mdadm: upgrade 4.4 -> 4.5
PATCH v2 1/1 mdadm: upgrade 4.4 -> 4.5
From: Ankur Tyagi <[email protected]>
Changelog:
https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/tree/CHANGELOG.md?h=mdadm-4.5
Replaced 0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch with 0001-raid6check.c-restripe.c-Use-64-bit-off_t-across-both.patch as partial changes from first patch are part of the upstream version[1].
Dropped patches that are part of upstream version
- xmalloc.patch [2]
- 0001-Makefile-install-mdcheck.patch [3]
Also refreshed other patches.
[1] https://github.com/md-raid-utilities/mdadm/commit/787cc1b60130b8031be59e49d54463c58cd8cf74
[2] https://git.kernel.org/pub/scm/utils/mdadm/mdadm.git/commit/raid6check.c?h=mdadm-4.5&id=e0df6c4c984d564e9e40913727e916a6cd8f466e
[3] https://github.com/md-raid-utilities/mdadm/commit/934d81184880d2231b0d7b3836b12358516ea35c
---
changes in v2:
- dropped mdcheck install patch to address reproducibility issue
- removed dropped patches from git
---
...-corosync-and-dlm-header-files-check.patch | 6 +-
.../files/0001-Makefile-install-mdcheck.patch | 27 ---
...tripe.c-Use-64-bit-off_t-across-both.patch | 54 ++++++
...FILE_OFFSET_BITS-to-enable-largefile.patch | 155 ------------------
...-linux-falloc.h-for-FALLOC_FL_ZERO_R.patch | 6 +-
.../mdadm/files/xmalloc.patch | 32 ----
.../mdadm/{mdadm_4.4.bb => mdadm_4.5.bb} | 8 +-
7 files changed, 63 insertions(+), 225 deletions(-)
delete mode 100644 meta/recipes-extended/mdadm/files/0001-Makefile-install-mdcheck.patch
create mode 100644 meta/recipes-extended/mdadm/files/0001-raid6check.c-restripe.c-Use-64-bit-off_t-across-both.patch
delete mode 100644 meta/recipes-extended/mdadm/files/0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch
delete mode 100644 meta/recipes-extended/mdadm/files/xmalloc.patch
rename meta/recipes-extended/mdadm/{mdadm_4.4.bb => mdadm_4.5.bb} (93%)meta/recipes-extended/mdadm/files/0001-Fix-the-path-of-corosync-and-dlm-header-files-check.patch +3 -3
@@ -1,4 +1,4 @@ -From 232101e2fcc6359fb6db2bc232570e373e368c75 Mon Sep 17 00:00:00 2001 +From 77e9527a0ea995ed377e00813eeeb17175dbce72 Mon Sep 17 00:00:00 2001 From: "Maxin B. John" <[email protected]> Date: Tue, 9 Feb 2016 11:44:01 +0200 Subject: [PATCH] Fix the path of corosync and dlm header files check
@@ -14,10 +14,10 @@ Signed-off-by: Maxin B. John <[email protected]> 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/Makefile b/Makefile -index bcd092de..0a78cd22 100644 +index 8031ce24..788ff5d9 100644 --- a/Makefile +++ b/Makefile -@@ -130,8 +130,8 @@ FAILED_SLOTS_DIR = $(RUN_DIR)/failed-slots +@@ -120,8 +120,8 @@ FAILED_SLOTS_DIR = $(RUN_DIR)/failed-slots SYSTEMD_DIR=/lib/systemd/system LIB_DIR=/usr/libexec/mdadm
meta/recipes-extended/mdadm/files/0001-Makefile-install-mdcheck.patch deleted +0 -27
