> -----Original Message-----
> From: Fijalkowski, Maciej <maciej.fijalkowski@xxxxxxxxx>
> Sent: Wednesday, March 12, 2025 3:41 AM
> To: Vyavahare, Tushar <tushar.vyavahare@xxxxxxxxx>
> Cc: bpf@xxxxxxxxxxxxxxx; netdev@xxxxxxxxxxxxxxx; bjorn@xxxxxxxxxx; Karlsson,
> Magnus <magnus.karlsson@xxxxxxxxx>; jonathan.lemon@xxxxxxxxx;
> davem@xxxxxxxxxxxxx; kuba@xxxxxxxxxx; pabeni@xxxxxxxxxx;
> ast@xxxxxxxxxx; daniel@xxxxxxxxxxxxx; Sarkar, Tirthendu
> <tirthendu.sarkar@xxxxxxxxx>
> Subject: Re: [PATCH bpf-next v3 2/2] selftests/xsk: Add tail adjustment tests
> and support check
>
> On Wed, Mar 05, 2025 at 02:18:13PM +0000, Tushar Vyavahare wrote:
> > Introduce tail adjustment functionality in xskxceiver using
> > bpf_xdp_adjust_tail(). Add `xsk_xdp_adjust_tail` to modify packet
> > sizes and drop unmodified packets. Implement
> > `is_adjust_tail_supported` to check helper availability. Develop
> > packet resizing tests, including shrinking and growing scenarios, with
> > functions for both single-buffer and multi-buffer cases. Update the
> > test framework to handle various scenarios and adjust MTU settings.
> > These changes enhance the testing of packet tail adjustments, improving
> AF_XDP framework reliability.
> >
> > Signed-off-by: Tushar Vyavahare <tushar.vyavahare@xxxxxxxxx>
> > ---
> > .../selftests/bpf/progs/xsk_xdp_progs.c | 49 ++++++++
> > tools/testing/selftests/bpf/xsk_xdp_common.h | 1 +
> > tools/testing/selftests/bpf/xskxceiver.c | 107 +++++++++++++++++-
> > tools/testing/selftests/bpf/xskxceiver.h | 2 +
> > 4 files changed, 157 insertions(+), 2 deletions(-)
> >
> > diff --git a/tools/testing/selftests/bpf/progs/xsk_xdp_progs.c
> > b/tools/testing/selftests/bpf/progs/xsk_xdp_progs.c
> > index ccde6a4c6319..34c832a5a98c 100644
> > --- a/tools/testing/selftests/bpf/progs/xsk_xdp_progs.c
> > +++ b/tools/testing/selftests/bpf/progs/xsk_xdp_progs.c
> > @@ -4,6 +4,8 @@
> > #include <linux/bpf.h>
> > #include <bpf/bpf_helpers.h>
> > #include <linux/if_ether.h>
> > +#include <linux/ip.h>
> > +#include <linux/errno.h>
> > #include "xsk_xdp_common.h"
> >
> > struct {
> > @@ -14,6 +16,7 @@ struct {
> > } xsk SEC(".maps");
> >
> > static unsigned int idx;
> > +int adjust_value = 0;
> > int count = 0;
> >
> > SEC("xdp.frags") int xsk_def_prog(struct xdp_md *xdp) @@ -70,4 +73,50
> > @@ SEC("xdp") int xsk_xdp_shared_umem(struct xdp_md *xdp)
> > return bpf_redirect_map(&xsk, idx, XDP_DROP); }
> >
> > +SEC("xdp.frags") int xsk_xdp_adjust_tail(struct xdp_md *xdp) {
> > + __u32 buff_len, curr_buff_len;
> > + int ret;
> > +
> > + buff_len = bpf_xdp_get_buff_len(xdp);
> > + if (buff_len == 0)
> > + return XDP_DROP;
> > +
> > + ret = bpf_xdp_adjust_tail(xdp, adjust_value);
> > + if (ret < 0) {
> > + /* Handle unsupported cases */
> > + if (ret == -EOPNOTSUPP) {
> > + /* Set adjust_value to -EOPNOTSUPP to indicate to
> userspace that this case
> > + * is unsupported
> > + */
> > + adjust_value = -EOPNOTSUPP;
> > + return bpf_redirect_map(&xsk, 0, XDP_DROP);
>
> so in this case you will proceed with running whole traffic? IMHO test should
> be stopped for very first notsupp case, but not sure if it's worth the hassle.
