diff --git a/list-objects-filter-options.c b/list-objects-filter-options.c
index 7c3e397d29..75d0236ee2 100644
--- a/list-objects-filter-options.c
+++ b/list-objects-filter-options.c
@@ -6,6 +6,12 @@
 #include "list-objects.h"
 #include "list-objects-filter.h"
 #include "list-objects-filter-options.h"
+#include "url.h"
+
+static int parse_combine_filter(
+	struct list_objects_filter_options *filter_options,
+	const char *arg,
+	struct strbuf *errbuf);
 
 /*
  * Parse value of the argument to the "filter" keyword.
@@ -35,8 +41,6 @@ static int gently_parse_list_objects_filter(
 		return 1;
 	}
 
-	filter_options->filter_spec = strdup(arg);
-
 	if (!strcmp(arg, "blob:none")) {
 		filter_options->choice = LOFC_BLOB_NONE;
 		return 0;
@@ -77,6 +81,10 @@ static int gently_parse_list_objects_filter(
 				_("sparse:path filters support has been dropped"));
 		}
 		return 1;
+
+	} else if (skip_prefix(arg, "combine:", &v0)) {
+		return parse_combine_filter(filter_options, v0, errbuf);
+
 	}
 	/*
 	 * Please update _git_fetch() in git-completion.bash when you
@@ -89,10 +97,95 @@ static int gently_parse_list_objects_filter(
 	return 1;
 }
 
+static const char *RESERVED_NON_WS = "~`!@#$^&*()[]{}\\;'\",<>?";
+
+static int has_reserved_character(
+	struct strbuf *sub_spec, struct strbuf *errbuf)
+{
+	const char *c = sub_spec->buf;
+	while (*c) {
+		if (*c <= ' ' || strchr(RESERVED_NON_WS, *c)) {
+			strbuf_addf(
+				errbuf,
+				_("must escape char in sub-filter-spec: '%c'"),
+				*c);
+			return 1;
+		}
+		c++;
+	}
+
+	return 0;
+}
+
+static int parse_combine_subfilter(
+	struct list_objects_filter_options *filter_options,
+	struct strbuf *subspec,
+	struct strbuf *errbuf)
+{
+	size_t new_index = filter_options->sub_nr++;
+	char *decoded;
+	int result;
+
+	ALLOC_GROW(filter_options->sub, filter_options->sub_nr,
+		   filter_options->sub_alloc);
+	memset(&filter_options->sub[new_index], 0,
+	       sizeof(*filter_options->sub));
+
+	decoded = url_percent_decode(subspec->buf);
+
+	result = has_reserved_character(subspec, errbuf) ||
+		gently_parse_list_objects_filter(
+			&filter_options->sub[new_index], decoded, errbuf);
+
+	free(decoded);
+	return result;
+}
+
+static int parse_combine_filter(
+	struct list_objects_filter_options *filter_options,
+	const char *arg,
+	struct strbuf *errbuf)
+{
+	struct strbuf **subspecs = strbuf_split_str(arg, '+', 0);
+	size_t sub;
+	int result = 0;
+
+	if (!subspecs[0]) {
+		strbuf_addstr(errbuf, _("expected something after combine:"));
+		result = 1;
+		goto cleanup;
+	}
+
+	for (sub = 0; subspecs[sub] && !result; sub++) {
+		if (subspecs[sub + 1]) {
+			/*
+			 * This is not the last subspec. Remove trailing "+" so
+			 * we can parse it.
+			 */
+			size_t last = subspecs[sub]->len - 1;
+			assert(subspecs[sub]->buf[last] == '+');
+			strbuf_remove(subspecs[sub], last, 1);
+		}
+		result = parse_combine_subfilter(
+			filter_options, subspecs[sub], errbuf);
+	}
+
+	filter_options->choice = LOFC_COMBINE;
+
+cleanup:
+	strbuf_list_free(subspecs);
+	if (result) {
+		list_objects_filter_release(filter_options);
+		memset(filter_options, 0, sizeof(*filter_options));
+	}
+	return result;
+}
+
 int parse_list_objects_filter(struct list_objects_filter_options *filter_options,
 			      const char *arg)
 {
 	struct strbuf buf = STRBUF_INIT;
+	filter_options->filter_spec = strdup(arg);
 	if (gently_parse_list_objects_filter(filter_options, arg, &buf))
 		die("%s", buf.buf);
 	return 0;
@@ -129,8 +222,15 @@ void expand_list_objects_filter_spec(
 void list_objects_filter_release(
 	struct list_objects_filter_options *filter_options)
 {
+	size_t sub;
+
+	if (!filter_options)
+		return;
 	free(filter_options->filter_spec);
 	free(filter_options->sparse_oid_value);
+	for (sub = 0; sub < filter_options->sub_nr; sub++)
+		list_objects_filter_release(&filter_options->sub[sub]);
+	free(filter_options->sub);
 	memset(filter_options, 0, sizeof(*filter_options));
 }
 
