtc_ccu.h, sys/arch/evbarm/conf GENERIC64

   sunxi: Add support for Allwinner A733 (sun60i-w2) RTC clock provider.

   Compile tested only at this point.
VersionDeltaFile
1.1+142-0sys/arch/arm/sunxi/sun60i_w2_rtc_ccu.c
1.1+46-0sys/arch/arm/sunxi/sun60i_w2_rtc_ccu.h
1.77+6-1sys/arch/arm/sunxi/files.sunxi
1.235+2-1sys/arch/evbarm/conf/GENERIC64
+196-24 files

NetBSD/src qymCE1d — sys/arch/arm/sunxi sunxi_ccu_gate.c sunxi_ccu.h

   sunxi: Allow gate clocks to specify a fixed output rate
VersionDeltaFile
1.26+18-11sys/arch/arm/sunxi/sunxi_ccu.h
1.3+23-2sys/arch/arm/sunxi/sunxi_ccu_gate.c
+41-132 files

NetBSD/src DVHLxcA — sys/kern vfs_wapbl.c

   The evaluation of resource shortage in wapbl_register_deallocation()
   is incomplete as only "wl_dealloccnt" is checked.  Replace this check with
   the doflush check from wapbl_begin() testing for 75% instead 50% to
   make sure the next wapbl_begin() will always flush.

   It is still a mitigation, for a fix the file system must provide an
   upper bound for every transaction to wapbl_begin().

   PR kern/47146 user-triggerable crash in wapbl_register_deallocation (confi)
   PR kern/60621 (panic: ufs_inactive: dirty filesystem?)
VersionDeltaFile
1.118+26-17sys/kern/vfs_wapbl.c
+26-171 files

NetBSD/src SUPT9mY — sys/arch/m68k/m68k pmap_68k.c

   Update comment describing the serialization mechanisms at play here.
VersionDeltaFile
1.109+69-56sys/arch/m68k/m68k/pmap_68k.c
+69-561 files

NetBSD/src LQEzj3N — sys/arch/m68k/include pmap_68k.h, sys/arch/m68k/m68k pmap_68k.c

   Implement a locking protocol for the 68k pmap structures.  This essentially
   replaces the old PMAP_CRIT_{ENTER,EXIT}() markers with PMAP_{LOCK,UNLOCK}()
   calls.

   On systems that aren't built for MP or for pmap lock debugging, this is
   all a no-op as before.
VersionDeltaFile
1.108+109-129sys/arch/m68k/m68k/pmap_68k.c
1.32+23-23sys/arch/m68k/include/pmap_68k.h
+132-1522 files

NetBSD/src ByJJZMw — sys/arch/m68k/m68k pmap_68k.c

   Document our locking rules re. PTE modification.  Make pmap stats updates
   MP-safe.
VersionDeltaFile
1.107+38-5sys/arch/m68k/m68k/pmap_68k.c
+38-51 files

NetBSD/src EFMw9kP — sys/arch/m68k/include pmap_68k.h, sys/arch/m68k/m68k pmap_68k.c

   Introduce a locking protocol around the PV lists.
VersionDeltaFile
1.106+76-55sys/arch/m68k/m68k/pmap_68k.c
1.31+38-1sys/arch/m68k/include/pmap_68k.h
+114-562 files

NetBSD/src SrSItcP — sys/arch/m68k/m68k pmap_68k.c

   Update the cached U/M attributes in pmap_pv_remove(), not all of the
   callers of pmap_pv_remove().
VersionDeltaFile
1.105+9-10sys/arch/m68k/m68k/pmap_68k.c
+9-101 files

NetBSD/src 5ab8ckM — sys/arch/m68k/m68k pmap_68k.c

   pmap_enter(): collect the garbage correctly in the "returning an error" case.
   pmap_activate(): KASSERT that there are no pending ATC invalidations.
VersionDeltaFile
1.104+5-3sys/arch/m68k/m68k/pmap_68k.c
+5-31 files

NetBSD/src LcRPsl8 — sys/arch/m68k/m68k pmap_68k.c

   pmap_atc_process_hp(): Add instrumentation.
VersionDeltaFile
1.103+29-3sys/arch/m68k/m68k/pmap_68k.c
+29-31 files

NetBSD/src lwQaoPl — sys/ufs/ffs ffs_snapshot.c

   Add UFS_WAPBL_UPDATE() after block allocations to group the
   allocation and the inode update into the same transaction.