>
When the expected packet length does not match, the test will fail on the
very first packet in receive_pkts(). After this initial failure, check if
the adjust_tail function is supported, but only for cases where the test
involves adjust_tail and fails.
> > + }
> > +
> > + return XDP_DROP;
> > + }
> > +
> > + curr_buff_len = bpf_xdp_get_buff_len(xdp);
> > + if (curr_buff_len != buff_len + adjust_value)
> > + return XDP_DROP;
> > +
> > + if (curr_buff_len > buff_len) {
> > + __u32 *pkt_data = (void *)(long)xdp->data;
> > + __u32 len, words_to_end, seq_num;
> > +
> > + len = curr_buff_len - PKT_HDR_ALIGN;
> > + words_to_end = len / sizeof(*pkt_data) - 1;
> > + seq_num = words_to_end;
> > +
> > + /* Convert sequence number to network byte order. Store this
> in the last 4 bytes of
> > + * the packet. Use 'adjust_value' to determine the position at
> the end of the
> > + * packet for storing the sequence number.
> > + */
> > + seq_num = __constant_htonl(words_to_end);
> > + bpf_xdp_store_bytes(xdp, curr_buff_len - adjust_value,
> &seq_num,
> > +sizeof(seq_num));
>
> what is the purpose of that? test suite expects seq_num to be at the end of
> the packet so you have to double it here?
>
The test suite expects the seq_num to be at the end of the packet, so it
needs to be doubled to ensure it is placed correctly in the final packet
structure.
> > + }
> > +
> > + return bpf_redirect_map(&xsk, 0, XDP_DROP); }
> > +
> > char _license[] SEC("license") = "GPL"; diff --git
> > a/tools/testing/selftests/bpf/xsk_xdp_common.h
> > b/tools/testing/selftests/bpf/xsk_xdp_common.h
> > index 5a6f36f07383..45810ff552da 100644
> > --- a/tools/testing/selftests/bpf/xsk_xdp_common.h
> > +++ b/tools/testing/selftests/bpf/xsk_xdp_common.h
> > @@ -4,6 +4,7 @@
> > #define XSK_XDP_COMMON_H_
> >
> > #define MAX_SOCKETS 2
> > +#define PKT_HDR_ALIGN (sizeof(struct ethhdr) + 2) /* Just to align
> > +the data in the packet */
> >
> > struct xdp_info {
> > __u64 count;
> > diff --git a/tools/testing/selftests/bpf/xskxceiver.c
> > b/tools/testing/selftests/bpf/xskxceiver.c
> > index d60ee6a31c09..bcc265fc784c 100644
> > --- a/tools/testing/selftests/bpf/xskxceiver.c
> > +++ b/tools/testing/selftests/bpf/xskxceiver.c
> > @@ -524,6 +524,8 @@ static void __test_spec_init(struct test_spec *test,
> struct ifobject *ifobj_tx,
> > test->nb_sockets = 1;
> > test->fail = false;
> > test->set_ring = false;
> > + test->adjust_tail = false;
> > + test->adjust_tail_support = false;
> > test->mtu = MAX_ETH_PKT_SIZE;
> > test->xdp_prog_rx = ifobj_rx->xdp_progs->progs.xsk_def_prog;
> > test->xskmap_rx = ifobj_rx->xdp_progs->maps.xsk; @@ -992,6 +994,31
> > @@ static bool is_metadata_correct(struct pkt *pkt, void *buffer, u64 addr)
> > return true;
> > }
> >
> > +static bool is_adjust_tail_supported(struct xsk_xdp_progs *skel_rx) {
> > + struct bpf_map *data_map;
> > + int adjust_value = 0;
> > + int key = 0;
> > + int ret;
> > +
> > + data_map = bpf_object__find_map_by_name(skel_rx->obj,
> "xsk_xdp_.bss");
> > + if (!data_map || !bpf_map__is_internal(data_map)) {
> > + ksft_print_msg("Error: could not find bss section of XDP
> program\n");
> > + exit_with_error(errno);
> > + }
> > +
> > + ret = bpf_map_lookup_elem(bpf_map__fd(data_map), &key,
> &adjust_value);
> > + if (ret) {
> > + ksft_print_msg("Error: bpf_map_lookup_elem failed with error
> %d\n", ret);
> > + return false;
>
> exit_with_error(errno) to be consistent?