@@ -174,6 +274,8 @@ void partial_clone_get_default_filter_spec(
 	 */
 	if (!core_partial_clone_filter_default)
 		return;
+
+	filter_options->filter_spec = strdup(core_partial_clone_filter_default);
 	gently_parse_list_objects_filter(filter_options,
 					 core_partial_clone_filter_default,
 					 &errbuf);
diff --git a/list-objects-filter-options.h b/list-objects-filter-options.h
index c54f0000fb..789faef1e5 100644
--- a/list-objects-filter-options.h
+++ b/list-objects-filter-options.h
@@ -13,6 +13,7 @@ enum list_objects_filter_choice {
 	LOFC_BLOB_LIMIT,
 	LOFC_TREE_DEPTH,
 	LOFC_SPARSE_OID,
+	LOFC_COMBINE,
 	LOFC__COUNT /* must be last */
 };
 
@@ -38,13 +39,23 @@ struct list_objects_filter_options {
 	unsigned int no_filter : 1;
 
 	/*
-	 * Parsed values (fields) from within the filter-spec.  These are
-	 * choice-specific; not all values will be defined for any given
-	 * choice.
+	 * BEGIN choice-specific parsed values from within the filter-spec. Only
+	 * some values will be defined for any given choice.
 	 */
+
 	struct object_id *sparse_oid_value;
 	unsigned long blob_limit_value;
 	unsigned long tree_exclude_depth;
+
+	/* LOFC_COMBINE values */
+
+	/* This array contains all the subfilters which this filter combines. */
+	size_t sub_nr, sub_alloc;
+	struct list_objects_filter_options *sub;
+
+	/*
+	 * END choice-specific parsed values.
+	 */
 };
 
 /* Normalized command line arguments */
diff --git a/list-objects-filter.c b/list-objects-filter.c
index 3b4b6764ca..d664264d65 100644
--- a/list-objects-filter.c
+++ b/list-objects-filter.c
@@ -26,6 +26,14 @@
  */
 #define FILTER_SHOWN_BUT_REVISIT (1<<21)
 
+struct subfilter {
+	struct filter *filter;
+	struct oidset seen;
+	struct oidset omits;
+	struct object_id skip_tree;
+	unsigned is_skipping_tree : 1;
+};
+
 struct filter {
 	enum list_objects_filter_result (*filter_object_fn)(
 		struct repository *r,
@@ -36,6 +44,23 @@ struct filter {
 		struct oidset *omits,
 		void *filter_data);
 
+	/*
+	 * Optional. If this function is supplied and the filter needs
+	 * to collect omits, then this function is called once before
+	 * free_fn is called.
+	 *
+	 * This is required because the following two conditions hold:
+	 *
+	 *   a. A tree filter can add and remove objects as an object
+	 *      graph is traversed.
+	 *   b. A combine filter's omit set is the union of all its
+	 *      subfilters, which may include tree: filters.
+	 *
+	 * As such, the omits sets must be separate sets, and can only
+	 * be unioned after the traversal is completed.
+	 */
+	void (*finalize_omits_fn)(struct oidset *omits, void *filter_data);
+
 	void (*free_fn)(void *filter_data);
 
 	void *filter_data;
@@ -471,6 +496,139 @@ static void filter_sparse_oid__init(
 	filter->free_fn = filter_sparse_free;
 }
 
