netdev - XFRM Compilation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <DM6PR02MB5913FBB71ABD0C78F238FAA8EE489@DM6PR02MB5913.namprd02.prod.outlook.com>
Date:   Thu, 22 Apr 2021 14:33:06 +0000
From:   Chad Woitas <CWoitas@...-engineering.com>
To:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "shemminger@...l.org" <shemminger@...l.org>
Subject: XFRM Compilation

Hello,

 I'm an automation developer using a pre-emptive kernel for robotics type applications. I believe I've found an issue with XFRM but as I have never dug into the kernel passed "hope it compiles". I wanted to make a comment here, in case it helps.

This was the same result on 5.4.109, 5.4.109-rt56-rc1, 5.10.25, 5.10.25-rt35, 5.10.27, 5.10.27-rt36, 5.10.30-rt37.

After running make -j12:

In file included from ./include/linux/mmzone.h:16,
                 from ./include/linux/gfp.h:6,
                 from ./include/linux/mm.h:10,
                 from ./include/linux/bvec.h:14,
                 from ./include/linux/skbuff.h:17,
                 from ./include/net/xfrm.h:9,
                 from net/xfrm/xfrm_state.c:18:
net/xfrm/xfrm_state.c: In function ‘xfrm_state_init’:
./include/linux/seqlock.h:178:36: error: initialization of ‘seqcount_spinlock_t *’ {aka ‘struct seqcount_spinlock *’} from incompatible pointer type ‘seqcount_t *’ {aka ‘struct seqcount *’} [-Werror=incompatible-pointer-types]
  178 |   seqcount_##lockname##_t *____s = (s);   \
      |                                    ^
./include/linux/seqlock.h:184:42: note: in expansion of macro ‘seqcount_LOCKNAME_init’
  184 | #define seqcount_spinlock_init(s, lock)  seqcount_LOCKNAME_init(s, lock, spinlock)
      |                                          ^~~~~~~~~~~~~~~~~~~~~~
net/xfrm/xfrm_state.c:2666:2: note: in expansion of macro ‘seqcount_spinlock_init’
 2666 |  seqcount_spinlock_init(&net->xfrm.xfrm_state_hash_generation,
      |  ^~~~~~~~~~~~~~~~~~~~~~
  CC [M]  drivers/net/ethernet/realtek/r8169_main.o
  CC [M]  drivers/media/usb/gspca/dtcs033.o
  CC [M]  drivers/gpu/drm/drm_blend.o
  CC [M]  drivers/net/ipvlan/ipvlan_main.o
  CC [M]  drivers/net/ethernet/qualcomm/emac/emac-sgmii-fsm9900.o
  CC [M]  net/ipv6/netfilter/nf_flow_table_ipv6.o
  CC [M]  drivers/media/usb/gspca/etoms.o
  CC [M]  drivers/net/ethernet/qlogic/qed/qed_mcp.o
  CC [M]  drivers/net/ethernet/qualcomm/emac/emac-sgmii-qdf2432.o
  CC [M]  net/ipv6/netfilter/ip6t_ah.o
  CC [M]  drivers/gpu/drm/drm_encoder.o
cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:279: net/xfrm/xfrm_state.o] Error 1
make[1]: *** [scripts/Makefile.build:496: net/xfrm] Error 2
make[1]: *** Waiting for unfinished jobs..

For now I've disabled the strict warning in the make file, line (947), and it works for me. 

# enforce correct pointer usage
KBUILD_CFLAGS   += $(call cc-option) # -Werror=incompatible-pointer-types)

Thanks,
Chad