* Re: [PATCH 03/19] fsinfo: Provide a bitmap of supported features [ver #16]
From: David Howells @ 2020-02-20 12:22 UTC (permalink / raw)
To: Darrick J. Wong
Cc: dhowells, viro, raven, mszeredi, christian, linux-api,
linux-fsdevel, linux-kernel
In-Reply-To: <20200219163705.GC9496@magnolia>
Darrick J. Wong <darrick.wong@oracle.com> wrote:
> > +struct fsinfo_features {
> > + __u8 features[(FSINFO_FEAT__NR + 7) / 8];
>
> Hm...the structure size is pretty small (56 bytes) and will expand as we
> add new _FEAT flags. Is this ok because the fsinfo code will truncate
> its response to userspace to whatever buffer size userspace tells it?
Yes. Also, you can ask fsinfo how many feature bits it supports.
I should put this in the struct first rather than putting it elsewhere.
David
^ permalink raw reply
* [Bug 206525] BUG: KASAN: stack-out-of-bounds in test_bit+0x30/0x44 (kernel 5.6-rc1)
From: bugzilla-daemon @ 2020-02-20 12:19 UTC (permalink / raw)
To: linuxppc-dev
In-Reply-To: <bug-206525-206035@https.bugzilla.kernel.org/>
https://bugzilla.kernel.org/show_bug.cgi?id=206525
--- Comment #6 from Nikolay Aleksandrov (nikolay@cumulusnetworks.com) ---
Note that the bug wasn't introduced by my commit, but instead has been there
since:
commit 4f520900522f
Author: Richard Guy Briggs <rgb@redhat.com>
Date: Tue Apr 22 21:31:54 2014 -0400
netlink: have netlink per-protocol bind function return an error code.
which moved the ngroups test_bit() to a local variable. My commit only exposed
the bug since it added the 33rd group. I'm currently preparing a fix and will
post it to netdev after verifying and testing it.
--
You are receiving this mail because:
You are watching the assignee of the bug.
^ permalink raw reply
* Re: [PATCH 5/5] arm64/vdso: Restrict splitting VVAR VMA
From: Vincenzo Frascino @ 2020-02-20 12:22 UTC (permalink / raw)
To: Andrei Vagin
Cc: linux-arm-kernel, linux-kernel, Thomas Gleixner, Dmitry Safonov
In-Reply-To: <20200204175913.74901-6-avagin@gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1939 bytes --]
Hi Andrei,
On 04/02/2020 17:59, Andrei Vagin wrote:
> Forbid splitting VVAR VMA resulting in a stricter ABI and reducing the
> amount of corner-cases to consider while working further on VDSO time
> namespace support.
>
> As the offset from timens to VVAR page is computed compile-time, the pages
> in VVAR should stay together and not being partically mremap()'ed.
>
I agree on the concept, but why do we need to redefine mremap?
special_mapping_mremap() (mm/mmap.c +3317) seems doing already the same thing if
we leave mremap == NULL as is.
> Signed-off-by: Andrei Vagin <avagin@gmail.com>
> ---
> arch/arm64/kernel/vdso.c | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
>
> diff --git a/arch/arm64/kernel/vdso.c b/arch/arm64/kernel/vdso.c
> index 2e553468b183..e6ebdc184c1e 100644
> --- a/arch/arm64/kernel/vdso.c
> +++ b/arch/arm64/kernel/vdso.c
> @@ -229,6 +229,17 @@ static vm_fault_t vvar_fault(const struct vm_special_mapping *sm,
> return vmf_insert_pfn(vma, vmf->address, pfn);
> }
>
> +static int vvar_mremap(const struct vm_special_mapping *sm,
> + struct vm_area_struct *new_vma)
> +{
> + unsigned long new_size = new_vma->vm_end - new_vma->vm_start;
> +
> + if (new_size != VVAR_NR_PAGES * PAGE_SIZE)
> + return -EINVAL;
> +
> + return 0;
> +}
> +
> static int __setup_additional_pages(enum arch_vdso_type arch_index,
> struct mm_struct *mm,
> struct linux_binprm *bprm,
> @@ -311,6 +322,7 @@ static struct vm_special_mapping aarch32_vdso_spec[C_PAGES] = {
> {
> .name = "[vvar]",
> .fault = vvar_fault,
> + .mremap = vvar_mremap,
> },
> {
> .name = "[vdso]",
> @@ -493,6 +505,7 @@ static struct vm_special_mapping vdso_spec[A_PAGES] __ro_after_init = {
> {
> .name = "[vvar]",
> .fault = vvar_fault,
> + .mremap = vvar_mremap,
> },
> {
> .name = "[vdso]",
>
--
Regards,
Vincenzo
[-- Attachment #2: pEpkey.asc --]
[-- Type: application/pgp-keys, Size: 14291 bytes --]
^ permalink raw reply
* Re: [PATCH 5/5] arm64/vdso: Restrict splitting VVAR VMA
From: Vincenzo Frascino @ 2020-02-20 12:22 UTC (permalink / raw)
To: Andrei Vagin
Cc: Thomas Gleixner, linux-kernel, linux-arm-kernel, Dmitry Safonov
In-Reply-To: <20200204175913.74901-6-avagin@gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1939 bytes --]
Hi Andrei,
On 04/02/2020 17:59, Andrei Vagin wrote:
> Forbid splitting VVAR VMA resulting in a stricter ABI and reducing the
> amount of corner-cases to consider while working further on VDSO time
> namespace support.
>
> As the offset from timens to VVAR page is computed compile-time, the pages
> in VVAR should stay together and not being partically mremap()'ed.
>
I agree on the concept, but why do we need to redefine mremap?
special_mapping_mremap() (mm/mmap.c +3317) seems doing already the same thing if
we leave mremap == NULL as is.
> Signed-off-by: Andrei Vagin <avagin@gmail.com>
> ---
> arch/arm64/kernel/vdso.c | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
>
> diff --git a/arch/arm64/kernel/vdso.c b/arch/arm64/kernel/vdso.c
> index 2e553468b183..e6ebdc184c1e 100644
> --- a/arch/arm64/kernel/vdso.c
> +++ b/arch/arm64/kernel/vdso.c
> @@ -229,6 +229,17 @@ static vm_fault_t vvar_fault(const struct vm_special_mapping *sm,
> return vmf_insert_pfn(vma, vmf->address, pfn);
> }
>
> +static int vvar_mremap(const struct vm_special_mapping *sm,
> + struct vm_area_struct *new_vma)
> +{
> + unsigned long new_size = new_vma->vm_end - new_vma->vm_start;
> +
> + if (new_size != VVAR_NR_PAGES * PAGE_SIZE)
> + return -EINVAL;
> +
> + return 0;
> +}
> +
> static int __setup_additional_pages(enum arch_vdso_type arch_index,
> struct mm_struct *mm,
> struct linux_binprm *bprm,
> @@ -311,6 +322,7 @@ static struct vm_special_mapping aarch32_vdso_spec[C_PAGES] = {
> {
> .name = "[vvar]",
> .fault = vvar_fault,
> + .mremap = vvar_mremap,
> },
> {
> .name = "[vdso]",
> @@ -493,6 +505,7 @@ static struct vm_special_mapping vdso_spec[A_PAGES] __ro_after_init = {
> {
> .name = "[vvar]",
> .fault = vvar_fault,
> + .mremap = vvar_mremap,
> },
> {
> .name = "[vdso]",
>
--
Regards,
Vincenzo
[-- Attachment #2: pEpkey.asc --]
[-- Type: application/pgp-keys, Size: 14291 bytes --]
[-- Attachment #3: Type: text/plain, Size: 176 bytes --]
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply
* Re: [dpdk-dev] [PATCH 4/4] ci: reorganise Travis jobs
From: David Marchand @ 2020-02-20 12:22 UTC (permalink / raw)
To: Thomas Monjalon, Aaron Conole; +Cc: dev, Michael Santana, Bruce Richardson
In-Reply-To: <3732162.jZfb76A358@xps>
On Thu, Feb 20, 2020 at 11:42 AM Thomas Monjalon <thomas@monjalon.net> wrote:
>
> 19/02/2020 22:39, Aaron Conole:
> > David Marchand <david.marchand@redhat.com> writes:
> >
> > > Let's prune the jobs list to limit the amount of time spent by the robot
> > > in Travis.
> > >
> > > Since meson enables automatically the relevant components, there is not
> > > much gain in testing with extra_packages vs required_packages only.
> > >
> > > For a given arch/compiler/env combination, compilation is first tested
> > > in all jobs that run tests or build the docs or run the ABI checks.
> > > In the same context, for jobs that accumulates running tests, building
> > > the docs etc..., those steps are independent and can be split to save
> > > some cpu on Travis.
> > >
> > > With this, we go down from 21 to 15 jobs.
> > >
> > > Note: this patch requires a flush of the existing caches in Travis.
> > >
> > > Signed-off-by: David Marchand <david.marchand@redhat.com>
> > > ---
> >
> > In general, I think the idea with required vs. extra was to have a build
> > that did the minimum required, and one that did all the packages (to
> > allow a minimum vs. full DPDK).
> >
> > At least, that's from
> > http://mails.dpdk.org/archives/dev/2019-January/124007.html
>
> I think the benefit of a minimum build is to have a quick report,
> and easy to setup.
Yes, Travis serves as a first gate when submitting patches.
But since Travis is best effort/free, we can't have a full coverage.
> > Not sure if that's still something anyone cares about.
>
> Given that Travis knows how to satisfy the dependencies,
> and that we must wait for all jobs to finish,
> I don't see any benefit of a minimal setup.
This minimal setup also tests that dpdk dependencies are correct.
If a change makes something rely on libX and libX is in the packages
always installed in Travis, the missing dependency would not get
caught.
But here, this adds too many jobs.
UNH, Intel and other CIs should step in and fill this kind of gap.