>
Will do.
> > + }
> > +
> > + /* Set the 'adjust_value' variable to -EOPNOTSUPP in the XDP program
> if the adjust_tail
> > + * helper is not supported. Skip the adjust_tail test case in this
> scenario.
> > + */
> > + return adjust_value != -EOPNOTSUPP;
> > +}
> > +
> > static bool is_frag_valid(struct xsk_umem_info *umem, u64 addr, u32 len,
> u32 expected_pkt_nb,
> > u32 bytes_processed)
> > {
> > @@ -1768,8 +1795,13 @@ static void *worker_testapp_validate_rx(void
> > *arg)
> >
> > if (!err && ifobject->validation_func)
> > err = ifobject->validation_func(ifobject);
> > - if (err)
> > - report_failure(test);
> > +
> > + if (err) {
> > + if (test->adjust_tail && !is_adjust_tail_supported(ifobject-
> >xdp_progs))
> > + test->adjust_tail_support = false;
>
> how about setting is_adjust_tail_supported() as validation_func? Would that
> work? Special casing this check specially for tail manipulation tests seems a bit
> odd.
>
Setting is_adjust_tail_supported() as the validation_func would not work
directly. This function is designed to check if the bpf_xdp_adjust_tail
helper is supported by the XDP program, rather than to validate test
results. It assesses the capability of the XDP program, not the outcome of
the tests.
> > + else
> > + report_failure(test);
> > + }
> >
> > pthread_exit(NULL);
> > }
> > @@ -2516,6 +2548,73 @@ static int testapp_hw_sw_max_ring_size(struct
> test_spec *test)
> > return testapp_validate_traffic(test); }
> >
> > +static int testapp_xdp_adjust_tail(struct test_spec *test, int
> > +adjust_value) {
> > + struct xsk_xdp_progs *skel_rx = test->ifobj_rx->xdp_progs;
> > + struct xsk_xdp_progs *skel_tx = test->ifobj_tx->xdp_progs;
> > +
> > + test_spec_set_xdp_prog(test, skel_rx->progs.xsk_xdp_adjust_tail,
> > + skel_tx->progs.xsk_xdp_adjust_tail,
> > + skel_rx->maps.xsk, skel_tx->maps.xsk);
> > +
> > + skel_rx->bss->adjust_value = adjust_value;
> > +
> > + return testapp_validate_traffic(test); }
> > +
> > +static int testapp_adjust_tail(struct test_spec *test, u32 value, u32
> > +pkt_len) {
> > + u32 pkt_cnt = DEFAULT_BATCH_SIZE;
>
> you don't need this variable
>
will do.
> > + int ret;
> > +
> > + test->adjust_tail_support = true;
> > + test->adjust_tail = true;
> > + test->total_steps = 1;
> > +
> > + pkt_stream_replace_ifobject(test->ifobj_tx, pkt_cnt, pkt_len);
> > + pkt_stream_replace_ifobject(test->ifobj_rx, pkt_cnt, pkt_len +
> > +value);
> > +
> > + ret = testapp_xdp_adjust_tail(test, value);
> > + if (ret)
> > + return ret;
> > +
> > + if (!test->adjust_tail_support) {
> > + ksft_test_result_skip("%s %sResize pkt with
> bpf_xdp_adjust_tail() not supported\n",
> > + mode_string(test), busy_poll_string(test));
> > + return TEST_SKIP;
>
> missing indent
>
will do.