@@ -1,27 +0,0 @@ -From 407691e11037709af888ce2cf6bd5eac6971ac61 Mon Sep 17 00:00:00 2001 -From: Chen Qi <[email protected]> -Date: Tue, 25 Jan 2022 16:25:01 +0800 -Subject: [PATCH] Makefile: install mdcheck - -The mdcheck_xxx.service files use mdcheck, but it's not installed. -We need to install this script. - -Upstream-Status: Submitted [Sent patch to maintainer] - -Signed-off-by: Chen Qi <[email protected]> ---- - Makefile | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/Makefile b/Makefile -index 981e16fa..b28e7855 100644 ---- a/Makefile -+++ b/Makefile -@@ -345,6 +345,7 @@ install-systemd: systemd/[email protected] - install-bin: mdadm mdmon - $(INSTALL) -D $(STRIP) -m 755 mdadm $(DESTDIR)$(BINDIR)/mdadm - $(INSTALL) -D $(STRIP) -m 755 mdmon $(DESTDIR)$(BINDIR)/mdmon -+ $(INSTALL) -D -m 755 misc/mdcheck $(DESTDIR)/usr/share/mdadm/mdcheck - - uninstall: - rm -f $(DESTDIR)$(MAN8DIR)/mdadm.8 $(DESTDIR)$(MAN8DIR)/mdmon.8 $(DESTDIR)$(MAN4DIR)/md.4 $(DESTDIR)$(MAN5DIR)/mdadm.conf.5 $(DESTDIR)$(BINDIR)/mdadm
meta/recipes-extended/mdadm/files/0001-raid6check.c-restripe.c-Use-64-bit-off_t-across-both.patch new file +54 -0
54 lines not on this page · raw message
meta/recipes-extended/mdadm/files/0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch deleted +0 -155
155 lines not on this page · raw message
meta/recipes-extended/mdadm/files/0002-Create.c-include-linux-falloc.h-for-FALLOC_FL_ZERO_R.patch +3 -3
@@ -1,4 +1,4 @@ -From 8de073c637a1ea968b7304f373a20b13a4a47bbc Mon Sep 17 00:00:00 2001 +From b7e51fbedc1fbc44af00398a4a7aef51cdf191ee Mon Sep 17 00:00:00 2001 From: Alexander Kanavin <[email protected]> Date: Tue, 12 Mar 2024 10:54:08 +0100 Subject: [PATCH] Create.c: include linux/falloc.h for FALLOC_FL_ZERO_RANGE
@@ -14,10 +14,10 @@ Signed-off-by: Alexander Kanavin <[email protected]> 1 file changed, 1 insertion(+) diff --git a/Create.c b/Create.c -index fd6c9215..3210a03e 100644 +index 4aa75b7e..5d5fc806 100644 --- a/Create.c +++ b/Create.c -@@ -33,6 +33,7 @@ +@@ -31,6 +31,7 @@ #include <signal.h> #include <sys/signalfd.h> #include <sys/wait.h>
meta/recipes-extended/mdadm/files/xmalloc.patch deleted +0 -32
@@ -1,32 +0,0 @@ -From c5cf5a83be4e3ce04ebf3250f77aeb465eeb53a1 Mon Sep 17 00:00:00 2001 -From: Xiao Ni <[email protected]> -Date: Fri, 17 Jan 2025 15:15:40 +0800 -Subject: [PATCH] mdadm/raid6check: add xmalloc.h to raid6check.c - -It reports building error: -raid6check.c:324:26: error: implicit declaration of function xmalloc - -Add xmalloc.h to raid6check.c file to fix this. - -Signed-off-by: Xiao Ni <[email protected]> -Link: https://lore.kernel.org/r/[email protected] -Signed-off-by: Song Liu <[email protected]> - -Upstream-Status: Backport [https://web.git.kernel.org/pub/scm/utils/mdadm/mdadm.git/commit/?h=main&id=e0df6c4c984d564e9e40913727e916a6cd8f466e] -Signed-off-by: Alexander Kanavin <[email protected]> ---- - raid6check.c | 1 + - 1 file changed, 1 insertion(+) - -diff --git a/raid6check.c b/raid6check.c -index 8e7f1421..486b10c6 100644 ---- a/raid6check.c -+++ b/raid6check.c -@@ -26,6 +26,7 @@ - #define _FILE_OFFSET_BITS 64 - - #include "mdadm.h" -+#include "xmalloc.h" - #include <stdint.h> - #include <sys/mman.h> -
meta/recipes-extended/mdadm/mdadm_4.5.bb renamed from meta/recipes-extended/mdadm/mdadm_4.4.bb +3 -5