--
David Marchand
^ permalink raw reply
* Re: [PATCH] libsolv: upgrade 0.7.10 -> 0.7.11
From: Richard Purdie @ 2020-02-20 12:24 UTC (permalink / raw)
To: Wang Mingyu, openembedded-core
In-Reply-To: <1582191660-130760-1-git-send-email-wangmy@cn.fujitsu.com>
On Thu, 2020-02-20 at 01:40 -0800, Wang Mingyu wrote:
> Signed-off-by: Wang Mingyu <wangmy@cn.fujitsu.com>
> ---
> .../libsolv/{libsolv_0.7.10.bb => libsolv_0.7.11.bb} | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
> rename meta/recipes-extended/libsolv/{libsolv_0.7.10.bb => libsolv_0.7.11.bb} (93%)
>
> diff --git a/meta/recipes-extended/libsolv/libsolv_0.7.10.bb b/meta/recipes-extended/libsolv/libsolv_0.7.11.bb
> similarity index 93%
> rename from meta/recipes-extended/libsolv/libsolv_0.7.10.bb
> rename to meta/recipes-extended/libsolv/libsolv_0.7.11.bb
> index 502f4e0e85..84349606f1 100644
> --- a/meta/recipes-extended/libsolv/libsolv_0.7.10.bb
> +++ b/meta/recipes-extended/libsolv/libsolv_0.7.11.bb
> @@ -10,7 +10,7 @@ DEPENDS = "expat zlib"
> SRC_URI = "git://github.com/openSUSE/libsolv.git \
> "
>
> -SRCREV = "605dd2645ef899e2b7c95709476fb51e28d7e378"
> +SRCREV = "102ac552669181d3e43731aa275bdbbc7ddcb94c"
>
> UPSTREAM_CHECK_GITTAGREGEX = "(?P<pver>\d+(\.\d+)+)"
>
I think this causes:
https://autobuilder.yoctoproject.org/typhoon/#/builders/61/builds/1591
ERROR: core-image-sato-1.0-r0 do_testimage: core-image-sato - FAILED - check the task log and the ssh log
ERROR: Logfile of failure stored in: /home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/build/tmp/work/qemux86-poky-linux/core-image-sato/1.0-r0/temp/log.do_testimage.5108
NOTE: recipe core-image-sato-1.0-r0: task do_testimage: Failed
ERROR: Task (/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/recipes-sato/images/core-image-sato.bb:do_testimage) failed with exit code '1'
Test requires gtk+3 to be installed
Started HTTPService on 0.0.0.0:35043
Traceback (most recent call last):
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/core/decorator/__init__.py", line 36, in wrapped_f
return func(*args, **kwargs)
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/dnf.py", line 181, in test_dnf_exclude
self.dnf_with_repo('install -y curl*')
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/dnf.py", line 69, in dnf_with_repo
output = self.dnf(" ".join(cmdlinerepoopts) + " --nogpgcheck " + command)
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/dnf.py", line 21, in dnf
self.assertEqual(status, expected, message)
AssertionError: 1 != 0 : dnf --repofrompath=oe-testimage-repo-qemux86,http://192.168.7.3:35043/qemux86 --repofrompath=oe-testimage-repo-noarch,http://192.168.7.3:35043/noarch --repofrompath=oe-testimage-repo-core2_32,http://192.168.7.3:35043/core2_32 --nogpgcheck install -y curl*
Added oe-testimage-repo-qemux86 repo from http://192.168.7.3:35043/qemux86
Added oe-testimage-repo-noarch repo from http://192.168.7.3:35043/noarch
Added oe-testimage-repo-core2_32 repo from http://192.168.7.3:35043/core2_32
Last metadata expiration check: 0:00:01 ago on Thu Feb 20 12:13:09 2020.
Package curl-7.68.0-r0.core2_32 is already installed.
Package curl-dev-7.68.0-r0.core2_32 is already installed.
Dependencies resolved.
================================================================================
Package Architecture Version Repository Size
================================================================================
Installing:
curl-dbg core2_32 7.68.0-r0 oe-testimage-repo-core2_32 36 k
curl-doc core2_32 7.68.0-r0 oe-testimage-repo-core2_32 269 k
Transaction Summary
================================================================================
Install 2 Packages
/usr/lib/python3.8/site-packages/dnf/package.py:57: DeprecationWarning: PY_SSIZE_T_CLEAN will be required for '#' formats
return super(Package, self).chksum
Total download size: 305 k
Installed size: 1.6 M
Downloading Packages:
(1/2): curl-dbg-7.68.0-r0.core2_32.rpm 2.6 MB/s | 36 kB 00:00
(2/2): curl-doc-7.68.0-r0.core2_32.rpm 16 MB/s | 269 kB 00:00
--------------------------------------------------------------------------------
Total 14 MB/s | 305 kB 00:00
Running transaction check
Transaction check succeeded.
Running transaction test
Transaction test succeeded.
Running transaction
Failed to obtain the transaction lock (logged in as: root).
The downloaded packages were saved in cache until the next successful transaction.
You can remove cached packages by executing 'dnf clean packages'.
Error: Could not run transaction.
Traceback (most recent call last):
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/core/decorator/__init__.py", line 36, in wrapped_f
return func(*args, **kwargs)
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/dnf.py", line 90, in test_dnf_install
self.dnf_with_repo('remove -y run-postinsts-dev')
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/dnf.py", line 69, in dnf_with_repo
output = self.dnf(" ".join(cmdlinerepoopts) + " --nogpgcheck " + command)
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/dnf.py", line 21, in dnf
self.assertEqual(status, expected, message)
AssertionError: 1 != 0 : dnf --repofrompath=oe-testimage-repo-qemux86,http://192.168.7.3:35043/qemux86 --repofrompath=oe-testimage-repo-noarch,http://192.168.7.3:35043/noarch --repofrompath=oe-testimage-repo-core2_32,http://192.168.7.3:35043/core2_32 --nogpgcheck remove -y run-postinsts-dev
Added oe-testimage-repo-qemux86 repo from http://192.168.7.3:35043/qemux86
Added oe-testimage-repo-noarch repo from http://192.168.7.3:35043/noarch
Added oe-testimage-repo-core2_32 repo from http://192.168.7.3:35043/core2_32
Dependencies resolved.
================================================================================
Package Architecture Version Repository Size
================================================================================
Removing:
run-postinsts-dev noarch 1.0-r10 @oe-repo 0
Transaction Summary
================================================================================
Remove 1 Package
Freed space: 0
Running transaction check
Transaction check succeeded.
Running transaction test
Transaction test succeeded.
Running transaction
Failed to obtain the transaction lock (logged in as: root).
Error: Could not run transaction.
Test case dnf.DnfRepoTest.test_dnf_install_dependency depends on dnf.DnfRepoTest.test_dnf_install but it didn't pass/run.
Test case dnf.DnfRepoTest.test_dnf_install_from_disk depends on dnf.DnfRepoTest.test_dnf_install_dependency but it didn't pass/run.
Test case dnf.DnfRepoTest.test_dnf_install_from_http depends on dnf.DnfRepoTest.test_dnf_install_from_disk but it didn't pass/run.
Test run when enable usrmege
Test case dnf.DnfRepoTest.test_dnf_reinstall depends on dnf.DnfRepoTest.test_dnf_install but it didn't pass/run.
Stopped HTTPService on 0.0.0.0:35043
Test requires opkg to be installed
Test requires pam to be in DISTRO_FEATURES
Traceback (most recent call last):
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/core/decorator/__init__.py", line 36, in wrapped_f
return func(*args, **kwargs)
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/parselogs.py", line 370, in test_parselogs
self.assertEqual(errcount, 0, msg=self.msg)
AssertionError: 6 != 0 : Log: /home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/build/tmp/work/qemux86-poky-linux/core-image-sato-sdk/1.0-r0/target_logs/dnf.log
-----------------------
Central error: 2020-02-20T12:13:10Z CRITICAL Failed to obtain the transaction lock (logged in as: root).
***********************
2020-02-20T12:13:10Z INFO Total 14 MB/s | 305 kB 00:00
2020-02-20T12:13:10Z INFO Running transaction check
2020-02-20T12:13:10Z INFO Transaction check succeeded.
2020-02-20T12:13:10Z INFO Running transaction test
2020-02-20T12:13:10Z INFO Transaction test succeeded.
2020-02-20T12:13:10Z DDEBUG timer: transaction test: 145 ms
2020-02-20T12:13:10Z INFO Running transaction
2020-02-20T12:13:10Z DEBUG RPMDB altered outside of DNF.
2020-02-20T12:13:10Z DDEBUG RPM transaction start.
2020-02-20T12:13:10Z DDEBUG RPM transaction over.
2020-02-20T12:13:10Z CRITICAL Failed to obtain the transaction lock (logged in as: root).
2020-02-20T12:13:10Z DDEBUG Cleaning up.
2020-02-20T12:13:10Z INFO The downloaded packages were saved in cache until the next successful transaction.
2020-02-20T12:13:10Z INFO You can remove cached packages by executing 'dnf clean packages'.
2020-02-20T12:13:10Z SUBDEBUG
Traceback (most recent call last):
File "/usr/lib/python3.8/site-packages/dnf/cli/main.py", line 64, in main
return _main(base, args, cli_class, option_parser_class)
File "/usr/lib/python3.8/site-packages/dnf/cli/main.py", line 99, in _main
return cli_run(cli, base)
File "/usr/lib/python3.8/site-packages/dnf/cli/main.py", line 123, in cli_run
***********************
Central error: raise dnf.exceptions.Error(msg)
***********************
return cli_run(cli, base)
File "/usr/lib/python3.8/site-packages/dnf/cli/main.py", line 123, in cli_run
ret = resolving(cli, base)
File "/usr/lib/python3.8/site-packages/dnf/cli/main.py", line 167, in resolving
base.do_transaction(display=displays)
File "/usr/lib/python3.8/site-packages/dnf/cli/cli.py", line 225, in do_transaction
tid = super(BaseCli, self).do_transaction(display)
File "/usr/lib/python3.8/site-packages/dnf/base.py", line 882, in do_transaction
tid = self._run_transaction(cb=cb)
File "/usr/lib/python3.8/site-packages/dnf/base.py", line 1005, in _run_transaction
raise dnf.exceptions.Error(msg)
dnf.exceptions.Error: Could not run transaction.
2020-02-20T12:13:10Z CRITICAL Error: Could not run transaction.
2020-02-20T12:13:10Z INFO --- logging initialized ---
2020-02-20T12:13:10Z DDEBUG timer: config: 3 msa
2020-02-20T12:13:10Z DEBUG DNF version: 4.2.2
2020-02-20T12:13:10Z DDEBUG Command: dnf --repofrompath=oe-testimage-repo-qemux86,http://192.168.7.3:35043/qemux86 --repofrompath=oe-testimage-repo-noarch,http://192.168.7.3:35043/noarch --repofrompath=oe-testimage-repo-core2_32,http://192.168.7.3:35043/core2_32 --nogpgcheck list run-postinsts-dev
2020-02-20T12:13:10Z DDEBUG Installroot: /
2020-02-20T12:13:10Z DDEBUG Releasever: master
2020-02-20T12:13:10Z DEBUG cachedir: /var/cache/dnf
2020-02-20T12:13:10Z DDEBUG Base command: list
est requires ptest-runner to be installed
Traceback (most recent call last):
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/core/decorator/__init__.py", line 36, in wrapped_f
return func(*args, **kwargs)
File "/home/pokybuild/yocto-worker/pkgman-rpm-non-rpm/build/meta/lib/oeqa/runtime/cases/rpm.py", line 100, in test_rpm_install
self.assertEqual(status, 0, msg=msg)
AssertionError: 1 != 0 : Failed to install base-passwd-doc package: error: db5 error(-30969) from dbenv->open: BDB0091 DB_VERSION_MISMATCH: Database environment version mismatch
error: cannot open Packages index using db5 - (-30969)
error: cannot open Packages database in /var/lib/rpm
^ permalink raw reply
* Re: [PATCH] firmware: imx: Align imx SC msg structs to 4
From: Leonard Crestez @ 2020-02-20 12:25 UTC (permalink / raw)
To: Stephen Boyd
Cc: Aisheng Dong, Shawn Guo, Fabio Estevam, Michael Turquette,
Stefan Agner, Linus Walleij, Alessandro Zummo, Alexandre Belloni,
Anson Huang, Abel Vesa, Franck Lenormand, kernel@pengutronix.de,
dl-linux-imx, linux-arm-kernel@lists.infradead.org,
linux-clk@vger.kernel.org, linux-gpio@vger.kernel.org,
linux-rtc@vger.kernel.org
In-Reply-To: <158215662160.184098.12475377410437970578@swboyd.mtv.corp.google.com>
On 20.02.2020 01:57, Stephen Boyd wrote:
> Quoting Leonard Crestez (2020-02-11 13:24:33)
>> The imx SC api strongly assumes that messages are composed out of
>> 4-bytes words but some of our message structs have sizeof "6" and "7".