+/* A filter which only shows objects shown by all sub-filters. */
+struct combine_filter_data {
+	struct subfilter *sub;
+	size_t nr;
+};
+
+static enum list_objects_filter_result process_subfilter(
+	struct repository *r,
+	enum list_objects_filter_situation filter_situation,
+	struct object *obj,
+	const char *pathname,
+	const char *filename,
+	struct subfilter *sub)
+{
+	enum list_objects_filter_result result;
+
+	/*
+	 * Check and update is_skipping_tree before oidset_contains so
+	 * that is_skipping_tree gets unset even when the object is
+	 * marked as seen.  As of this writing, no filter uses
+	 * LOFR_MARK_SEEN on trees that also uses LOFR_SKIP_TREE, so the
+	 * ordering is only theoretically important. Be cautious if you
+	 * change the order of the below checks and more filters have
+	 * been added!
+	 */
+	if (sub->is_skipping_tree) {
+		if (filter_situation == LOFS_END_TREE &&
+		    oideq(&obj->oid, &sub->skip_tree))
+			sub->is_skipping_tree = 0;
+		else
+			return LOFR_ZERO;
+	}
+	if (oidset_contains(&sub->seen, &obj->oid))
+		return LOFR_ZERO;
+
+	result = list_objects_filter__filter_object(
+		r, filter_situation, obj, pathname, filename, sub->filter);
+
+	if (result & LOFR_MARK_SEEN)
+		oidset_insert(&sub->seen, &obj->oid);
+
+	if (result & LOFR_SKIP_TREE) {
+		sub->is_skipping_tree = 1;
+		sub->skip_tree = obj->oid;
+	}
+
+	return result;
+}
+
+static enum list_objects_filter_result filter_combine(
+	struct repository *r,
+	enum list_objects_filter_situation filter_situation,
+	struct object *obj,
+	const char *pathname,
+	const char *filename,
+	struct oidset *omits,
+	void *filter_data)
+{
+	struct combine_filter_data *d = filter_data;
+	enum list_objects_filter_result combined_result =
+		LOFR_DO_SHOW | LOFR_MARK_SEEN | LOFR_SKIP_TREE;
+	size_t sub;
+
+	for (sub = 0; sub < d->nr; sub++) {
+		enum list_objects_filter_result sub_result = process_subfilter(
+			r, filter_situation, obj, pathname, filename,
+			&d->sub[sub]);
+		if (!(sub_result & LOFR_DO_SHOW))
+			combined_result &= ~LOFR_DO_SHOW;
+		if (!(sub_result & LOFR_MARK_SEEN))
+			combined_result &= ~LOFR_MARK_SEEN;
+		if (!d->sub[sub].is_skipping_tree)
+			combined_result &= ~LOFR_SKIP_TREE;
+	}
+
+	return combined_result;
+}
+
+static void filter_combine__free(void *filter_data)
+{
+	struct combine_filter_data *d = filter_data;
+	size_t sub;
+	for (sub = 0; sub < d->nr; sub++) {
+		list_objects_filter__free(d->sub[sub].filter);
+		oidset_clear(&d->sub[sub].seen);
+		if (d->sub[sub].omits.set.size)
+			BUG("expected oidset to be cleared already");
+	}
+	free(d->sub);
+}
+
+static void add_all(struct oidset *dest, struct oidset *src) {
+	struct oidset_iter iter;
+	struct object_id *src_oid;
+
+	oidset_iter_init(src, &iter);
+	while ((src_oid = oidset_iter_next(&iter)) != NULL)
+		oidset_insert(dest, src_oid);
+}
+
+static void filter_combine__finalize_omits(
+	struct oidset *omits,
+	void *filter_data)
+{
+	struct combine_filter_data *d = filter_data;
+	size_t sub;
+
+	for (sub = 0; sub < d->nr; sub++) {
+		add_all(omits, &d->sub[sub].omits);
+		oidset_clear(&d->sub[sub].omits);
+	}
+}
+
+static void filter_combine__init(
+	struct list_objects_filter_options *filter_options,
+	struct filter* filter)
+{
+	struct combine_filter_data *d = xcalloc(1, sizeof(*d));
+	size_t sub;
+
+	d->nr = filter_options->sub_nr;
+	d->sub = xcalloc(d->nr, sizeof(*d->sub));
+	for (sub = 0; sub < d->nr; sub++)
+		d->sub[sub].filter = list_objects_filter__init(
+			filter->omits ? &d->sub[sub].omits : NULL,
+			&filter_options->sub[sub]);
+
+	filter->filter_data = d;
+	filter->filter_object_fn = filter_combine;
+	filter->free_fn = filter_combine__free;
+	filter->finalize_omits_fn = filter_combine__finalize_omits;
+}
+
 typedef void (*filter_init_fn)(
 	struct list_objects_filter_options *filter_options,
 	struct filter *filter);
@@ -484,6 +642,7 @@ static filter_init_fn s_filters[] = {
 	filter_blobs_limit__init,
 	filter_trees_depth__init,
 	filter_sparse_oid__init,
+	filter_combine__init,
 };
 
