libaio.git
2 years ago[PATCH libaio 16/16] harness: Disable failing test 23
Guillem Jover [Tue, 19 Apr 2022 13:45:39 +0000 (15:45 +0200)]
[PATCH libaio 16/16] harness: Disable failing test 23

From b108f741b6ebd1e68a0392a72ba8e35537689ba0 Mon Sep 17 00:00:00 2001
Origin: vendor
Forwarded: not-needed

This new test fails on several architectures. But it is a new test
intended to check for a Linux kernel regression, so we are not worse
off than were before. Add it to the EXTRACASES variables, which has
currently the side effect of disabling it, as it is not handled
explicitly in the make target responsible for that variable.

Sent-Upstream: not-relevant
Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0016-harness-Disable-failing-test-23.patch

2 years ago[PATCH libaio 15/16] Fix io_pgetevents() syscall wrapper on 32-bit userland on 64...
Guillem Jover [Fri, 16 Aug 2019 00:33:46 +0000 (02:33 +0200)]
[PATCH libaio 15/16] Fix io_pgetevents() syscall wrapper on 32-bit userland on 64-bit kernels

From d3fe396f569d113e1259a77f875cde0ed52104ab Mon Sep 17 00:00:00 2001
Origin: vendor
Forwarded: not-needed

The kernel compat syscall in the kernel got introduced with a broken
layout, which requires a pointer to the actual sigset_t variable but
with the size of the running kernel, not the size of the compat code.

This means that when the wrapper sends the expected compat (32-bit)
pointer, the kernel reads a 64-bit pointer, eating with it also the
sigset size member. And then proceeds to fail the comparison of the
sigset_t size and returns EINVAL.

This really needs to be fixed in the kernel, as there's no apparent
user of the broken compat layout (from codesearch.debian.org, nor a
quick github.com search). But we have to workaround it in libaio so
that we can use kernels that have not yet been fixed.

We do that, by trying the non-broken layout (that would be used with
a 32-bit userland on a 32-bit kernel), and if that fails with -EINVAL
we retry with a structure padded to what the kernel expects.

Sent-Upstream: rejected
Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0015-Fix-io_pgetevents-syscall-wrapper-on-32-bit-userland.patch

2 years ago[PATCH libaio 14/16] Add x32 support
Guillem Jover [Sun, 12 Apr 2020 01:36:39 +0000 (03:36 +0200)]
[PATCH libaio 14/16] Add x32 support

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0014-Add-x32-support.patch

2 years ago[PATCH libaio 13/16] Add SH support
Nobuhiro Iwamatsu [Sun, 12 Apr 2020 01:30:41 +0000 (03:30 +0200)]
[PATCH libaio 13/16] Add SH support

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0013-Add-SH-support.patch

2 years ago[PATCH libaio 12/16] Add m68k support
Guillem Jover [Wed, 14 Aug 2019 01:25:22 +0000 (03:25 +0200)]
[PATCH libaio 12/16] Add m68k support

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0012-Add-m68k-support.patch

2 years ago[PATCH libaio 11/16] Add PARISC support
Guillem Jover [Wed, 14 Aug 2019 01:22:10 +0000 (03:22 +0200)]
[PATCH libaio 11/16] Add PARISC support

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0011-Add-PARISC-support.patch

2 years ago[PATCH libaio 10/16] Fix and improve MIPS support
Guillem Jover [Wed, 14 Aug 2019 01:18:50 +0000 (03:18 +0200)]
[PATCH libaio 10/16] Fix and improve MIPS support

Fix structure padding for MIPS 64, add syscall definitions, and the
kernel R/W pointer.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0010-Fix-and-improve-MIPS-support.patch

2 years ago[PATCH libaio 09/16] harness: Add KERNEL_RW_POINTER for various 32-bit architectures
Guillem Jover [Wed, 14 Aug 2019 01:30:09 +0000 (03:30 +0200)]
[PATCH libaio 09/16] harness: Add KERNEL_RW_POINTER for various 32-bit architectures

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0009-harness-Add-KERNEL_RW_POINTER-for-various-32-bit-arc.patch

2 years ago[PATCH libaio 08/16] syscall: Fix ia64 definitions to match exactly the kernel
Guillem Jover [Sun, 12 Apr 2020 01:44:43 +0000 (03:44 +0200)]
[PATCH libaio 08/16] syscall: Fix ia64 definitions to match exactly the kernel

Otherwise we get preprocessor warnings due to the redefinitions.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0008-syscall-Fix-ia64-definitions-to-match-exactly-the-ke.patch

2 years ago[PATCH libaio 07/16] syscall: Fix ARM definitions to match exactly the kernel
Guillem Jover [Sun, 12 Apr 2020 01:39:47 +0000 (03:39 +0200)]
[PATCH libaio 07/16] syscall: Fix ARM definitions to match exactly the kernel

Otherwise we get preprocessor warnings due to the redefinitions.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0007-syscall-Fix-ARM-definitions-to-match-exactly-the-ker.patch

2 years ago[PATCH libaio 06/16] Use new symver function attribute to support LTO builds
Guillem Jover [Tue, 2 Nov 2021 21:37:32 +0000 (22:37 +0100)]
[PATCH libaio 06/16] Use new symver function attribute to support LTO builds

From d2d9bd8ec04c3ddbea1f816da47475248135c515 Mon Sep 17 00:00:00 2001
Origin: vendor
Forwarded: https://marc.info/?l=linux-aio&m=164999309520544

The LTO support cannot work properly when there are versioned symbols
via asm statements, as those are not seen by the compiler. Use the new
function attributes if supported instead of the asm statements.