>>
>> This produces many oopses with CONFIG_KASAN=y:
>>
>> BUG: KASAN: stack-out-of-bounds in imx_mu_send_data+0x108/0x1f0
>
> Can you share the full kasan bug report instead of the single line?
[ 1.606708] imx-scu scu: NXP i.MX SCU Initialized
[ 1.635265] random: fast init done
[ 1.652200]
==================================================================
[ 1.659118] BUG: KASAN: stack-out-of-bounds in
imx_mu_send_data+0x108/0x1f0
[ 1.666046] Read of size 4 at addr ffff0008c80e6bc4 by task swapper/0/1
[ 1.672642]
[ 1.674134] CPU: 0 PID: 1 Comm: swapper/0 Not tainted
5.4.3-03848-g13efcd6 #54
[ 1.681335] Hardware name: Freescale i.MX8QM MEK (DT)
[ 1.686373] Call trace:
[ 1.688815] dump_backtrace+0x0/0x1e8
[ 1.692458] show_stack+0x14/0x20
[ 1.695766] dump_stack+0xe8/0x140
[ 1.699155] print_address_description.isra.11+0x64/0x348
[ 1.704532] __kasan_report+0x11c/0x230
[ 1.708356] kasan_report+0xc/0x18
[ 1.711743] __asan_load4+0x90/0xb0
[ 1.715218] imx_mu_send_data+0x108/0x1f0
[ 1.719215] msg_submit+0x104/0x180
[ 1.722689] mbox_send_message+0xa8/0x1a0
[ 1.726696] imx_scu_call_rpc+0x168/0x310
[ 1.730679] imx_sc_pd_power+0x180/0x1e0
[ 1.734589] imx_sc_pd_power_on+0x10/0x18
[ 1.738598] genpd_power_on.part.23+0x118/0x2a8
[ 1.743105] genpd_runtime_resume+0x138/0x320
[ 1.747454] __rpm_callback+0xb0/0x1a0
[ 1.751184] rpm_callback+0x34/0xe0
[ 1.754659] rpm_resume+0x5b8/0x7e8
[ 1.758137] __pm_runtime_resume+0x38/0x90
[ 1.762222] imx_clk_scu_probe+0x5c/0x1c8
[ 1.766218] platform_drv_probe+0x6c/0xc8
[ 1.770217] really_probe+0x148/0x428
[ 1.773861] driver_probe_device+0x74/0x130
[ 1.778031] __device_attach_driver+0xc4/0xe8
[ 1.782380] bus_for_each_drv+0xf0/0x158
[ 1.786283] __device_attach+0x158/0x1d8
[ 1.790195] device_initial_probe+0x10/0x18
[ 1.794362] bus_probe_device+0xe0/0xf0
[ 1.798185] device_add+0x660/0x998
[ 1.801659] platform_device_add+0x198/0x340
[ 1.805916] imx_clk_scu_alloc_dev+0x1b8/0x1e8
[ 1.810347] imx8qxp_clk_probe+0x19d0/0x28b8
[ 1.814601] platform_drv_probe+0x6c/0xc8
[ 1.818601] really_probe+0x148/0x428
[ 1.822250] driver_probe_device+0x74/0x130
[ 1.826423] __device_attach_driver+0xc4/0xe8
[ 1.830763] bus_for_each_drv+0xf0/0x158
[ 1.834675] __device_attach+0x158/0x1d8
[ 1.838583] device_initial_probe+0x10/0x18
[ 1.842752] bus_probe_device+0xe0/0xf0
[ 1.846572] device_add+0x660/0x998
[ 1.850058] of_device_add+0x74/0x98
[ 1.853610] of_platform_device_create_pdata+0x11c/0x178
[ 1.858908] of_platform_bus_create+0x404/0x4f0
[ 1.863425] of_platform_populate+0x74/0x110
[ 1.867688] devm_of_platform_populate+0x54/0xb8
[ 1.872291] imx_scu_probe+0x1b8/0x220
[ 1.876022] platform_drv_probe+0x6c/0xc8
[ 1.880021] really_probe+0x148/0x428
[ 1.883671] driver_probe_device+0x74/0x130
[ 1.887841] device_driver_attach+0x94/0xa0
[ 1.892010] __driver_attach+0x70/0x110
[ 1.895832] bus_for_each_dev+0xe8/0x158
[ 1.899741] driver_attach+0x30/0x40
[ 1.903303] bus_add_driver+0x1b0/0x2b8
[ 1.907129] driver_register+0xbc/0x1d0
[ 1.910947] __platform_driver_register+0x7c/0x88
[ 1.915653] imx_scu_driver_init+0x18/0x20
[ 1.919725] do_one_initcall+0xd4/0x244
[ 1.923552] kernel_init_freeable+0x238/0x2d4
[ 1.927889] kernel_init+0x10/0x114
[ 1.931365] ret_from_fork+0x10/0x18
[ 1.934917]
[ 1.936399] The buggy address belongs to the page:
[ 1.941184] page:fffffe0023003980 refcount:0 mapcount:0
mapping:0000000000000000 index:0x0
[ 1.949447] raw: 1ffff00000000000 fffffe0023003988 fffffe0023003988
0000000000000000
[ 1.957170] raw: 0000000000000000 0000000000000000 00000000ffffffff
0000000000000000
[ 1.964894] page dumped because: kasan: bad access detected
[ 1.970449]
[ 1.971933] addr ffff0008c80e6bc4 is located in stack of task
swapper/0/1 at offset 36 in frame:
[ 1.980708] imx_sc_pd_power+0x0/0x1e0
[ 1.984442]
[ 1.985917] this frame has 1 object:
[ 1.989481] [32, 39) 'msg'
[ 1.989484]
[ 1.993732] Memory state around the buggy address:
[ 1.998520] ffff0008c80e6a80: 00 00 00 00 00 00 00 00 00 00 f1 f1 f1
f1 00 00
[ 2.005730] ffff0008c80e6b00: 00 00 f3 f3 f3 f3 00 00 00 00 00 00 00
00 00 00
[ 2.012940] >ffff0008c80e6b80: 00 00 00 00 f1 f1 f1 f1 07 f2 f2 f2 00
00 00 00
[ 2.020141] ^
[ 2.025448] ffff0008c80e6c00: 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00
[ 2.032659] ffff0008c80e6c80: 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00
[ 2.039862]
==================================================================
This is actually from an NXP release branch but code is very close up
upstream.
>> It shouldn't cause an issues in normal use because these structs are
>> always allocated on the stack.
>
> Is packed necessary for these? I thought that if the beginning of the
> struct was naturally aligned and there was sometimes a byte or two at
> the end then having __packed wasn't useful. So maybe it's better to just
> drop __packed on all these structs and let the compiler decide it can
> add some padding on the stack? Or do we have arrays of these structs
> sitting in memory all next to each other and they need to be that way to
> be sent through the mailbox?
I'm not sure I understand the question: the structs are __packed because
they represent the binary protocol for communicating with the "System
Controller".
Without __packed the compiler could insert padding inside the structs
and break the protocol.
As far as I understand compilers are still allowed to use padding on
stack since that padding is outside the message struct itself.
--
Regards,
Leonard
^ permalink raw reply
* Re: [PATCH] firmware: imx: Align imx SC msg structs to 4
From: Leonard Crestez @ 2020-02-20 12:25 UTC (permalink / raw)
To: Stephen Boyd
Cc: Aisheng Dong, Alessandro Zummo, Alexandre Belloni, Abel Vesa,
Anson Huang, Michael Turquette, Stefan Agner,
linux-clk@vger.kernel.org, Franck Lenormand,
linux-gpio@vger.kernel.org, dl-linux-imx, kernel@pengutronix.de,
Fabio Estevam, Shawn Guo, Linus Walleij,
linux-arm-kernel@lists.infradead.org, linux-rtc@vger.kernel.org
In-Reply-To: <158215662160.184098.12475377410437970578@swboyd.mtv.corp.google.com>
On 20.02.2020 01:57, Stephen Boyd wrote:
> Quoting Leonard Crestez (2020-02-11 13:24:33)
>> The imx SC api strongly assumes that messages are composed out of
>> 4-bytes words but some of our message structs have sizeof "6" and "7".
>>
>> This produces many oopses with CONFIG_KASAN=y:
>>
>> BUG: KASAN: stack-out-of-bounds in imx_mu_send_data+0x108/0x1f0
>
> Can you share the full kasan bug report instead of the single line?