 struct filter *list_objects_filter__init(
@@ -536,6 +695,8 @@ void list_objects_filter__free(struct filter *filter)
 {
 	if (!filter)
 		return;
+	if (filter->finalize_omits_fn && filter->omits)
+		filter->finalize_omits_fn(filter->omits, filter->filter_data);
 	filter->free_fn(filter->filter_data);
 	free(filter);
 }
diff --git a/list-objects-filter.h b/list-objects-filter.h
index 6908954266..cfd784e203 100644
--- a/list-objects-filter.h
+++ b/list-objects-filter.h
@@ -62,7 +62,13 @@ enum list_objects_filter_situation {
 
 struct filter;
 
-/* Constructor for the set of defined list-objects filters. */
+/*
+ * Constructor for the set of defined list-objects filters.
+ * The `omitted` set is optional. It is populated with objects that the
+ * filter excludes. This set should not be considered finalized until
+ * after list_objects_filter__free is called on the returned `struct
+ * filter *`.
+ */
 struct filter *list_objects_filter__init(
 	struct oidset *omitted,
 	struct list_objects_filter_options *filter_options);
@@ -80,7 +86,10 @@ enum list_objects_filter_result list_objects_filter__filter_object(
 	const char *filename,
 	struct filter *filter);
 
-/* Destroys `filter`. Does nothing if `filter` is null. */
+/*
+ * Destroys `filter` and finalizes the `omitted` set, if present. Does
+ * nothing if `filter` is null.
+ */
 void list_objects_filter__free(struct filter *filter);
 
 #endif /* LIST_OBJECTS_FILTER_H */
diff --git a/t/t6112-rev-list-filters-objects.sh b/t/t6112-rev-list-filters-objects.sh
index acd7f5ab80..05d4f2e9c2 100755
--- a/t/t6112-rev-list-filters-objects.sh
+++ b/t/t6112-rev-list-filters-objects.sh
@@ -278,7 +278,19 @@ test_expect_success 'verify skipping tree iteration when not collecting omits' '
 	test_line_count = 2 actual &&
 
 	# Make sure no other trees were considered besides the root.
-	! grep "Skipping contents of tree [^.]" filter_trace
+	! grep "Skipping contents of tree [^.]" filter_trace &&
+
+	# Try this again with "combine:". If both sub-filters are skipping
+	# trees, the composite filter should also skip trees. This is not
+	# important unless the user does combine:tree:X+tree:Y or another filter
+	# besides "tree:" is implemented in the future which can skip trees.
+	GIT_TRACE=1 git -C r3 rev-list \
+		--objects --filter=combine:tree:1+tree:3 HEAD 2>filter_trace &&
+
+	# Only skip the dir1/ tree, which is shared between the two commits.
+	grep "Skipping contents of tree " filter_trace >actual &&
+	test_write_lines "Skipping contents of tree dir1/..." >expected &&
+	test_cmp expected actual
 '
 