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +static int testapp_adjust_tail_common(struct test_spec *test, int
> adjust_value, u32 len,
> > + bool set_mtu)
> > +{
> > + if (set_mtu)
> > + test->mtu = MAX_ETH_JUMBO_SIZE;
>
> could you remove this and instead of bool_set_mtu just pass the value of
> mtu?
>
will do.
> static int testapp_adjust_tail(struct test_spec *test, u32 value, u32 pkt_len,
> u32 mtu) {
> (...)
> if (test->mtu != mtu)
> test->mtu = mtu;
> (...)
> }
>
> static int testapp_adjust_tail_shrink(struct test_spec *test) {
> return testapp_adjust_tail(test, -4, MIN_PKT_SIZE,
> MAX_ETH_PKT_SIZE); }
>
> static int testapp_adjust_tail_shrink_mb(struct test_spec *test) {
> return testapp_adjust_tail(test, -4,
> XSK_RING_PROD__DEFAULT_NUM_DESCS * 3,
> MAX_ETH_JUMBO_SIZE);
> }
>
> no need for _common func.
>
will do.
> > + return testapp_adjust_tail(test, adjust_value, len); }
> > +
> > +static int testapp_adjust_tail_shrink(struct test_spec *test) {
> > + return testapp_adjust_tail_common(test, -4, MIN_PKT_SIZE, false); }
> > +
> > +static int testapp_adjust_tail_shrink_mb(struct test_spec *test) {
> > + return testapp_adjust_tail_common(test, -4,
> > +XSK_RING_PROD__DEFAULT_NUM_DESCS * 3, true);
>
> Am I reading this right that you are modifying the size by just 4 bytes?
> The bugs that drivers had were for cases when packets got modified by value
> bigger than frag size which caused for example underlying page being freed.
>
> If that is the case tests do nothing valuable from my perspective.
>
In the v4 patchset, I have updated the code to modify the packet size by
1024 bytes instead of just 4 bytes.
I will send v4.
> > +}
> > +
> > +static int testapp_adjust_tail_grow(struct test_spec *test) {
> > + return testapp_adjust_tail_common(test, 4, MIN_PKT_SIZE, false); }
> > +
> > +static int testapp_adjust_tail_grow_mb(struct test_spec *test) {
> > + return testapp_adjust_tail_common(test, 4,
> > +XSK_RING_PROD__DEFAULT_NUM_DESCS * 3, true); }
> > +
> > static void run_pkt_test(struct test_spec *test) {
> > int ret;
> > @@ -2622,6 +2721,10 @@ static const struct test_spec tests[] = {
> > {.name = "TOO_MANY_FRAGS", .test_func = testapp_too_many_frags},
> > {.name = "HW_SW_MIN_RING_SIZE", .test_func =
> testapp_hw_sw_min_ring_size},
> > {.name = "HW_SW_MAX_RING_SIZE", .test_func =
> > testapp_hw_sw_max_ring_size},
> > + {.name = "XDP_ADJUST_TAIL_SHRINK", .test_func =
> testapp_adjust_tail_shrink},
> > + {.name = "XDP_ADJUST_TAIL_SHRINK_MULTI_BUFF", .test_func =
> testapp_adjust_tail_shrink_mb},
> > + {.name = "XDP_ADJUST_TAIL_GROW", .test_func =
> testapp_adjust_tail_grow},
> > + {.name = "XDP_ADJUST_TAIL_GROW_MULTI_BUFF", .test_func =
> > +testapp_adjust_tail_grow_mb},
> > };
> >
> > static void print_tests(void)
> > diff --git a/tools/testing/selftests/bpf/xskxceiver.h
> > b/tools/testing/selftests/bpf/xskxceiver.h
> > index e46e823f6a1a..67fc44b2813b 100644
> > --- a/tools/testing/selftests/bpf/xskxceiver.h
> > +++ b/tools/testing/selftests/bpf/xskxceiver.h
> > @@ -173,6 +173,8 @@ struct test_spec {
> > u16 nb_sockets;
> > bool fail;
> > bool set_ring;
> > + bool adjust_tail;
> > + bool adjust_tail_support;
> > enum test_mode mode;
> > char name[MAX_TEST_NAME_SIZE];
> > };
> > --
> > 2.34.1
> >