[ 1.606708] imx-scu scu: NXP i.MX SCU Initialized
[ 1.635265] random: fast init done
[ 1.652200]
==================================================================
[ 1.659118] BUG: KASAN: stack-out-of-bounds in
imx_mu_send_data+0x108/0x1f0
[ 1.666046] Read of size 4 at addr ffff0008c80e6bc4 by task swapper/0/1
[ 1.672642]
[ 1.674134] CPU: 0 PID: 1 Comm: swapper/0 Not tainted
5.4.3-03848-g13efcd6 #54
[ 1.681335] Hardware name: Freescale i.MX8QM MEK (DT)
[ 1.686373] Call trace:
[ 1.688815] dump_backtrace+0x0/0x1e8
[ 1.692458] show_stack+0x14/0x20
[ 1.695766] dump_stack+0xe8/0x140
[ 1.699155] print_address_description.isra.11+0x64/0x348
[ 1.704532] __kasan_report+0x11c/0x230
[ 1.708356] kasan_report+0xc/0x18
[ 1.711743] __asan_load4+0x90/0xb0
[ 1.715218] imx_mu_send_data+0x108/0x1f0
[ 1.719215] msg_submit+0x104/0x180
[ 1.722689] mbox_send_message+0xa8/0x1a0
[ 1.726696] imx_scu_call_rpc+0x168/0x310
[ 1.730679] imx_sc_pd_power+0x180/0x1e0
[ 1.734589] imx_sc_pd_power_on+0x10/0x18
[ 1.738598] genpd_power_on.part.23+0x118/0x2a8
[ 1.743105] genpd_runtime_resume+0x138/0x320
[ 1.747454] __rpm_callback+0xb0/0x1a0
[ 1.751184] rpm_callback+0x34/0xe0
[ 1.754659] rpm_resume+0x5b8/0x7e8
[ 1.758137] __pm_runtime_resume+0x38/0x90
[ 1.762222] imx_clk_scu_probe+0x5c/0x1c8
[ 1.766218] platform_drv_probe+0x6c/0xc8
[ 1.770217] really_probe+0x148/0x428
[ 1.773861] driver_probe_device+0x74/0x130
[ 1.778031] __device_attach_driver+0xc4/0xe8
[ 1.782380] bus_for_each_drv+0xf0/0x158
[ 1.786283] __device_attach+0x158/0x1d8
[ 1.790195] device_initial_probe+0x10/0x18
[ 1.794362] bus_probe_device+0xe0/0xf0
[ 1.798185] device_add+0x660/0x998
[ 1.801659] platform_device_add+0x198/0x340
[ 1.805916] imx_clk_scu_alloc_dev+0x1b8/0x1e8
[ 1.810347] imx8qxp_clk_probe+0x19d0/0x28b8
[ 1.814601] platform_drv_probe+0x6c/0xc8
[ 1.818601] really_probe+0x148/0x428
[ 1.822250] driver_probe_device+0x74/0x130
[ 1.826423] __device_attach_driver+0xc4/0xe8
[ 1.830763] bus_for_each_drv+0xf0/0x158
[ 1.834675] __device_attach+0x158/0x1d8
[ 1.838583] device_initial_probe+0x10/0x18
[ 1.842752] bus_probe_device+0xe0/0xf0
[ 1.846572] device_add+0x660/0x998
[ 1.850058] of_device_add+0x74/0x98
[ 1.853610] of_platform_device_create_pdata+0x11c/0x178
[ 1.858908] of_platform_bus_create+0x404/0x4f0
[ 1.863425] of_platform_populate+0x74/0x110
[ 1.867688] devm_of_platform_populate+0x54/0xb8
[ 1.872291] imx_scu_probe+0x1b8/0x220
[ 1.876022] platform_drv_probe+0x6c/0xc8
[ 1.880021] really_probe+0x148/0x428
[ 1.883671] driver_probe_device+0x74/0x130
[ 1.887841] device_driver_attach+0x94/0xa0
[ 1.892010] __driver_attach+0x70/0x110
[ 1.895832] bus_for_each_dev+0xe8/0x158
[ 1.899741] driver_attach+0x30/0x40
[ 1.903303] bus_add_driver+0x1b0/0x2b8
[ 1.907129] driver_register+0xbc/0x1d0
[ 1.910947] __platform_driver_register+0x7c/0x88
[ 1.915653] imx_scu_driver_init+0x18/0x20
[ 1.919725] do_one_initcall+0xd4/0x244
[ 1.923552] kernel_init_freeable+0x238/0x2d4
[ 1.927889] kernel_init+0x10/0x114
[ 1.931365] ret_from_fork+0x10/0x18
[ 1.934917]
[ 1.936399] The buggy address belongs to the page:
[ 1.941184] page:fffffe0023003980 refcount:0 mapcount:0
mapping:0000000000000000 index:0x0
[ 1.949447] raw: 1ffff00000000000 fffffe0023003988 fffffe0023003988
0000000000000000
[ 1.957170] raw: 0000000000000000 0000000000000000 00000000ffffffff
0000000000000000
[ 1.964894] page dumped because: kasan: bad access detected
[ 1.970449]
[ 1.971933] addr ffff0008c80e6bc4 is located in stack of task
swapper/0/1 at offset 36 in frame:
[ 1.980708] imx_sc_pd_power+0x0/0x1e0
[ 1.984442]
[ 1.985917] this frame has 1 object:
[ 1.989481] [32, 39) 'msg'
[ 1.989484]
[ 1.993732] Memory state around the buggy address:
[ 1.998520] ffff0008c80e6a80: 00 00 00 00 00 00 00 00 00 00 f1 f1 f1
f1 00 00
[ 2.005730] ffff0008c80e6b00: 00 00 f3 f3 f3 f3 00 00 00 00 00 00 00
00 00 00
[ 2.012940] >ffff0008c80e6b80: 00 00 00 00 f1 f1 f1 f1 07 f2 f2 f2 00
00 00 00
[ 2.020141] ^
[ 2.025448] ffff0008c80e6c00: 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00
[ 2.032659] ffff0008c80e6c80: 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00
[ 2.039862]
==================================================================
This is actually from an NXP release branch but code is very close up
upstream.
>> It shouldn't cause an issues in normal use because these structs are
>> always allocated on the stack.
>
> Is packed necessary for these? I thought that if the beginning of the
> struct was naturally aligned and there was sometimes a byte or two at
> the end then having __packed wasn't useful. So maybe it's better to just
> drop __packed on all these structs and let the compiler decide it can
> add some padding on the stack? Or do we have arrays of these structs
> sitting in memory all next to each other and they need to be that way to
> be sent through the mailbox?
I'm not sure I understand the question: the structs are __packed because
they represent the binary protocol for communicating with the "System
Controller".
Without __packed the compiler could insert padding inside the structs
and break the protocol.
As far as I understand compilers are still allowed to use padding on
stack since that padding is outside the message struct itself.
--
Regards,
Leonard
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
^ permalink raw reply
* Re: [PATCH v3 05/37] s390/mm: provide memory management functions for protected KVM guests
From: David Hildenbrand @ 2020-02-20 12:25 UTC (permalink / raw)
To: Christian Borntraeger, Janosch Frank
Cc: KVM, Cornelia Huck, Thomas Huth, Ulrich Weigand, Claudio Imbrenda,
linux-s390, Michael Mueller, Vasily Gorbik, Andrea Arcangeli,
linux-mm
In-Reply-To: <20200220104020.5343-6-borntraeger@de.ibm.com>
On 20.02.20 11:39, Christian Borntraeger wrote:
> From: Claudio Imbrenda <imbrenda@linux.ibm.com>
>
> This provides the basic ultravisor calls and page table handling to cope
> with secure guests:
> - provide arch_make_page_accessible
> - make pages accessible after unmapping of secure guests
> - provide the ultravisor commands convert to/from secure
> - provide the ultravisor commands pin/unpin shared
> - provide callbacks to make pages secure (inacccessible)
> - we check for the expected pin count to only make pages secure if the
> host is not accessing them
> - we fence hugetlbfs for secure pages
>
> Co-developed-by: Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
> Signed-off-by: Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
> Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
> [borntraeger@de.ibm.com: patch merging, splitting, fixing]
> Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Acked-by: David Hildenbrand <david@redhat.com>
--
Thanks,
David / dhildenb
^ permalink raw reply
* [PATCH 1/4] ceph: always renew caps if mds_wanted is insufficient
From: Yan, Zheng @ 2020-02-20 12:26 UTC (permalink / raw)
To: ceph-devel; +Cc: jlayton, Yan, Zheng
Not only after mds closes session and caps get dropped. This is
preparation patch for not requesting caps for idle open files.
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
---
fs/ceph/caps.c | 36 +++++++++++++++---------------------
fs/ceph/mds_client.c | 5 -----
fs/ceph/super.h | 1 -
3 files changed, 15 insertions(+), 27 deletions(-)
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index d05717397c2a..293920d013ff 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2659,6 +2659,7 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
}
} else {
int session_readonly = false;
+ int mds_wanted;
if (ci->i_auth_cap &&
(need & (CEPH_CAP_FILE_WR | CEPH_CAP_FILE_EXCL))) {
struct ceph_mds_session *s = ci->i_auth_cap->session;
@@ -2667,32 +2668,27 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
spin_unlock(&s->s_cap_lock);
}
if (session_readonly) {
- dout("get_cap_refs %p needed %s but mds%d readonly\n",
+ dout("get_cap_refs %p need %s but mds%d readonly\n",
inode, ceph_cap_string(need), ci->i_auth_cap->mds);
ret = -EROFS;
goto out_unlock;
}
- if (ci->i_ceph_flags & CEPH_I_CAP_DROPPED) {
- int mds_wanted;
- if (READ_ONCE(mdsc->fsc->mount_state) ==
- CEPH_MOUNT_SHUTDOWN) {
- dout("get_cap_refs %p forced umount\n", inode);
- ret = -EIO;
- goto out_unlock;
- }
- mds_wanted = __ceph_caps_mds_wanted(ci, false);
- if (need & ~(mds_wanted & need)) {
- dout("get_cap_refs %p caps were dropped"
- " (session killed?)\n", inode);
- ret = -ESTALE;
- goto out_unlock;
- }
- if (!(file_wanted & ~mds_wanted))
- ci->i_ceph_flags &= ~CEPH_I_CAP_DROPPED;
+ if (READ_ONCE(mdsc->fsc->mount_state) == CEPH_MOUNT_SHUTDOWN) {
+ dout("get_cap_refs %p forced umount\n", inode);
+ ret = -EIO;
+ goto out_unlock;
+ }
+ mds_wanted = __ceph_caps_mds_wanted(ci, false);
+ if (need & ~mds_wanted) {
+ dout("get_cap_refs %p need %s > mds_wanted %s\n",
+ inode, ceph_cap_string(need),
+ ceph_cap_string(mds_wanted));
+ ret = -ESTALE;
+ goto out_unlock;
}
- dout("get_cap_refs %p have %s needed %s\n", inode,
+ dout("get_cap_refs %p have %s need %s\n", inode,
ceph_cap_string(have), ceph_cap_string(need));
}
out_unlock:
@@ -3646,8 +3642,6 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex,
goto out_unlock;
if (target < 0) {
- if (cap->mds_wanted | cap->issued)
- ci->i_ceph_flags |= CEPH_I_CAP_DROPPED;
__ceph_remove_cap(cap, false);
goto out_unlock;
}
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index fab9d6461a65..98d746b3bb53 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1411,8 +1411,6 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
dout("removing cap %p, ci is %p, inode is %p\n",
cap, ci, &ci->vfs_inode);
spin_lock(&ci->i_ceph_lock);
- if (cap->mds_wanted | cap->issued)
- ci->i_ceph_flags |= CEPH_I_CAP_DROPPED;
__ceph_remove_cap(cap, false);
if (!ci->i_auth_cap) {
struct ceph_cap_flush *cf;
@@ -1578,9 +1576,6 @@ static int wake_up_session_cb(struct inode *inode, struct ceph_cap *cap,
/* mds did not re-issue stale cap */
spin_lock(&ci->i_ceph_lock);
cap->issued = cap->implemented = CEPH_CAP_PIN;
- /* make sure mds knows what we want */
- if (__ceph_caps_file_wanted(ci) & ~cap->mds_wanted)
- ci->i_ceph_flags |= CEPH_I_CAP_DROPPED;
spin_unlock(&ci->i_ceph_lock);
}
} else if (ev == FORCE_RO) {
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 37dc1ac8f6c3..d370f89df358 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -517,7 +517,6 @@ static inline struct inode *ceph_find_inode(struct super_block *sb,
#define CEPH_I_POOL_RD (1 << 4) /* can read from pool */
#define CEPH_I_POOL_WR (1 << 5) /* can write to pool */
#define CEPH_I_SEC_INITED (1 << 6) /* security initialized */
-#define CEPH_I_CAP_DROPPED (1 << 7) /* caps were forcibly dropped */
#define CEPH_I_KICK_FLUSH (1 << 8) /* kick flushing caps */
#define CEPH_I_FLUSH_SNAPS (1 << 9) /* need flush snapss */
#define CEPH_I_ERROR_WRITE (1 << 10) /* have seen write errors */
--
2.21.1
^ permalink raw reply related
* [PATCH 2/4] ceph: consider file's last read/write when calculating wanted caps
From: Yan, Zheng @ 2020-02-20 12:26 UTC (permalink / raw)
To: ceph-devel; +Cc: jlayton, Yan, Zheng
In-Reply-To: <20200220122630.63170-1-zyan@redhat.com>
When getting caps for read/write, update corresponding file's last
read/write. If a file hasn't been read/write for 'caps_wanted_delay_max'
seconds, ignore the file when calculating wanted caps.