 # Test tree:# filters.
@@ -330,6 +342,112 @@ test_expect_success 'verify tree:3 includes everything expected' '
 	test_line_count = 10 actual
 '
 
+test_expect_success 'combine:... for a simple combination' '
+	git -C r3 rev-list --objects --filter=combine:tree:2+blob:none HEAD \
+		>actual &&
+
+	expect_has HEAD "" &&
+	expect_has HEAD~1 "" &&
+	expect_has HEAD dir1 &&
+
+	# There are also 2 commit objects
+	test_line_count = 5 actual
+'
+
+test_expect_success 'combine:... with URL encoding' '
+	git -C r3 rev-list --objects \
+		--filter=combine:tree%3a2+blob:%6Eon%65 HEAD >actual &&
+
+	expect_has HEAD "" &&
+	expect_has HEAD~1 "" &&
+	expect_has HEAD dir1 &&
+
+	# There are also 2 commit objects
+	test_line_count = 5 actual
+'
+
+expect_invalid_filter_spec () {
+	spec="$1" &&
+	err="$2" &&
+
+	test_must_fail git -C r3 rev-list --objects --filter="$spec" HEAD \
+		>actual 2>actual_stderr &&
+	test_must_be_empty actual &&
+	test_i18ngrep "$err" actual_stderr
+}
+
+test_expect_success 'combine:... while URL-encoding things that should not be' '
+	expect_invalid_filter_spec combine%3Atree:2+blob:none \
+		"invalid filter-spec"
+'
+
+test_expect_success 'combine: with nothing after the :' '
+	expect_invalid_filter_spec combine: "expected something after combine:"
+'
+
+test_expect_success 'parse error in first sub-filter in combine:' '
+	expect_invalid_filter_spec combine:tree:asdf+blob:none \
+		"expected .tree:<depth>."
+'
+
+test_expect_success 'combine:... with non-encoded reserved chars' '
+	expect_invalid_filter_spec combine:tree:2+sparse:@xyz \
+		"must escape char in sub-filter-spec: .@." &&
+	expect_invalid_filter_spec combine:tree:2+sparse:\` \
+		"must escape char in sub-filter-spec: .\`." &&
+	expect_invalid_filter_spec combine:tree:2+sparse:~abc \
+		"must escape char in sub-filter-spec: .\~."
+'
+
+test_expect_success 'validate err msg for "combine:<valid-filter>+"' '
+	expect_invalid_filter_spec combine:tree:2+ "expected .tree:<depth>."
+'
+
+test_expect_success 'combine:... with edge-case hex digits: Ff Aa 0 9' '
+	git -C r3 rev-list --objects --filter="combine:tree:2+bl%6Fb:n%6fne" \
+		HEAD >actual &&
+	test_line_count = 5 actual &&
+	git -C r3 rev-list --objects --filter="combine:tree%3A2+blob%3anone" \
+		HEAD >actual &&
+	test_line_count = 5 actual &&
+	git -C r3 rev-list --objects --filter="combine:tree:%30" HEAD >actual &&
+	test_line_count = 2 actual &&
+	git -C r3 rev-list --objects --filter="combine:tree:%39+blob:none" \
+		HEAD >actual &&
+	test_line_count = 5 actual
+'
+
+test_expect_success 'add a sparse pattern blob whose path has reserved chars' '
+	cp r3/pattern r3/pattern1+renamed% &&
+	git -C r3 add pattern1+renamed% &&
+	git -C r3 commit -m "add sparse pattern file with reserved chars"
+'
+
+test_expect_success 'combine:... with more than two sub-filters' '
+	git -C r3 rev-list --objects \
+		--filter=combine:tree:3+blob:limit=40+sparse:oid=master:pattern \
+		HEAD >actual &&
+
+	expect_has HEAD "" &&
+	expect_has HEAD~1 "" &&
+	expect_has HEAD~2 "" &&
+	expect_has HEAD dir1 &&
+	expect_has HEAD dir1/sparse1 &&
+	expect_has HEAD dir1/sparse2 &&
+
+	# Should also have 3 commits
+	test_line_count = 9 actual &&
+
+	# Try again, this time making sure the last sub-filter is only
+	# URL-decoded once.
+	cp actual expect &&
+
+	git -C r3 rev-list --objects \
+		--filter=combine:tree:3+blob:limit=40+sparse:oid=master:pattern1%2brenamed%25 \
+		HEAD >actual &&
+	test_cmp expect actual
+'
+
 # Test provisional omit collection logic with a repo that has objects appearing
 # at multiple depths - first deeper than the filter's threshold, then shallow.
 