@@ -12,13 +12,11 @@ SRC_URI = "git://git.kernel.org/pub/scm/utils/mdadm/mdadm.git;protocol=https;bra file://run-ptest \ file://0001-Fix-the-path-of-corosync-and-dlm-header-files-check.patch \ file://mdadm.init \ - file://0001-Makefile-install-mdcheck.patch \ - file://0001-restripe.c-Use-_FILE_OFFSET_BITS-to-enable-largefile.patch \ file://0002-Create.c-include-linux-falloc.h-for-FALLOC_FL_ZERO_R.patch \ - file://xmalloc.patch \ + file://0001-raid6check.c-restripe.c-Use-64-bit-off_t-across-both.patch \ " -SRCREV = "8e56efac9afd7080bb42bae4b77cdad5f345633a" +SRCREV = "8f0c7692d48414ff7b3fe927ce75799c65ef24b1" inherit ptest systemd
@@ -37,7 +35,7 @@ CFLAGS:append:mipsarchn32 = ' -D__SANE_USERSPACE_TYPES__' EXTRA_OEMAKE = 'CHECK_RUN_DIR=0 CWFLAGS="" CXFLAGS="${CFLAGS}" SYSTEMD_DIR=${systemd_system_unitdir} \ BINDIR="${base_sbindir}" UDEVDIR="${nonarch_base_libdir}/udev" LDFLAGS="${LDFLAGS}" \ - SYSROOT="${STAGING_DIR_TARGET}" STRIP=' + SYSROOT="${STAGING_DIR_TARGET}" MISCDIR="${datadir}/${BPN}" STRIP=' DEBUG_OPTIMIZATION:append = " -Wno-error"
meta/recipes-extended/mdadm/mdadm_4.5.bb CFLAGS:append:mipsarchn32 = ' -D__SANE_USERSPACE_TYPES__' · line 38
Quoted 7 lines · Ankur Tyagi via lists.openembedded.org
- dropped mdcheck install patch to address reproducibility issue
EXTRA_OEMAKE = 'CHECK_RUN_DIR=0 CWFLAGS="" CXFLAGS="${CFLAGS}" SYSTEMD_DIR=${systemd_system_unitdir} \ BINDIR="${base_sbindir}" UDEVDIR="${nonarch_base_libdir}/udev" LDFLAGS="${LDFLAGS}" \meta/recipes-extended/mdadm/mdadm_4.5.bb renamed from meta/recipes-extended/mdadm/mdadm_4.4.bb +1
-1- SYSROOT="${STAGING_DIR_TARGET}" STRIP=' + SYSROOT="${STAGING_DIR_TARGET}" MISCDIR="${datadir}/${BPN}" STRIP='
This has merged, and unfortunately the same reproducibility issue appeared in nightly master:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/3317
The issue seems to be that sometimes the sed command to patch BINDIR in mdcheck works, and sometimes it does not. I am baffled as to why that would be, but we can't have such fails in master, so for now I'm afraid this should be reverted.
Alex
Quoted 14 lines · Alexander Kanavin
meta/recipes-extended/mdadm/mdadm_4.5.bb CFLAGS:append:mipsarchn32 = ' -D__SANE_USERSPACE_TYPES__' · line 38
Quoted 7 lines · Ankur Tyagi via lists.openembedded.org
- dropped mdcheck install patch to address reproducibility issue
EXTRA_OEMAKE = 'CHECK_RUN_DIR=0 CWFLAGS="" CXFLAGS="${CFLAGS}" SYSTEMD_DIR=${systemd_system_unitdir} \ BINDIR="${base_sbindir}" UDEVDIR="${nonarch_base_libdir}/udev" LDFLAGS="${LDFLAGS}" \meta/recipes-extended/mdadm/mdadm_4.5.bb renamed from meta/recipes-extended/mdadm/mdadm_4.4.bb +1
-1- SYSROOT="${STAGING_DIR_TARGET}" STRIP=' + SYSROOT="${STAGING_DIR_TARGET}" MISCDIR="${datadir}/${BPN}" STRIP='This has merged, and unfortunately the same reproducibility issue appeared in nightly master:
Quoted 7 lines
https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/3317
The issue seems to be that sometimes the sed command to patch BINDIR in mdcheck works, and sometimes it does not. I am baffled as to why that would be, but we can't have such fails in master, so for now I'm afraid this should be reverted.
yeah, that's odd.