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
---
fs/ceph/caps.c | 149 ++++++++++++++++++++++++-----------
fs/ceph/file.c | 23 +++---
fs/ceph/inode.c | 15 +++-
fs/ceph/ioctl.c | 4 +-
fs/ceph/super.h | 16 +++-
include/linux/ceph/ceph_fs.h | 1 +
6 files changed, 145 insertions(+), 63 deletions(-)
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 293920d013ff..ccdc47bd7cf0 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -971,18 +971,44 @@ int __ceph_caps_used(struct ceph_inode_info *ci)
return used;
}
+#define FMODE_WAIT_BIAS 1000
+
/*
* wanted, by virtue of open file modes
*/
int __ceph_caps_file_wanted(struct ceph_inode_info *ci)
{
+ struct ceph_mount_options *opt =
+ ceph_inode_to_client(&ci->vfs_inode)->mount_options;
+ unsigned long used_cutoff =
+ round_jiffies(jiffies - opt->caps_wanted_delay_max * HZ);
+ unsigned long idle_cutoff =
+ round_jiffies(jiffies - opt->caps_wanted_delay_min * HZ);
int i, bits = 0;
+
for (i = 0; i < CEPH_FILE_MODE_BITS; i++) {
- if (ci->i_nr_by_mode[i])
+ if (ci->i_file_by_mode[i].nr >= FMODE_WAIT_BIAS) {
+ /* there are cap waiters or lots of open files */
bits |= 1 << i;
+ } else if (ci->i_file_by_mode[i].nr > 0) {
+ if (i == CEPH_FILE_MODE_PIN ||
+ time_after(ci->i_file_by_mode[i].last_used,
+ used_cutoff))
+ bits |= 1 << i;
+ } else if ((ci->i_file_by_mode[i].last_used & 1)) {
+ if (time_after(ci->i_file_by_mode[i].last_used,
+ idle_cutoff)) {
+ bits |= 1 << i;
+ } else {
+ ci->i_file_by_mode[i].last_used &= ~1UL;
+ }
+ }
}
if (bits == 0)
return 0;
+ if (bits == 1 && !S_ISDIR(ci->vfs_inode.i_mode))
+ return 0;
+
return ceph_caps_for_mode(bits >> 1);
}
@@ -1021,14 +1047,6 @@ int __ceph_caps_mds_wanted(struct ceph_inode_info *ci, bool check)
return mds_wanted;
}
-/*
- * called under i_ceph_lock
- */
-static int __ceph_is_single_caps(struct ceph_inode_info *ci)
-{
- return rb_first(&ci->i_caps) == rb_last(&ci->i_caps);
-}
-
int ceph_is_any_caps(struct inode *inode)
{
struct ceph_inode_info *ci = ceph_inode(inode);
@@ -1856,10 +1874,6 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
if (ci->i_ceph_flags & CEPH_I_FLUSH)
flags |= CHECK_CAPS_FLUSH;
- if (!(flags & CHECK_CAPS_AUTHONLY) ||
- (ci->i_auth_cap && __ceph_is_single_caps(ci)))
- __cap_delay_cancel(mdsc, ci);
-
goto retry_locked;
retry:
spin_lock(&ci->i_ceph_lock);
@@ -2081,9 +2095,16 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
goto retry; /* retake i_ceph_lock and restart our cap scan. */
}
- /* Reschedule delayed caps release if we delayed anything */
- if (delayed)
- __cap_delay_requeue(mdsc, ci, false);
+ if (list_empty(&ci->i_cap_delay_list)) {
+ if (delayed) {
+ /* Reschedule delayed caps release if we delayed anything */
+ __cap_delay_requeue(mdsc, ci, false);
+ } else if ((file_wanted & ~CEPH_CAP_PIN) &&
+ !(used & (CEPH_CAP_FILE_RD | CEPH_CAP_ANY_FILE_WR))) {
+ /* periodically re-calculate caps wanted by open files */
+ __cap_delay_requeue(mdsc, ci, true);
+ }
+ }
spin_unlock(&ci->i_ceph_lock);
@@ -2549,8 +2570,9 @@ static void __take_cap_refs(struct ceph_inode_info *ci, int got,
* FIXME: how does a 0 return differ from -EAGAIN?
*/
enum {
- NON_BLOCKING = 1,
- CHECK_FILELOCK = 2,
+ /* first 8 bits are reserved for CEPH_FILE_MODE_FOO */
+ NON_BLOCKING = (1 << 8),
+ CHECK_FILELOCK = (1 << 9),
};
static int try_get_cap_refs(struct inode *inode, int need, int want,
@@ -2560,7 +2582,6 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
int ret = 0;
int have, implemented;
- int file_wanted;
bool snap_rwsem_locked = false;
dout("get_cap_refs %p need %s want %s\n", inode,
@@ -2576,15 +2597,6 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
goto out_unlock;
}
- /* make sure file is actually open */
- file_wanted = __ceph_caps_file_wanted(ci);
- if ((file_wanted & need) != need) {
- dout("try_get_cap_refs need %s file_wanted %s, EBADF\n",
- ceph_cap_string(need), ceph_cap_string(file_wanted));
- ret = -EBADF;
- goto out_unlock;
- }
-
/* finish pending truncate */
while (ci->i_truncate_pending) {
spin_unlock(&ci->i_ceph_lock);
@@ -2692,6 +2704,9 @@ static int try_get_cap_refs(struct inode *inode, int need, int want,
ceph_cap_string(have), ceph_cap_string(need));
}
out_unlock:
+
+ __ceph_touch_fmode(ci, mdsc, flags);
+
spin_unlock(&ci->i_ceph_lock);
if (snap_rwsem_locked)
up_read(&mdsc->snap_rwsem);
@@ -2729,10 +2744,22 @@ static void check_max_size(struct inode *inode, loff_t endoff)
ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL);
}
+static inline int get_used_file_mode(int need, int want)
+{
+ int fmode = 0;
+ if (need & CEPH_CAP_FILE_RD)
+ fmode |= CEPH_FILE_MODE_RD;
+ if (need & CEPH_CAP_FILE_WR)
+ fmode |= CEPH_FILE_MODE_WR;
+ if (want & CEPH_CAP_FILE_LAZYIO)
+ fmode |= CEPH_FILE_MODE_LAZY;
+ return fmode;
+}
+
int ceph_try_get_caps(struct inode *inode, int need, int want,
bool nonblock, int *got)
{
- int ret;
+ int ret, flags;
BUG_ON(need & ~CEPH_CAP_FILE_RD);
BUG_ON(want & ~(CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO|CEPH_CAP_FILE_SHARED));
@@ -2740,8 +2767,11 @@ int ceph_try_get_caps(struct inode *inode, int need, int want,
if (ret < 0)
return ret;
- ret = try_get_cap_refs(inode, need, want, 0,
- (nonblock ? NON_BLOCKING : 0), got);
+ flags = get_used_file_mode(need, want);
+ if (nonblock)
+ flags |= NON_BLOCKING;
+
+ ret = try_get_cap_refs(inode, need, want, 0, flags, got);
return ret == -EAGAIN ? 0 : ret;
}
@@ -2767,11 +2797,15 @@ int ceph_get_caps(struct file *filp, int need, int want,
fi->filp_gen != READ_ONCE(fsc->filp_gen))
return -EBADF;
+ flags = get_used_file_mode(need, want);
+
while (true) {
if (endoff > 0)
check_max_size(inode, endoff);
- flags = atomic_read(&fi->num_locks) ? CHECK_FILELOCK : 0;
+ flags &= CEPH_FILE_MODE_MASK;
+ if (atomic_read(&fi->num_locks))
+ flags |= CHECK_FILELOCK;
_got = 0;
ret = try_get_cap_refs(inode, need, want, endoff,
flags, &_got);
@@ -2791,6 +2825,8 @@ int ceph_get_caps(struct file *filp, int need, int want,
list_add(&cw.list, &mdsc->cap_wait_list);
spin_unlock(&mdsc->caps_list_lock);
+ /* make sure used fmode not timeout */
+ ceph_get_fmode(ci, flags, FMODE_WAIT_BIAS);
add_wait_queue(&ci->i_cap_wq, &wait);
flags |= NON_BLOCKING;
@@ -2804,6 +2840,7 @@ int ceph_get_caps(struct file *filp, int need, int want,
}
remove_wait_queue(&ci->i_cap_wq, &wait);
+ ceph_put_fmode(ci, flags, FMODE_WAIT_BIAS);
spin_lock(&mdsc->caps_list_lock);
list_del(&cw.list);
@@ -2823,7 +2860,7 @@ int ceph_get_caps(struct file *filp, int need, int want,
if (ret < 0) {
if (ret == -ESTALE) {
/* session was killed, try renew caps */
- ret = ceph_renew_caps(inode);
+ ret = ceph_renew_caps(inode, flags);
if (ret == 0)
continue;
}
@@ -4121,13 +4158,41 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc)
dout("flush_dirty_caps done\n");
}
+void __ceph_touch_fmode(struct ceph_inode_info *ci,
+ struct ceph_mds_client *mdsc, int fmode)
+{
+ int i;
+ int bits = (fmode << 1);
+ unsigned long now = jiffies | 1;
+ for (i = 1; i < CEPH_FILE_MODE_BITS; i++) {
+ if (bits & (1 << i))
+ ci->i_file_by_mode[i].last_used = now;
+ }
+
+ /* queue periodic check */
+ if (bits && list_empty(&ci->i_cap_delay_list))
+ __cap_delay_requeue(mdsc, ci, true);
+}
+
+void ceph_get_fmode(struct ceph_inode_info *ci, int fmode, int count)
+{
+ int i;
+ int bits = (fmode << 1) | 1;
+ spin_lock(&ci->i_ceph_lock);
+ for (i = 0; i < CEPH_FILE_MODE_BITS; i++) {
+ if (bits & (1 << i))
+ ci->i_file_by_mode[i].nr += count;
+ }
+ spin_unlock(&ci->i_ceph_lock);
+}
+
void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode)
{
int i;
int bits = (fmode << 1) | 1;
for (i = 0; i < CEPH_FILE_MODE_BITS; i++) {
if (bits & (1 << i))
- ci->i_nr_by_mode[i]++;
+ ci->i_file_by_mode[i].nr++;
}
}
@@ -4136,26 +4201,18 @@ void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode)
* we may need to release capabilities to the MDS (or schedule
* their delayed release).
