On Tue, Aug 1, 2023 at 12:19 PM Bruce Richardson
<bruce.richard...@intel.com> wrote:
>
> On Mon, Jul 31, 2023 at 08:39:31PM -0400, Patrick Robb wrote:
> > Hi Bruce,
> > I see some failures for this series for our Ubuntu 20.04 containers.
> > And, our DTS testbeds which are on ubuntu 20.04 are skipping running
> > testsuites because they can't compile DPDK. So, that's why it has some
> > missing results for a couple of the Intel NICs. For context, I'll paste
> > below where the compile job terminates in one of our containerized
> > compile test runs. The GCC in use here is version 9.4, so it meets the
> > requirements as described in your patch as far as I can tell. I'll
> > check it out more tomorrow to see whether it's an infra failure, like
> > some missing dependencies. Please let me know if we expect to have no
> > issues with 20.04 or if this is anticipated.
> > Thanks!
> > [1638/2730] Generating symbol file
> > 'drivers/a715181@@rte_net_ixgbe@sha/librte_net_ixgbe.so.24.0.symbols'.
> > [1639/2730] Compiling C object
> > 'drivers/a715181@@tmp_rte_net_mlx4@sta/net_mlx4_mlx4.c.o'.
> > FAILED: drivers/a715181@@tmp_rte_net_mlx4@sta/net_mlx4_mlx4.c.o
> > ccache cc -Idrivers/a715181@@tmp_rte_net_mlx4@sta -Idrivers
> > -I../drivers -Idrivers/net/mlx4 -I../drivers/net/mlx4 -Ilib/ethdev
> > -I../lib/ethdev -I. -I../ -Iconfig -I../config -Ilib/eal/include
> > -I../lib/eal/include -Ilib/eal/linux/include -I../lib/eal/linux/include
> > -Ilib/eal/x86/include -I../lib/eal/x86/include -Ilib/eal/common
> > -I../lib/eal/common -Ilib/eal -I../lib/eal -Ilib/kvargs -I../lib/kvargs
> > -Ilib/telemetry/../metrics -I../lib/telemetry/../metrics
> > -Ilib/telemetry -I../lib/telemetry -Ilib/net -I../lib/net -Ilib/mbuf
> > -I../lib/mbuf -Ilib/mempool -I../lib/mempool -Ilib/ring -I../lib/ring
> > -Ilib/meter -I../lib/meter -Idrivers/bus/pci -I../drivers/bus/pci
> > -I../drivers/bus/pci/linux -Ilib/pci -I../lib/pci -Idrivers/bus/vdev
> > -I../drivers/bus/vdev -I/usr/include/libnl3 -fdiagnostics-color=always
> > -pipe -D_FILE_OFFSET_BITS=64 -Wall -Winvalid-pch -Wextra -Werror
> > -std=c11 -O3 -include rte_config.h -Wcast-qual -Wdeprecated -Wformat
> > -Wformat-nonliteral -Wformat-security -Wmissing-declarations
> > -Wmissing-prototypes -Wnested-externs -Wold-style-definition
> > -Wpointer-arith -Wsign-compare -Wstrict-prototypes -Wundef
> > -Wwrite-strings -Wno-address-of-packed-member -Wno-packed-not-aligned
> > -Wno-missing-field-initializers -D_GNU_SOURCE -fPIC -march=native
> > -DALLOW_EXPERIMENTAL_API -DALLOW_INTERNAL_API -Wno-format-truncation
> > -std=c11 -Wno-strict-prototypes -D_BSD_SOURCE -D_DEFAULT_SOURCE
> > -D_XOPEN_SOURCE=600 -UPEDANTIC -DRTE_LOG_DEFAULT_LOGTYPE=pmd.net.mlx4
> > -MD -MQ 'drivers/a715181@@tmp_rte_net_mlx4@sta/net_mlx4_mlx4.c.o' -MF
> > 'drivers/a715181@@tmp_rte_net_mlx4@sta/net_mlx4_mlx4.c.o.d' -o
> > 'drivers/a715181@@tmp_rte_net_mlx4@sta/net_mlx4_mlx4.c.o' -c
> > ../drivers/net/mlx4/mlx4.c
> > In file included from ../drivers/net/mlx4/mlx4_rxtx.h:27,
> > from ../drivers/net/mlx4/mlx4.c:49:
> > ../drivers/net/mlx4/mlx4_prm.h:111:8: error: redefinition of 'struct
> > mlx4_wqe_lso_seg'
> > 111 | struct mlx4_wqe_lso_seg {
> > | ^~~~~~~~~~~~~~~~
> > In file included from ../drivers/net/mlx4/mlx4_glue.h:16,
> > from ../drivers/net/mlx4/mlx4.c:46:
> > /usr/include/infiniband/mlx4dv.h:410:8: note: originally defined here
> > 410 | struct mlx4_wqe_lso_seg {
> > | ^~~~~~~~~~~~~~~~
> > ninja: build stopped: subcommand failed.
>
> Hi again,
>
> I've attempted to reproduce this on my Ubuntu 20.04 VM and failed,
> everything seems to build ok.
>
> Looking through the logs, though, there does appear to be a difference in
> the configurations in the two cases. I suspect my Ubuntu has an updated
> verbs package compared to the image you are using. In the log of the failed
> build, I see:
>
> Checking whether type "struct mlx4_wqe_lso_seg" has member
> "mss_hdr_size" with dependencies libmlx4, libibverbs: NO
> Configuring mlx4_autoconf.h using configuration
>
> While building in my VM, I have:
>
> Checking whether type "struct mlx4_wqe_lso_seg" has member
> "mss_hdr_size" with dependencies libmlx4, libibverbs: YES (cached)
> Configuring mlx4_autoconf.h using configuration
>
> So my verbs mlx4 header has got a different set of definitions to those in
> the CI machine. My Ubuntu reports as 20.04.6 with libibverbs-dev package
> version "28.0-1ubuntu1"
>
> Can the CI image be updated to latest 20.04 packages?
>
> /Bruce
>
I can reproduce the issue seen at UNH, with a 20.04.6 container and
the same libibverbs as you:
ii libibverbs-dev:amd64 28.0-1ubuntu1
amd64 Development files for the libibverbs library
So I suspect something is different in container images..
Pasting the (hopefully) relevant meson logs:
Running compile:
Working directory: /root/dpdk/build/meson-private/tmp0ovvvd9g
Command line: ccache cc -I/usr/include/libnl3
/root/dpdk/build/meson-private/tmp0ovvvd9g/testfile.c -o
/root/dpdk/build/meson-private/tmp0ovvvd9g/output.obj -pipe -c
-D_FILE_OFFSET_BITS=64 -O0 -std=c11
Code:
#include <infiniband/mlx4dv.h>
void bar(void) {
struct mlx4_wqe_lso_seg foo;
foo.mss_hdr_size;
};
Compiler stdout:
Compiler stderr:
In file included from /root/dpdk/build/meson-private/tmp0ovvvd9g/testfile.c:1:
/usr/include/infiniband/mlx4dv.h:176:2: error: unknown type name 'off_t'
176 | off_t uar_mmap_offset;
| ^~~~~
Checking whether type "struct mlx4_wqe_lso_seg" has member
"mss_hdr_size" with dependencies libmlx4, libibverbs: NO
--
David Marchand
