Hi,
On 26.10.23 09:11, kernel test robot wrote:
Hello,
kernel test robot noticed "kernel-selftests.uevent.uevent_filtering.fail" on:
commit: 5b45a753776be5d21cf395ec97e81c9187fbeaca ("selftests: uevent filtering: fix return on error in uevent_listener") https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
[test failed on linux-next/master 2030579113a1b1b5bfd7ff24c0852847836d8fd1]
in testcase: kernel-selftests version: kernel-selftests-x86_64-60acb023-1_20230329 with following parameters:
group: group-03
compiler: gcc-12 test machine: 36 threads 1 sockets Intel(R) Core(TM) i9-10980XE CPU @ 3.00GHz (Cascade Lake) with 32G memory
(please refer to attached dmesg/kmsg for entire log/backtrace)
we also noticed this issue does not always happen. as below, we saw 15 failures out of 50 runs. however, parent keeps passing.
37013b557b7f39e6 5b45a753776be5d21cf395ec97e
fail:runs %reproduction fail:runs | | | :50 30% 15:50 kernel-selftests.uevent.uevent_filtering.failIf you fix the issue in a separate patch/commit (i.e. not just a new version of the same patch/commit), kindly add following tags | Reported-by: kernel test robot oliver.sang@intel.com | Closes: https://lore.kernel.org/oe-lkp/202310261454.46082aaa-oliver.sang@intel.com
TAP version 13 1..1 # timeout set to 300 # selftests: uevent: uevent_filtering # TAP version 13 # 1..1 # # Starting 1 tests from 1 test cases. # # RUN global.uevent_filtering ... # add@/devices/virtual/mem/fullACTION=addDEVPATH=/devices/virtual/mem/fullSUBSYSTEM=memSYNTH_UUID=0MAJOR=1MINOR=7DEVNAME=fullDEVMODE=0666SEQNUM=3532 # add@/devices/virtual/mem/fullACTION=addDEVPATH=/devices/virtual/mem/fullSUBSYSTEM=memSYNTH_UUID=0MAJOR=1MINOR=7DEVNAME=fullDEVMODE=0666SEQNUM=3546 # add@/devices/virtual/mem/fullACTION=addDEVPATH=/devices/virtual/mem/fullSUBSYSTEM=memSYNTH_UUID=0MAJOR=1MINOR=7DEVNAME=fullDEVMODE=0666SEQNUM=3556 # add@/devices/virtual/mem/fullACTION=addDEVPATH=/devices/virtual/mem/fullSUBSYSTEM=memSYNTH_UUID=0MAJOR=1MINOR=7DEVNAME=fullDEVMODE=0666SEQNUM=3585 # add@/devices/virtual/mem/fullACTION=addDEVPATH=/devices/virtual/mem/fullSUBSYSTEM=memSYNTH_UUID=0MAJOR=1MINOR=7DEVNAME=fullDEVMODE=0666SEQNUM=3595 # No buffer space available - Failed to receive uevent # # uevent_filtering.c:479:uevent_filtering:Expected 0 (0) == ret (-1) # # uevent_filtering: Test failed at step #10 # # FAIL global.uevent_filtering # not ok 1 global.uevent_filtering # # FAILED: 0 / 1 tests passed. # # Totals: pass:0 fail:1 xfail:0 xpass:0 skip:0 error:0 not ok 1 selftests: uevent: uevent_filtering # exit=1
The kernel config and materials to reproduce are available at: https://download.01.org/0day-ci/archive/20231026/202310261454.46082aaa-olive...
What is apparently happening here is that the test returns now a proper error value in the uevent_listener function. In this case recvmsg() is returning ENOBUFS - No buffer space available. This occurs because buf[_UEVENT_BUFFER_SIZE] is not big enough, being _UEVENT_BUFFER_SIZE = (2048 *2).
Therefore the error is not in the fix commit as such because the error was happening and it was not returned properly (previously ret was assigned to -1 but it was not returned).
The obvious fix would be a larger buffer, but when I see de definition of UEVENT_BUFFER_SIZE in kobject.h, its value is just 2048. Then I wonder why 2048 *2 was used in the first place, why that is not enough and what size is then meaningful.
After a quick search I found a discussion on Bugzilla about the UEVENT_BUFFER_SIZE being resized to 4096:
https://bugzilla.kernel.org/show_bug.cgi?id=207695
But apparently it has not been resized upstream.
Best regards, Javier Carrasco