*/
-void ceph_put_fmode(struct ceph_inode_info *ci, int fmode)
+void ceph_put_fmode(struct ceph_inode_info *ci, int fmode, int count)
{
- int i, last = 0;
+ int i;
int bits = (fmode << 1) | 1;
spin_lock(&ci->i_ceph_lock);
for (i = 0; i < CEPH_FILE_MODE_BITS; i++) {
if (bits & (1 << i)) {
- BUG_ON(ci->i_nr_by_mode[i] == 0);
- if (--ci->i_nr_by_mode[i] == 0)
- last++;
+ BUG_ON(ci->i_file_by_mode[i].nr < count);
+ ci->i_file_by_mode[i].nr -= count;
}
}
- dout("put_fmode %p fmode %d {%d,%d,%d,%d}\n",
- &ci->vfs_inode, fmode,
- ci->i_nr_by_mode[0], ci->i_nr_by_mode[1],
- ci->i_nr_by_mode[2], ci->i_nr_by_mode[3]);
spin_unlock(&ci->i_ceph_lock);
-
- if (last && ci->i_vino.snap == CEPH_NOSNAP)
- ceph_check_caps(ci, 0, NULL);
}
/*
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 7e0190b1f821..f28f420bad23 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -213,7 +213,7 @@ static int ceph_init_file_info(struct inode *inode, struct file *file,
struct ceph_dir_file_info *dfi =
kmem_cache_zalloc(ceph_dir_file_cachep, GFP_KERNEL);
if (!dfi) {
- ceph_put_fmode(ci, fmode); /* clean up */
+ ceph_put_fmode(ci, fmode, 1); /* clean up */
return -ENOMEM;
}
@@ -224,7 +224,7 @@ static int ceph_init_file_info(struct inode *inode, struct file *file,
} else {
fi = kmem_cache_zalloc(ceph_file_cachep, GFP_KERNEL);
if (!fi) {
- ceph_put_fmode(ci, fmode); /* clean up */
+ ceph_put_fmode(ci, fmode, 1); /* clean up */
return -ENOMEM;
}
@@ -263,7 +263,7 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
case S_IFLNK:
dout("init_file %p %p 0%o (symlink)\n", inode, file,
inode->i_mode);
- ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */
+ ceph_put_fmode(ceph_inode(inode), fmode, 1); /* clean up */
break;
default:
@@ -273,7 +273,7 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
* we need to drop the open ref now, since we don't
* have .release set to ceph_release.
*/
- ceph_put_fmode(ceph_inode(inode), fmode); /* clean up */
+ ceph_put_fmode(ceph_inode(inode), fmode, 1); /* clean up */
BUG_ON(inode->i_fop->release == ceph_release);
/* call the proper open fop */
@@ -285,14 +285,15 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
/*
* try renew caps after session gets killed.
*/
-int ceph_renew_caps(struct inode *inode)
+int ceph_renew_caps(struct inode *inode, int fmode)
{
- struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc;
+ struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
struct ceph_inode_info *ci = ceph_inode(inode);
struct ceph_mds_request *req;
int err, flags, wanted;
spin_lock(&ci->i_ceph_lock);
+ __ceph_touch_fmode(ci, mdsc, fmode);
wanted = __ceph_caps_file_wanted(ci);
if (__ceph_is_any_real_caps(ci) &&
(!(wanted & CEPH_CAP_ANY_WR) || ci->i_auth_cap)) {
@@ -405,6 +406,7 @@ int ceph_open(struct inode *inode, struct file *file)
} else if (ceph_snap(inode) != CEPH_NOSNAP &&
(ci->i_snap_caps & wanted) == wanted) {
__ceph_get_fmode(ci, fmode);
+ __ceph_touch_fmode(ci, mdsc, fmode);
spin_unlock(&ci->i_ceph_lock);
return ceph_init_file(inode, file, fmode);
}
@@ -525,7 +527,7 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
}
out_req:
if (!req->r_err && req->r_target_inode)
- ceph_put_fmode(ceph_inode(req->r_target_inode), req->r_fmode);
+ ceph_put_fmode(ceph_inode(req->r_target_inode), req->r_fmode, 1);
ceph_mdsc_put_request(req);
out_ctx:
ceph_release_acl_sec_ctx(&as_ctx);
@@ -542,7 +544,7 @@ int ceph_release(struct inode *inode, struct file *file)
dout("release inode %p dir file %p\n", inode, file);
WARN_ON(!list_empty(&dfi->file_info.rw_contexts));
- ceph_put_fmode(ci, dfi->file_info.fmode);
+ ceph_put_fmode(ci, dfi->file_info.fmode, 1);
if (dfi->last_readdir)
ceph_mdsc_put_request(dfi->last_readdir);
@@ -554,7 +556,8 @@ int ceph_release(struct inode *inode, struct file *file)
dout("release inode %p regular file %p\n", inode, file);
WARN_ON(!list_empty(&fi->rw_contexts));
- ceph_put_fmode(ci, fi->fmode);
+ ceph_put_fmode(ci, fi->fmode, 1);
+
kmem_cache_free(ceph_file_cachep, fi);
}
@@ -1560,7 +1563,7 @@ static ssize_t ceph_write_iter(struct kiocb *iocb, struct iov_iter *from)
if (dirty)
__mark_inode_dirty(inode, dirty);
if (ceph_quota_is_max_bytes_approaching(inode, iocb->ki_pos))
- ceph_check_caps(ci, CHECK_CAPS_NODELAY, NULL);
+ ceph_check_caps(ci, CHECK_CAPS_AUTHONLY, NULL);
}
dout("aio_write %p %llx.%llx %llu~%u dropping cap refs on %s\n",
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 094b8fc37787..b279bd8e168e 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -478,8 +478,10 @@ struct inode *ceph_alloc_inode(struct super_block *sb)
ci->i_head_snapc = NULL;
ci->i_snap_caps = 0;
- for (i = 0; i < CEPH_FILE_MODE_BITS; i++)
- ci->i_nr_by_mode[i] = 0;
+ for (i = 0; i < CEPH_FILE_MODE_BITS; i++) {
+ ci->i_file_by_mode[i].nr = 0;
+ ci->i_file_by_mode[i].last_used = 0;
+ }
mutex_init(&ci->i_truncate_mutex);
ci->i_truncate_seq = 0;
@@ -637,7 +639,7 @@ int ceph_fill_file_size(struct inode *inode, int issued,
if ((issued & (CEPH_CAP_FILE_CACHE|
CEPH_CAP_FILE_BUFFER)) ||
mapping_mapped(inode->i_mapping) ||
- __ceph_caps_file_wanted(ci)) {
+ __ceph_is_file_opened(ci)) {
ci->i_truncate_pending++;
queue_trunc = 1;
}
@@ -1010,6 +1012,13 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
fill_inline = true;
}
+ if (cap_fmode >= 0) {
+ if (!info_caps)
+ pr_warn("mds issued no caps on %llx.%llx\n",
+ ceph_vinop(inode));
+ __ceph_touch_fmode(ci, mdsc, cap_fmode);
+ }
+
spin_unlock(&ci->i_ceph_lock);
if (fill_inline)
diff --git a/fs/ceph/ioctl.c b/fs/ceph/ioctl.c
index c90f03beb15d..da0ee54ae5bc 100644
--- a/fs/ceph/ioctl.c
+++ b/fs/ceph/ioctl.c
@@ -243,11 +243,13 @@ static long ceph_ioctl_lazyio(struct file *file)
struct ceph_file_info *fi = file->private_data;
struct inode *inode = file_inode(file);
struct ceph_inode_info *ci = ceph_inode(inode);
+ struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
if ((fi->fmode & CEPH_FILE_MODE_LAZY) == 0) {
spin_lock(&ci->i_ceph_lock);
fi->fmode |= CEPH_FILE_MODE_LAZY;
- ci->i_nr_by_mode[ffs(CEPH_FILE_MODE_LAZY)]++;
+ ci->i_file_by_mode[ffs(CEPH_FILE_MODE_LAZY)].nr++;
+ __ceph_touch_fmode(ci, mdsc, CEPH_FILE_MODE_LAZY);
spin_unlock(&ci->i_ceph_lock);
dout("ioctl_layzio: file %p marked lazy\n", file);
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index d370f89df358..029823643b8b 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -361,7 +361,10 @@ struct ceph_inode_info {
dirty|flushing caps */
unsigned i_snap_caps; /* cap bits for snapped files */
- int i_nr_by_mode[CEPH_FILE_MODE_BITS]; /* open file counts */
+ struct {
+ int nr;
+ unsigned long last_used;
+ } i_file_by_mode[CEPH_FILE_MODE_BITS]; /* open file counts */
struct mutex i_truncate_mutex;
u32 i_truncate_seq; /* last truncate to smaller size */
@@ -673,6 +676,10 @@ extern int __ceph_caps_revoking_other(struct ceph_inode_info *ci,
extern int ceph_caps_revoking(struct ceph_inode_info *ci, int mask);
extern int __ceph_caps_used(struct ceph_inode_info *ci);
+static inline bool __ceph_is_file_opened(struct ceph_inode_info *ci)
+{
+ return ci->i_file_by_mode[0].nr;
+}
extern int __ceph_caps_file_wanted(struct ceph_inode_info *ci);
extern int __ceph_caps_wanted(struct ceph_inode_info *ci);
@@ -1074,7 +1081,10 @@ extern int ceph_try_get_caps(struct inode *inode,
/* for counting open files by mode */
extern void __ceph_get_fmode(struct ceph_inode_info *ci, int mode);
-extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode);
+extern void ceph_get_fmode(struct ceph_inode_info *ci, int mode, int count);
+extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode, int count);
+extern void __ceph_touch_fmode(struct ceph_inode_info *ci,
+ struct ceph_mds_client *mdsc, int fmode);
/* addr.c */
extern const struct address_space_operations ceph_aops;
@@ -1086,7 +1096,7 @@ extern void ceph_pool_perm_destroy(struct ceph_mds_client* mdsc);
/* file.c */
extern const struct file_operations ceph_file_fops;
-extern int ceph_renew_caps(struct inode *inode);
+extern int ceph_renew_caps(struct inode *inode, int fmode);
extern int ceph_open(struct inode *inode, struct file *file);
extern int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
struct file *file, unsigned flags, umode_t mode);
diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h
index cb21c5cf12c3..8017130a08a1 100644
--- a/include/linux/ceph/ceph_fs.h
+++ b/include/linux/ceph/ceph_fs.h
@@ -564,6 +564,7 @@ struct ceph_filelock {
#define CEPH_FILE_MODE_RDWR 3 /* RD | WR */
#define CEPH_FILE_MODE_LAZY 4 /* lazy io */
#define CEPH_FILE_MODE_BITS 4
+#define CEPH_FILE_MODE_MASK ((1 << CEPH_FILE_MODE_BITS) - 1)
int ceph_flags_to_mode(int flags);
--
2.21.1
^ permalink raw reply related
* [PATCH 3/4] ceph: simplify calling of ceph_get_fmode()
From: Yan, Zheng @ 2020-02-20 12:26 UTC (permalink / raw)
To: ceph-devel; +Cc: jlayton, Yan, Zheng
In-Reply-To: <20200220122630.63170-1-zyan@redhat.com>
Call ceph_get_fmode() when initializing file. Because fill_inode()
already calls ceph_touch_fmode() for open file request. It affects
__ceph_caps_file_wanted() for 'caps_wanted_delay_min' seconds, enough
for ceph_get_fmode() to get called by ceph_init_file_info().