We need to move the SYMVER calls after the function definitions
otherwise it will reference symbols not yet seen.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0006-Use-new-symver-function-attribute-to-support-LTO-bui.patch

2 years ago[PATCH libaio 05/16] Move semicolon to SYMVER and DEFSYMVER call sites
Guillem Jover [Tue, 2 Nov 2021 23:20:50 +0000 (00:20 +0100)]
[PATCH libaio 05/16] Move semicolon to SYMVER and DEFSYMVER call sites

From ed76466ea5e551a916e3003b8ed3d7afd5248e9a Mon Sep 17 00:00:00 2001
Origin: vendor
Forwarded: https://marc.info/?l=linux-aio&m=164999309420541

These look like function macros, so let's unify their call sites to use
function syntax with a final semicolon.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0005-Move-semicolon-to-SYMVER-and-DEFSYMVER-call-sites.patch

2 years ago[PATCH libaio 04/16] build: Remove all test artifacts on clean
Guillem Jover [Sun, 6 Aug 2023 18:56:10 +0000 (20:56 +0200)]
[PATCH libaio 04/16] build: Remove all test artifacts on clean

We create these artifacts when running the test suite, so should remove
them when running the clean target.

Reported-by: Lucas Nussbaum <lucas@debian.org>
Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0004-build-Remove-all-test-artifacts-on-clean.patch

2 years ago[PATCH libaio 03/16] build: Honor user build flags
Guillem Jover [Sun, 12 Apr 2020 02:01:27 +0000 (04:01 +0200)]
[PATCH libaio 03/16] build: Honor user build flags

CPPFLAGS, CFLAGS and LDFLAGS are user flags, which we need to preserve,
regardless of the user passing them over the environment or the
command-line. Any required flag that the build system needs, has to be
set in some other flag (which we will namespace with «MK_») so that the
build will use even if the user flags are passed, even though the user
should be able to override by appending after them. We pass CPPFLAGS to
any compilation command that missed them.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0003-build-Honor-user-build-flags.patch

2 years ago[PATCH libaio 02/16] build: Fix DESTDIR and pathname variables handling
Guillem Jover [Sun, 12 Apr 2020 01:58:12 +0000 (03:58 +0200)]
[PATCH libaio 02/16] build: Fix DESTDIR and pathname variables handling

The convention for DESTDIR is that it is only prefixed during the
install target. In this case we postpone that to the inner Makefile
so that it can be invoked directly while preserving the expected
semantics. We should also prefer the user specified pathname variables
if already set in any way, and do not need to pass them explicitly
to the sub-make.

Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0002-build-Fix-DESTDIR-and-pathname-variables-handling.patch

2 years ago[PATCH libaio 01/16] man: Fix typos
Guillem Jover [Sun, 10 Oct 2021 04:14:51 +0000 (06:14 +0200)]
[PATCH libaio 01/16] man: Fix typos

From f95e1898b6efd3eaf50273bd1c5a8a40816b5160 Mon Sep 17 00:00:00 2001
Origin: vendor
Forwarded: https://marc.info/?l=linux-aio&m=164999309220532

Warned-by: lintian
Signed-off-by: Guillem Jover <guillem@hadrons.org>
Gbp-Pq: Name 0001-man-Fix-typos.patch

2 years agolibaio (0.3.113-5) unstable; urgency=medium
Guillem Jover [Sun, 6 Aug 2023 19:12:34 +0000 (20:12 +0100)]
libaio (0.3.113-5) unstable; urgency=medium

  * Wrap fields in debian/tests/control.
  * Fix upstream build system to remove all test artifacts on clean.
    Reported by Lucas Nussbaum <lucas@debian.org>.

[dgit import unpatched libaio 0.3.113-5]

2 years agoImport libaio_0.3.113-5.debian.tar.xz
Guillem Jover [Sun, 6 Aug 2023 19:12:34 +0000 (20:12 +0100)]
Import libaio_0.3.113-5.debian.tar.xz

[dgit import tarball libaio 0.3.113-5 libaio_0.3.113-5.debian.tar.xz]

3 years agoImport libaio_0.3.113.orig.tar.gz
Guillem Jover [Tue, 19 Apr 2022 01:22:49 +0000 (02:22 +0100)]
Import libaio_0.3.113.orig.tar.gz

[dgit import orig libaio_0.3.113.orig.tar.gz]