Let's revert it and I will spend some time understanding what's going on here.
Quoted 1 line
Alex
Quoted 13 lines · Ankur Tyagi
Quoted 13 lines
meta/recipes-extended/mdadm/mdadm_4.5.bb CFLAGS:append:mipsarchn32 = ' -D__SANE_USERSPACE_TYPES__' · line 38
Quoted 7 lines
- dropped mdcheck install patch to address reproducibility issue
EXTRA_OEMAKE = 'CHECK_RUN_DIR=0 CWFLAGS="" CXFLAGS="${CFLAGS}" SYSTEMD_DIR=${systemd_system_unitdir} \ BINDIR="${base_sbindir}" UDEVDIR="${nonarch_base_libdir}/udev" LDFLAGS="${LDFLAGS}" \meta/recipes-extended/mdadm/mdadm_4.5.bb renamed from meta/recipes-extended/mdadm/mdadm_4.4.bb +1
-1- SYSROOT="${STAGING_DIR_TARGET}" STRIP=' + SYSROOT="${STAGING_DIR_TARGET}" MISCDIR="${datadir}/${BPN}" STRIP='This has merged, and unfortunately the same reproducibility issue appeared in nightly master:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/3317
Quoted 8 lines
Quoted 5 lines
The issue seems to be that sometimes the sed command to patch BINDIR in mdcheck works, and sometimes it does not. I am baffled as to why that would be, but we can't have such fails in master, so for now I'm afraid this should be reverted.
yeah, that's odd.
Let's revert it and I will spend some time understanding what's going on here.
One idea is to try to catch the issue right when it happens, e.g. add a test patch that greps the offending file for BINDIR, and if that's a yes, then it loudly fails, stopping the build. We can take it into a staging branch and then once it fails someone can look into why sed didn't work.
Alex
Quoted 21 lines · Alexander Kanavin
Quoted 19 lines · Ankur Tyagi
Quoted 18 lines
meta/recipes-extended/mdadm/mdadm_4.5.bb CFLAGS:append:mipsarchn32 = ' -D__SANE_USERSPACE_TYPES__' · line 38
Quoted 7 lines
- dropped mdcheck install patch to address reproducibility issue
EXTRA_OEMAKE = 'CHECK_RUN_DIR=0 CWFLAGS="" CXFLAGS="${CFLAGS}" SYSTEMD_DIR=${systemd_system_unitdir} \ BINDIR="${base_sbindir}" UDEVDIR="${nonarch_base_libdir}/udev" LDFLAGS="${LDFLAGS}" \meta/recipes-extended/mdadm/mdadm_4.5.bb renamed from meta/recipes-extended/mdadm/mdadm_4.4.bb +1
-1- SYSROOT="${STAGING_DIR_TARGET}" STRIP=' + SYSROOT="${STAGING_DIR_TARGET}" MISCDIR="${datadir}/${BPN}" STRIP='This has merged, and unfortunately the same reproducibility issue appeared in nightly master:
https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/3317
The issue seems to be that sometimes the sed command to patch BINDIR in mdcheck works, and sometimes it does not. I am baffled as to why that would be, but we can't have such fails in master, so for now I'm afraid this should be reverted.
Quoted 9 lines
Quoted 2 lines
yeah, that's odd.
Let's revert it and I will spend some time understanding what's going on here.One idea is to try to catch the issue right when it happens, e.g. add a test patch that greps the offending file for BINDIR, and if that's a yes, then it loudly fails, stopping the build. We can take it into a staging branch and then once it fails someone can look into why sed didn't work.
Thanks Alex, I will try it out in the coming days and send a patch to monitor the situation.
Quoted 1 line
Alex