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
---
fs/ceph/caps.c | 26 +++-----------------------
fs/ceph/file.c | 21 +++++----------------
fs/ceph/inode.c | 8 +-------
fs/ceph/super.h | 3 +--
4 files changed, 10 insertions(+), 48 deletions(-)
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index ccdc47bd7cf0..2f4ff7e9508e 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -606,7 +606,7 @@ static void __check_cap_issue(struct ceph_inode_info *ci, struct ceph_cap *cap,
*/
void ceph_add_cap(struct inode *inode,
struct ceph_mds_session *session, u64 cap_id,
- int fmode, unsigned issued, unsigned wanted,
+ unsigned issued, unsigned wanted,
unsigned seq, unsigned mseq, u64 realmino, int flags,
struct ceph_cap **new_cap)
{
@@ -622,13 +622,6 @@ void ceph_add_cap(struct inode *inode,
dout("add_cap %p mds%d cap %llx %s seq %d\n", inode,
session->s_mds, cap_id, ceph_cap_string(issued), seq);
- /*
- * If we are opening the file, include file mode wanted bits
- * in wanted.
- */
- if (fmode >= 0)
- wanted |= ceph_caps_for_mode(fmode);
-
spin_lock(&session->s_gen_ttl_lock);
gen = session->s_cap_gen;
spin_unlock(&session->s_gen_ttl_lock);
@@ -753,9 +746,6 @@ void ceph_add_cap(struct inode *inode,
cap->issue_seq = seq;
cap->mseq = mseq;
cap->cap_gen = gen;
-
- if (fmode >= 0)
- __ceph_get_fmode(ci, fmode);
}
/*
@@ -3726,7 +3716,7 @@ static void handle_cap_export(struct inode *inode, struct ceph_mds_caps *ex,
/* add placeholder for the export tagert */
int flag = (cap == ci->i_auth_cap) ? CEPH_CAP_FLAG_AUTH : 0;
tcap = new_cap;
- ceph_add_cap(inode, tsession, t_cap_id, -1, issued, 0,
+ ceph_add_cap(inode, tsession, t_cap_id, issued, 0,
t_seq - 1, t_mseq, (u64)-1, flag, &new_cap);
if (!list_empty(&ci->i_cap_flush_list) &&
@@ -3831,7 +3821,7 @@ static void handle_cap_import(struct ceph_mds_client *mdsc,
__ceph_caps_issued(ci, &issued);
issued |= __ceph_caps_dirty(ci);
- ceph_add_cap(inode, session, cap_id, -1, caps, wanted, seq, mseq,
+ ceph_add_cap(inode, session, cap_id, caps, wanted, seq, mseq,
realmino, CEPH_CAP_FLAG_AUTH, &new_cap);
ocap = peer >= 0 ? __get_cap_for_mds(ci, peer) : NULL;
@@ -4186,16 +4176,6 @@ void ceph_get_fmode(struct ceph_inode_info *ci, int fmode, int count)
spin_unlock(&ci->i_ceph_lock);
}
-void __ceph_get_fmode(struct ceph_inode_info *ci, int fmode)
-{
- int i;
- int bits = (fmode << 1) | 1;
- for (i = 0; i < CEPH_FILE_MODE_BITS; i++) {
- if (bits & (1 << i))
- ci->i_file_by_mode[i].nr++;
- }
-}
-
/*
* Drop open file reference. If we were the last open file,
* we may need to release capabilities to the MDS (or schedule
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index f28f420bad23..60a2dfa02ba2 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -212,10 +212,8 @@ static int ceph_init_file_info(struct inode *inode, struct file *file,
if (isdir) {
struct ceph_dir_file_info *dfi =
kmem_cache_zalloc(ceph_dir_file_cachep, GFP_KERNEL);
- if (!dfi) {
- ceph_put_fmode(ci, fmode, 1); /* clean up */
+ if (!dfi)
return -ENOMEM;
- }
file->private_data = dfi;
fi = &dfi->file_info;
@@ -223,15 +221,15 @@ static int ceph_init_file_info(struct inode *inode, struct file *file,
dfi->readdir_cache_idx = -1;
} else {
fi = kmem_cache_zalloc(ceph_file_cachep, GFP_KERNEL);
- if (!fi) {
- ceph_put_fmode(ci, fmode, 1); /* clean up */
+ if (!fi)
return -ENOMEM;
- }
file->private_data = fi;
}
+ ceph_get_fmode(ci, fmode, 1);
fi->fmode = fmode;
+
spin_lock_init(&fi->rw_contexts_lock);
INIT_LIST_HEAD(&fi->rw_contexts);
fi->meta_err = errseq_sample(&ci->i_meta_err);
@@ -263,7 +261,6 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
case S_IFLNK:
dout("init_file %p %p 0%o (symlink)\n", inode, file,
inode->i_mode);
- ceph_put_fmode(ceph_inode(inode), fmode, 1); /* clean up */
break;
default:
@@ -273,7 +270,6 @@ static int ceph_init_file(struct inode *inode, struct file *file, int fmode)
* we need to drop the open ref now, since we don't
* have .release set to ceph_release.
*/
- ceph_put_fmode(ceph_inode(inode), fmode, 1); /* clean up */
BUG_ON(inode->i_fop->release == ceph_release);
/* call the proper open fop */
@@ -327,7 +323,6 @@ int ceph_renew_caps(struct inode *inode, int fmode)
req->r_inode = inode;
ihold(inode);
req->r_num_caps = 1;
- req->r_fmode = -1;
err = ceph_mdsc_do_request(mdsc, NULL, req);
ceph_mdsc_put_request(req);
@@ -373,9 +368,6 @@ int ceph_open(struct inode *inode, struct file *file)
/* trivially open snapdir */
if (ceph_snap(inode) == CEPH_SNAPDIR) {
- spin_lock(&ci->i_ceph_lock);
- __ceph_get_fmode(ci, fmode);
- spin_unlock(&ci->i_ceph_lock);
return ceph_init_file(inode, file, fmode);
}
@@ -393,7 +385,7 @@ int ceph_open(struct inode *inode, struct file *file)
dout("open %p fmode %d want %s issued %s using existing\n",
inode, fmode, ceph_cap_string(wanted),
ceph_cap_string(issued));
- __ceph_get_fmode(ci, fmode);
+ __ceph_touch_fmode(ci, mdsc, fmode);
spin_unlock(&ci->i_ceph_lock);
/* adjust wanted? */
@@ -405,7 +397,6 @@ int ceph_open(struct inode *inode, struct file *file)
return ceph_init_file(inode, file, fmode);
} else if (ceph_snap(inode) != CEPH_NOSNAP &&
(ci->i_snap_caps & wanted) == wanted) {
- __ceph_get_fmode(ci, fmode);
__ceph_touch_fmode(ci, mdsc, fmode);
spin_unlock(&ci->i_ceph_lock);
return ceph_init_file(inode, file, fmode);
@@ -526,8 +517,6 @@ int ceph_atomic_open(struct inode *dir, struct dentry *dentry,
err = finish_open(file, dentry, ceph_open);
}
out_req:
- if (!req->r_err && req->r_target_inode)
- ceph_put_fmode(ceph_inode(req->r_target_inode), req->r_fmode, 1);
ceph_mdsc_put_request(req);
out_ctx:
ceph_release_acl_sec_ctx(&as_ctx);
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index b279bd8e168e..bb73b0c8c4d9 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -969,7 +969,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
if (ceph_snap(inode) == CEPH_NOSNAP) {
ceph_add_cap(inode, session,
le64_to_cpu(info->cap.cap_id),
- cap_fmode, info_caps,
+ info_caps,
le32_to_cpu(info->cap.wanted),
le32_to_cpu(info->cap.seq),
le32_to_cpu(info->cap.mseq),
@@ -994,13 +994,7 @@ static int fill_inode(struct inode *inode, struct page *locked_page,
dout(" %p got snap_caps %s\n", inode,
ceph_cap_string(info_caps));
ci->i_snap_caps |= info_caps;
- if (cap_fmode >= 0)
- __ceph_get_fmode(ci, cap_fmode);
}
- } else if (cap_fmode >= 0) {
- pr_warn("mds issued no caps on %llx.%llx\n",
- ceph_vinop(inode));
- __ceph_get_fmode(ci, cap_fmode);
}
if (iinfo->inline_version > 0 &&
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 029823643b8b..1ea76466efcb 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -1038,7 +1038,7 @@ extern struct ceph_cap *ceph_get_cap(struct ceph_mds_client *mdsc,
struct ceph_cap_reservation *ctx);
extern void ceph_add_cap(struct inode *inode,
struct ceph_mds_session *session, u64 cap_id,
- int fmode, unsigned issued, unsigned wanted,
+ unsigned issued, unsigned wanted,
unsigned cap, unsigned seq, u64 realmino, int flags,
struct ceph_cap **new_cap);
extern void __ceph_remove_cap(struct ceph_cap *cap, bool queue_release);
@@ -1080,7 +1080,6 @@ extern int ceph_try_get_caps(struct inode *inode,
int need, int want, bool nonblock, int *got);
/* for counting open files by mode */
-extern void __ceph_get_fmode(struct ceph_inode_info *ci, int mode);
extern void ceph_get_fmode(struct ceph_inode_info *ci, int mode, int count);
extern void ceph_put_fmode(struct ceph_inode_info *ci, int mode, int count);
extern void __ceph_touch_fmode(struct ceph_inode_info *ci,
--
2.21.1
^ permalink raw reply related
* [PATCH 4/4] ceph: remove delay check logic from ceph_check_caps()
From: Yan, Zheng @ 2020-02-20 12:26 UTC (permalink / raw)
To: ceph-devel; +Cc: jlayton, Yan, Zheng
In-Reply-To: <20200220122630.63170-1-zyan@redhat.com>
__ceph_caps_file_wanted() already checks 'caps_wanted_delay_min' and
'caps_wanted_delay_max'. There is no need to duplicte the logic in
ceph_check_caps() and __send_cap()
