tjh.dev / kernel
Linux kernel mirror (for testing) git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel os linux
1
fork

Configure Feed

Select the types of activity you want to include in your feed.

setlocalversion: work around "git describe" performance

Contrary to expectations, passing a single candidate tag to "git
describe" is slower than not passing any --match options.

$ time git describe --debug
...
traversed 10619 commits
...
v6.12-rc5-63-g0fc810ae3ae1

real 0m0.169s

$ time git describe --match=v6.12-rc5 --debug
...
traversed 1310024 commits
v6.12-rc5-63-g0fc810ae3ae1

real 0m1.281s

In fact, the --debug output shows that git traverses all or most of
history. For some repositories and/or git versions, those 1.3s are
actually 10-15 seconds.

This has been acknowledged as a performance bug in git [1], and a fix
is on its way [2]. However, no solution is yet in git.git, and even
when one lands, it will take quite a while before it finds its way to
a release and for $random_kernel_developer to pick that up.

So rewrite the logic to use plumbing commands. For each of the
candidate values of $tag, we ask: (1) is $tag even an annotated
tag? (2) Is it eligible to describe HEAD, i.e. an ancestor of
HEAD? (3) If so, how many commits are in $tag..HEAD?

I have tested that this produces the same output as the current script
for ~700 random commits between v6.9..v6.10. For those 700 commits,
and in my git repo, the 'make -s kernelrelease' command is on average
~4 times faster with this patch applied (geometric mean of ratios).

For the commit mentioned in Josh's original report [3], the
time-consuming part of setlocalversion goes from

$ time git describe --match=v6.12-rc5 c1e939a21eb1
v6.12-rc5-44-gc1e939a21eb1

real 0m1.210s

to

$ time git rev-list --count --left-right v6.12-rc5..c1e939a21eb1
0 44

real 0m0.037s

[1] https://lore.kernel.org/git/20241101113910.GA2301440@coredump.intra.peff.net/
[2] https://lore.kernel.org/git/20241106192236.GC880133@coredump.intra.peff.net/
[3] https://lore.kernel.org/lkml/309549cafdcfe50c4fceac3263220cc3d8b109b2.1730337435.git.jpoimboe@kernel.org/

Reported-by: Sean Christopherson <seanjc@google.com>
Closes: https://lore.kernel.org/lkml/ZPtlxmdIJXOe0sEy@google.com/
Reported-by: Josh Poimboeuf <jpoimboe@kernel.org>
Closes: https://lore.kernel.org/lkml/309549cafdcfe50c4fceac3263220cc3d8b109b2.1730337435.git.jpoimboe@kernel.org/
Tested-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

authored by

Rasmus Villemoes and committed by
Masahiro Yamada 523f3dbc e397a603

+38 -16
+38 -16
scripts/setlocalversion
··· 30 30 usage 31 31 fi 32 32 33 + try_tag() { 34 + tag="$1" 35 + 36 + # Is $tag an annotated tag? 37 + [ "$(git cat-file -t "$tag" 2> /dev/null)" = tag ] || return 1 38 + 39 + # Is it an ancestor of HEAD, and if so, how many commits are in $tag..HEAD? 40 + # shellcheck disable=SC2046 # word splitting is the point here 41 + set -- $(git rev-list --count --left-right "$tag"...HEAD 2> /dev/null) 42 + 43 + # $1 is 0 if and only if $tag is an ancestor of HEAD. Use 44 + # string comparison, because $1 is empty if the 'git rev-list' 45 + # command somehow failed. 46 + [ "$1" = 0 ] || return 1 47 + 48 + # $2 is the number of commits in the range $tag..HEAD, possibly 0. 49 + count="$2" 50 + 51 + return 0 52 + } 53 + 33 54 scm_version() 34 55 { 35 56 local short=false ··· 82 61 # stable kernel: 6.1.7 -> v6.1.7 83 62 version_tag=v$(echo "${KERNELVERSION}" | sed -E 's/^([0-9]+\.[0-9]+)\.0(.*)$/\1\2/') 84 63 64 + # try_tag initializes count if the tag is usable. 65 + count= 66 + 85 67 # If a localversion* file exists, and the corresponding 86 68 # annotated tag exists and is an ancestor of HEAD, use 87 69 # it. This is the case in linux-next. 88 - tag=${file_localversion#-} 89 - desc= 90 - if [ -n "${tag}" ]; then 91 - desc=$(git describe --match=$tag 2>/dev/null) 70 + if [ -n "${file_localversion#-}" ] ; then 71 + try_tag "${file_localversion#-}" 92 72 fi 93 73 94 74 # Otherwise, if a localversion* file exists, and the tag 95 75 # obtained by appending it to the tag derived from 96 76 # KERNELVERSION exists and is an ancestor of HEAD, use 97 77 # it. This is e.g. the case in linux-rt. 98 - if [ -z "${desc}" ] && [ -n "${file_localversion}" ]; then 99 - tag="${version_tag}${file_localversion}" 100 - desc=$(git describe --match=$tag 2>/dev/null) 78 + if [ -z "${count}" ] && [ -n "${file_localversion}" ]; then 79 + try_tag "${version_tag}${file_localversion}" 101 80 fi 102 81 103 82 # Otherwise, default to the annotated tag derived from KERNELVERSION. 104 - if [ -z "${desc}" ]; then 105 - tag="${version_tag}" 106 - desc=$(git describe --match=$tag 2>/dev/null) 83 + if [ -z "${count}" ]; then 84 + try_tag "${version_tag}" 107 85 fi 108 86 109 - # If we are at the tagged commit, we ignore it because the version is 110 - # well-defined. 111 - if [ "${tag}" != "${desc}" ]; then 87 + # If we are at the tagged commit, we ignore it because the 88 + # version is well-defined. If none of the attempted tags exist 89 + # or were usable, $count is still empty. 90 + if [ -z "${count}" ] || [ "${count}" -gt 0 ]; then 112 91 113 92 # If only the short version is requested, don't bother 114 93 # running further git commands ··· 116 95 echo "+" 117 96 return 118 97 fi 98 + 119 99 # If we are past the tagged commit, we pretty print it. 120 100 # (like 6.1.0-14595-g292a089d78d3) 121 - if [ -n "${desc}" ]; then 122 - echo "${desc}" | awk -F- '{printf("-%05d", $(NF-1))}' 101 + if [ -n "${count}" ]; then 102 + printf "%s%05d" "-" "${count}" 123 103 fi 124 104 125 105 # Add -g and exactly 12 hex chars. 126 - printf '%s%s' -g "$(echo $head | cut -c1-12)" 106 + printf '%s%.12s' -g "$head" 127 107 fi 128 108 129 109 if ${no_dirty}; then