@@ -373,6 +491,37 @@ test_expect_success 'verify skipping tree iteration when collecting omits' '
 	test_cmp expect actual
 '
 
+test_expect_success 'setup r5' '
+	git init r5 &&
+	mkdir -p r5/subdir &&
+
+	echo 1     >r5/short-root          &&
+	echo 12345 >r5/long-root           &&
+	echo a     >r5/subdir/short-subdir &&
+	echo abcde >r5/subdir/long-subdir  &&
+
+	git -C r5 add short-root long-root subdir &&
+	git -C r5 commit -m "commit msg"
+'
+
+test_expect_success 'verify collecting omits in combined: filter' '
+	# Note that this test guards against the naive implementation of simply
+	# giving both filters the same "omits" set and expecting it to
+	# automatically merge them.
+	git -C r5 rev-list --objects --quiet --filter-print-omitted \
+		--filter=combine:tree:2+blob:limit=3 HEAD >actual &&
+
+	# Expect 0 trees/commits, 3 blobs omitted (all blobs except short-root)
+	omitted_1=$(echo 12345 | git hash-object --stdin) &&
+	omitted_2=$(echo a     | git hash-object --stdin) &&
+	omitted_3=$(echo abcde | git hash-object --stdin) &&
+
+	grep ~$omitted_1 actual &&
+	grep ~$omitted_2 actual &&
+	grep ~$omitted_3 actual &&
+	test_line_count = 3 actual
+'
+
 # Test tree:<depth> where a tree is iterated to twice - once where a subentry is
 # too deep to be included, and again where the blob inside it is shallow enough
 # to be included. This makes sure we don't use LOFR_MARK_SEEN incorrectly (we
diff --git a/url.c b/url.c
index 1b8ef78cea..e34e5e7517 100644
--- a/url.c
+++ b/url.c
@@ -86,6 +86,12 @@ char *url_decode_mem(const char *url, int len)
 	return url_decode_internal(&url, len, NULL, &out, 0);
 }
 
+char *url_percent_decode(const char *encoded)
+{
+	struct strbuf out = STRBUF_INIT;
+	return url_decode_internal(&encoded, strlen(encoded), NULL, &out, 0);
+}
+
 char *url_decode_parameter_name(const char **query)
 {
 	struct strbuf out = STRBUF_INIT;
diff --git a/url.h b/url.h
index 00b7d58c33..2a27c34277 100644
--- a/url.h
+++ b/url.h
@@ -7,6 +7,14 @@ int is_url(const char *url);
 int is_urlschemechar(int first_flag, int ch);
 char *url_decode(const char *url);
 char *url_decode_mem(const char *url, int len);
+
+/*
+ * Similar to the url_decode_{,mem} methods above, but doesn't assume there
+ * is a scheme followed by a : at the start of the string. Instead, %-sequences
+ * before any : are also parsed.
+ */
+char *url_percent_decode(const char *encoded);
+
 char *url_decode_parameter_name(const char **query);
 char *url_decode_parameter_value(const char **query);