   PR kern/60621 (panic: ufs_inactive: dirty filesystem?)
VersionDeltaFile
1.158+4-2sys/ufs/ffs/ffs_snapshot.c
+4-21 files

NetBSD/src K4m1HVi — sys/arch/m68k/include cpu.h

   Clarify a comment and remove a no-longer-relevant prototype.
VersionDeltaFile
1.37+2-3sys/arch/m68k/include/cpu.h
+2-31 files

NetBSD/src Z7Wk8Ss — sys/dev/nvmm nvmm.c

   nvmm: don't hardcode cmajor
VersionDeltaFile
1.51+3-3sys/dev/nvmm/nvmm.c
+3-31 files

NetBSD/src XioJSCN — common/lib/libc/arch/m68k/atomic Makefile.inc membar_ops.S

   The 68010, 68020, and 68030 are have strictly ordered memory operations.
   The 68040 and 68060 do not, however -- stores can be observed after
   subsequent loads.

   This is to say: all loads are load-acquire, all stores are store-release,
   but store-before-load must be explicitly serialized.  The 68040 User's
   Manual states that a NOP insn forces this serialization.

   And so: membar_aquire -> do nothing, membar_release -> do nothing,
   membar_sync -> do nothing *with feeling*

   Chapter and verse of the manual quoted in comments.
VersionDeltaFile
1.1+111-0common/lib/libc/arch/m68k/atomic/membar_ops.S
1.19+3-4common/lib/libc/arch/m68k/atomic/Makefile.inc
+114-42 files

NetBSD/src RMC4w8a — external/bsd/tmux/usr.bin/tmux Makefile

   tmux: fix TMUX_CONF

   Addresses PR 61034.
VersionDeltaFile
1.47+2-2external/bsd/tmux/usr.bin/tmux/Makefile
+2-21 files

NetBSD/src wz86sD9 — sys/arch/m68k/m68k pmap_68k.c

   The MMU-specific functions that load the user root pointer already take
   care of any necessary ATC / cache invalidations, so we don't need to do
   it again here.
VersionDeltaFile
1.102+4-9sys/arch/m68k/m68k/pmap_68k.c
+4-91 files

NetBSD/src xk8OlgI — sys/arch/m68k/m68k mmu_subr.s

   mmu_load_urp51(): We're switching the user address space by loading
   %crp, so only invalidate user ATC entries, not the entire ATC.
VersionDeltaFile
1.8+3-3sys/arch/m68k/m68k/mmu_subr.s
+3-31 files

NetBSD/src jVKNnTQ — sys/arch/m68k/m68k mmu_subr.s

   mmu_load_urp20hp(): No need to explicitly invalidate the external VAC
   or the MMU's TLB -- user VAC entries and user TLB entries are automatically
   invalidated when the USTP register is written, and there's no need to
   invalidate supervisor entries in either of those when switching the user
   address space.
VersionDeltaFile
1.7+4-9sys/arch/m68k/m68k/mmu_subr.s
+4-91 files

NetBSD/src DQdSzJC — sys/dev/acpi pckbc_acpi.c

   pckbc(4) at acpi: Fix build with -Os.

   Fixes:

   /home/riastradh/netbsd/current/src/sys/dev/acpi/pckbc_acpi.c: In function 'pckbc_acpi_intr_establish':
   /home/riastradh/netbsd/current/src/sys/dev/acpi/pckbc_acpi.c:242:63: error: 'handle' may be used uninitialized [-Werror=maybe-uninitialized]
     242 |                 rv = acpi_intr_establish(sc->sc_dv, (uint64_t)(uintptr_t)handle,
         |                                                               ^~~~~~~~~~~~~~~~~
   /home/riastradh/netbsd/current/src/sys/dev/acpi/pckbc_acpi.c:224:21: note: 'handle' was declared here
     224 |         ACPI_HANDLE handle;
         |                     ^~~~~~

   No functional change intended -- the code wasn't _wrong_ before; this
   just makes it clearer, with less analysis effort, that handle is
   initialized in the path in question.

   While here, rename local variable `rv' to `ih' to make it clearer
   what it is (not just some arbitrary return value of a subroutine but
   specifically an interrupt handler cookie).

    [3 lines not shown]
VersionDeltaFile
1.41+9-9sys/dev/acpi/pckbc_acpi.c
+9-91 files

NetBSD/src be2gn3U — usr.sbin/cpuctl/arch aarch64.c

   cpuctl: aarch64: name ID_AA64DFR0_EL1 correctly in a comment
VersionDeltaFile
1.32+3-3usr.sbin/cpuctl/arch/aarch64.c
+3-31 files