463,284 commits found in 112 milliseconds
Enforce locking order lfs_iflock > lfs_lock, preventing a deadlock between
LFS_*ENTRY() and lfs_prelock. This may bear on PR kern/60811 .
tools/compat: remove obsolete requirement about ncurses-devel on Linux
texinfo/info required it, but it has no longer been built during
tools build in src/tools/texinfo since texinfo 4.8a import, by
patching SUBDIRS lines in external/gpl2/texinfo/dist/Makefile.in.
Zero the entire inode block, not just the inode numbers. Fixes a kernel
memory data leak reported in kern/60803 . Fix confirmed by submitter.
new versions
update to latest (no update only date)
NetBSD /src C22sy5M — distrib/evbsh3/rom/ramdiskcommon ramdiskbin.conf, distrib/utils README distrib: update per changes I made 24 years ago
I removed init_s and x_dd on 2002-10-17, so remove the documentation
and the last references.
bump GCC version.
regen.
dwc2: Minimal change to avoid shifty UB.
PR port-mips/60780 : UBSan complains about left-shifting outside of int
type range
Add 5300* to the RADEON_RX5500 entry. Add some DASH serial ports.
update GCC entry to note 14.4.0 is in-tree now.
sysvbfs: Fix mistake in previous ctime fix.
Third time's the charm, right?
PR kern/60800 : missing ctime updates
NetBSD /src Hq8jBFy — external/gpl3/gcc/usr.bin/gcc/arch/powerpc plugin-version.h, external/gpl3/gcc/usr.bin/gcc/arch/powerpc64 plugin-version.h regen mknative-gcc for GCC 14.4.0.
msdosfs: Stop confusing birthtime and ctime.
PR kern/60801 : confusion of ctime with birthtime
hfs: Stop confusing birthtime and ctime.
XXX Need some automatic tests for hfs!
PR kern/60801 : confusion of ctime with birthtime
sysvbfs, v7fs: More ctime fixes.
PR kern/60800 : missing ctime updates
t_vnops: Verify ctime changes but birthtime doesn't.
PR kern/60800 : missing ctime updates
PR kern/60801 : confusion of ctime with birthtime
regenerate for GCC 14.4.0.
NetBSD /src zkJzcIO — external/gpl3/gcc/dist MD5SUMS NEWS, external/gpl3/gcc/dist/INSTALL configure.html merge GCC 14.4.
Version Delta File 1.20 +1,926 -1,922 external/gpl3/gcc/dist/libgomp/libgomp.info 1.23 +1,434 -1,430 external/gpl3/gcc/dist/gcc/doc/gccinstall.info 1.29 +2,400 -0 external/gpl3/gcc/dist/gcc/ChangeLog 1.20 +1,054 -1,081 external/gpl3/gcc/dist/INSTALL/configure.html 1.21 +1,074 -1,060 external/gpl3/gcc/dist/NEWS 1.23 +1,010 -550 external/gpl3/gcc/dist/MD5SUMS +8,898 -6,043 51 files not shown +11,531 -9,153 57 files
NetBSD /src Us2gtpl — external/gpl3/gcc/dist/INSTALL configure.html, external/gpl3/gcc/dist/gcc ChangeLog initial import of GCC 14.4.0 sources.
changes include fixes for at least 113 GCC bugs. the ones found
to be interesting include:
122943 - Wrong code for switch on int64_t on ilp32 targets at -O2 since r14-8302
123271 - -ftrapv fails to detect overflow on ARM32 since GCC 10 since r10-3995
and there are a lot of less scary ICE and runtime fixes.
mknative-gcc coming up next.
Version Delta File 1.1.1.24 +10,071 -10,137 external/gpl3/gcc/dist/gcc/doc/gcc.1 1.1.1.24 +10,071 -10,137 external/gpl3/gcc/dist/gcc/doc/g++.1 1.1.1.23 +1,926 -1,922 external/gpl3/gcc/dist/libgomp/libgomp.info 1.1.1.24 +1,434 -1,430 external/gpl3/gcc/dist/gcc/doc/gccinstall.info 1.1.1.27 +2,400 -0 external/gpl3/gcc/dist/gcc/ChangeLog 1.1.1.23 +1,054 -1,081 external/gpl3/gcc/dist/INSTALL/configure.html +26,956 -24,707 391 files not shown +43,288 -33,882 397 files
Also build aligned_alloc.c if USE_JEMALLOC is no (i.e. using phkmalloc).
dtfs: Update ctime on utimes(2) .
PR kern/60800 : tmpfs: missing ctime updates
v7fs: Update ctime on utimes(2) .
PR kern/60800 : tmpfs: missing ctime updates
sysvbfs: Update ctime on utimes(2) .
PR kern/60800 : tmpfs: missing ctime updates
tmpfs: Update ctime on utimes(2) .
PR kern/60800 : tmpfs: missing ctime updates
t_vnops: Verify that utimes updates ctime, if applicable.
PR kern/60800 : tmpfs: missing ctime updates
switch everyone to jemalloc-5.4.0 for testing
Fix build.
Handle two causes of the panic "lfs_ifile: looping". Repurpose the lfs_iflock
to be locked with RW_WRITER while the segment lock is held, preventing
updates to the Ifile while its blocks are being gathered; and allow more
iterations than 10 before panicking, since we might need to rewrite the
entire segment table one block per iteration. Issue uncovered in the course
of PR kern/60746 .
re(4) : handle errors while initializing the RX ring
Properly propagate errors from re_newbuf() while initializing the RX
descriptor ring.