qat: Export missing symbols in qat_common
ClosedPublic
Actions

Authored by rupeshpilania_gmail.com on Sat, Dec 6, 7:23 PM.

Details

Reviewers

emaste
mav
zlei
rbranco_suse.com
kib
jhb
markj

Commits

rG6ea05372bea5: qat: Export missing symbols in qat_common

Summary

PR 291271: Some symbols required by dependent modules were missing from qat_common.ko and qat_api.ko, causing load failures when local symbol resolution is disabled.

Test Plan

Verified relevant sysctl settings:

sysctl debug.link_elf_leak_locals

debug.link_elf_leak_locals: 0

sysctl debug.link_elf_obj_leak_locals

debug.link_elf_obj_leak_locals: 0

Loaded QAT modules and confirmed all dependencies load correctly:

kldload /boot/kernel/qat.ko
kldstat

Id Refs Address Size Name
1 34 0xffffffff80200000 21268a0 kernel
2 1 0xffffffff82d10000 3220 intpm.ko
3 1 0xffffffff82d14000 2178 smbus.ko
4 1 0xffffffff82d17000 2a80 mac_ntpd.ko
5 1 0xffffffff82d1a000 4390 qat.ko
6 6 0xffffffff82d1f000 15dd0 qat_hw.ko
7 9 0xffffffff82d35000 30020 qat_common.ko
8 8 0xffffffff82d66000 68cd8 qat_api.ko

All symbols exported correctly; no module loading errors observed.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Passed

Unit

No Test Coverage

Build Status

Buildable 69105
Build 65988: arc lint + arc unit

Event Timeline

rupeshpilania_gmail.com created this revision.Sat, Dec 6, 7:23 PM

Herald added a subscriber: imp. · View Herald TranscriptSat, Dec 6, 7:23 PM

rupeshpilania_gmail.com requested review of this revision.Sat, Dec 6, 7:23 PM

Harbormaster completed remote builds in B69105: Diff 167652.Sat, Dec 6, 7:23 PM

pr-291271-test-logs.txt1 KBDownload

rupeshpilania_gmail.com attached a referenced file: F139031812: pr-291271-test-logs.txt. (Show Details)Sat, Dec 6, 7:33 PM

ziaee added a reviewer: markj.Sat, Dec 6, 7:55 PM

ziaee added a subscriber: ziaee.

Thank you for working on this, but for this driver I think it would be better to just have EXPORT_SYMS=yes rather than enumerating and maintaining the list of symbols. Otherwise this will be a headache for the upstream maintainer.

In D54107#1236540, @markj wrote:

Thank you for working on this, but for this driver I think it would be better to just have EXPORT_SYMS=yes rather than enumerating and maintaining the list of symbols. Otherwise this will be a headache for the upstream maintainer.

Agreed. Too many symbols to export and I think it is painful to maintain.

In D54107#1237875, @zlei wrote:

In D54107#1236540, @markj wrote:

Thank you for working on this, but for this driver I think it would be better to just have EXPORT_SYMS=yes rather than enumerating and maintaining the list of symbols. Otherwise this will be a headache for the upstream maintainer.

Agreed. Too many symbols to export and I think it is painful to maintain.