Signed-off-by: "Yan, Zheng" <zyan@redhat.com>
---
fs/ceph/caps.c | 146 ++++++++++++------------------------------------
fs/ceph/file.c | 5 +-
fs/ceph/inode.c | 1 -
fs/ceph/super.h | 8 +--
4 files changed, 41 insertions(+), 119 deletions(-)
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 2f4ff7e9508e..39bf41d0fbdb 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -490,13 +490,10 @@ static void __cap_set_timeouts(struct ceph_mds_client *mdsc,
struct ceph_inode_info *ci)
{
struct ceph_mount_options *opt = mdsc->fsc->mount_options;
-
- ci->i_hold_caps_min = round_jiffies(jiffies +
- opt->caps_wanted_delay_min * HZ);
ci->i_hold_caps_max = round_jiffies(jiffies +
opt->caps_wanted_delay_max * HZ);
- dout("__cap_set_timeouts %p min %lu max %lu\n", &ci->vfs_inode,
- ci->i_hold_caps_min - jiffies, ci->i_hold_caps_max - jiffies);
+ dout("__cap_set_timeouts %p %lu\n", &ci->vfs_inode,
+ ci->i_hold_caps_max - jiffies);
}
/*
@@ -508,8 +505,7 @@ static void __cap_set_timeouts(struct ceph_mds_client *mdsc,
* -> we take mdsc->cap_delay_lock
*/
static void __cap_delay_requeue(struct ceph_mds_client *mdsc,
- struct ceph_inode_info *ci,
- bool set_timeout)
+ struct ceph_inode_info *ci)
{
dout("__cap_delay_requeue %p flags %d at %lu\n", &ci->vfs_inode,
ci->i_ceph_flags, ci->i_hold_caps_max);
@@ -520,8 +516,7 @@ static void __cap_delay_requeue(struct ceph_mds_client *mdsc,
goto no_change;
list_del_init(&ci->i_cap_delay_list);
}
- if (set_timeout)
- __cap_set_timeouts(mdsc, ci);
+ __cap_set_timeouts(mdsc, ci);
list_add_tail(&ci->i_cap_delay_list, &mdsc->cap_delay_list);
no_change:
spin_unlock(&mdsc->cap_delay_lock);
@@ -719,7 +714,7 @@ void ceph_add_cap(struct inode *inode,
dout(" issued %s, mds wanted %s, actual %s, queueing\n",
ceph_cap_string(issued), ceph_cap_string(wanted),
ceph_cap_string(actual_wanted));
- __cap_delay_requeue(mdsc, ci, true);
+ __cap_delay_requeue(mdsc, ci);
}
if (flags & CEPH_CAP_FLAG_AUTH) {
@@ -1299,7 +1294,6 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
struct cap_msg_args arg;
int held, revoking;
int wake = 0;
- int delayed = 0;
int ret;
held = cap->issued | cap->implemented;
@@ -1312,28 +1306,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
ceph_cap_string(revoking));
BUG_ON((retain & CEPH_CAP_PIN) == 0);
- arg.session = cap->session;
-
- /* don't release wanted unless we've waited a bit. */
- if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0 &&
- time_before(jiffies, ci->i_hold_caps_min)) {
- dout(" delaying issued %s -> %s, wanted %s -> %s on send\n",
- ceph_cap_string(cap->issued),
- ceph_cap_string(cap->issued & retain),
- ceph_cap_string(cap->mds_wanted),
- ceph_cap_string(want));
- want |= cap->mds_wanted;
- retain |= cap->issued;
- delayed = 1;
- }
- ci->i_ceph_flags &= ~(CEPH_I_NODELAY | CEPH_I_FLUSH);
- if (want & ~cap->mds_wanted) {
- /* user space may open/close single file frequently.
- * This avoids droping mds_wanted immediately after
- * requesting new mds_wanted.
- */
- __cap_set_timeouts(mdsc, ci);
- }
+ ci->i_ceph_flags &= ~CEPH_I_FLUSH;
cap->issued &= retain; /* drop bits we don't want */
if (cap->implemented & ~cap->issued) {
@@ -1348,6 +1321,7 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
cap->implemented &= cap->issued | used;
cap->mds_wanted = want;
+ arg.session = cap->session;
arg.ino = ceph_vino(inode).ino;
arg.cid = cap->cap_id;
arg.follows = flushing ? ci->i_head_snapc->seq : 0;
@@ -1408,14 +1382,19 @@ static int __send_cap(struct ceph_mds_client *mdsc, struct ceph_cap *cap,
ret = send_cap_msg(&arg);
if (ret < 0) {
- dout("error sending cap msg, must requeue %p\n", inode);
- delayed = 1;
+ pr_err("error sending cap msg, ino (%llx.%llx) "
+ "flushing %s tid %llu, requeue\n",
+ ceph_vinop(inode), ceph_cap_string(flushing),
+ flush_tid);
+ spin_lock(&ci->i_ceph_lock);
+ __cap_delay_requeue(mdsc, ci);
+ spin_unlock(&ci->i_ceph_lock);
}
if (wake)
wake_up_all(&ci->i_cap_wq);
- return delayed;
+ return ret;
}
static inline int __send_flush_snap(struct inode *inode,
@@ -1679,7 +1658,7 @@ int __ceph_mark_dirty_caps(struct ceph_inode_info *ci, int mask,
if (((was | ci->i_flushing_caps) & CEPH_CAP_FILE_BUFFER) &&
(mask & CEPH_CAP_FILE_BUFFER))
dirty |= I_DIRTY_DATASYNC;
- __cap_delay_requeue(mdsc, ci, true);
+ __cap_delay_requeue(mdsc, ci);
return dirty;
}
@@ -1830,8 +1809,6 @@ bool __ceph_should_report_size(struct ceph_inode_info *ci)
* versus held caps. Release, flush, ack revoked caps to mds as
* appropriate.
*
- * CHECK_CAPS_NODELAY - caller is delayed work and we should not delay
- * cap release further.
* CHECK_CAPS_AUTHONLY - we should only check the auth cap
* CHECK_CAPS_FLUSH - we should flush any dirty caps immediately, without
* further delay.
@@ -1850,17 +1827,10 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
int mds = -1; /* keep track of how far we've gone through i_caps list
to avoid an infinite loop on retry */
struct rb_node *p;
- int delayed = 0, sent = 0;
- bool no_delay = flags & CHECK_CAPS_NODELAY;
bool queue_invalidate = false;
bool tried_invalidate = false;
- /* if we are unmounting, flush any unused caps immediately. */
- if (mdsc->stopping)
- no_delay = true;
-
spin_lock(&ci->i_ceph_lock);
-
if (ci->i_ceph_flags & CEPH_I_FLUSH)
flags |= CHECK_CAPS_FLUSH;
@@ -1906,14 +1876,13 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
}
dout("check_caps %p file_want %s used %s dirty %s flushing %s"
- " issued %s revoking %s retain %s %s%s%s\n", inode,
+ " issued %s revoking %s retain %s %s%s\n", inode,
ceph_cap_string(file_wanted),
ceph_cap_string(used), ceph_cap_string(ci->i_dirty_caps),
ceph_cap_string(ci->i_flushing_caps),
ceph_cap_string(issued), ceph_cap_string(revoking),
ceph_cap_string(retain),
(flags & CHECK_CAPS_AUTHONLY) ? " AUTHONLY" : "",
- (flags & CHECK_CAPS_NODELAY) ? " NODELAY" : "",
(flags & CHECK_CAPS_FLUSH) ? " FLUSH" : "");
/*
@@ -1921,7 +1890,7 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
* have cached pages, but don't want them, then try to invalidate.
* If we fail, it's because pages are locked.... try again later.
*/
- if ((!no_delay || mdsc->stopping) &&
+ if ((!(flags & CHECK_CAPS_NOINVAL) || mdsc->stopping) &&
S_ISREG(inode->i_mode) &&
!(ci->i_wb_ref || ci->i_wrbuffer_ref) && /* no dirty pages... */
inode->i_data.nrpages && /* have cached pages */
@@ -2001,21 +1970,6 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
if ((cap->issued & ~retain) == 0)
continue; /* nope, all good */
- if (no_delay)
- goto ack;
-
- /* delay? */
- if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0 &&
- time_before(jiffies, ci->i_hold_caps_max)) {
- dout(" delaying issued %s -> %s, wanted %s -> %s\n",
- ceph_cap_string(cap->issued),
- ceph_cap_string(cap->issued & retain),
- ceph_cap_string(cap->mds_wanted),
- ceph_cap_string(want));
- delayed++;
- continue;
- }
-
ack:
if (session && session != cap->session) {
dout("oops, wrong session %p mutex\n", session);
@@ -2076,24 +2030,18 @@ void ceph_check_caps(struct ceph_inode_info *ci, int flags,
}
mds = cap->mds; /* remember mds, so we don't repeat */
- sent++;
/* __send_cap drops i_ceph_lock */
- delayed += __send_cap(mdsc, cap, CEPH_CAP_OP_UPDATE, 0,
- cap_used, want, retain, flushing,
- flush_tid, oldest_flush_tid);
+ __send_cap(mdsc, cap, CEPH_CAP_OP_UPDATE, 0, cap_used, want,
+ retain, flushing, flush_tid, oldest_flush_tid);
goto retry; /* retake i_ceph_lock and restart our cap scan. */
}
- if (list_empty(&ci->i_cap_delay_list)) {
- if (delayed) {
- /* Reschedule delayed caps release if we delayed anything */
- __cap_delay_requeue(mdsc, ci, false);
- } else if ((file_wanted & ~CEPH_CAP_PIN) &&
- !(used & (CEPH_CAP_FILE_RD | CEPH_CAP_ANY_FILE_WR))) {
- /* periodically re-calculate caps wanted by open files */
- __cap_delay_requeue(mdsc, ci, true);
- }
+ /* periodically re-calculate caps wanted by open files */
+ if (list_empty(&ci->i_cap_delay_list) &&
+ (file_wanted & ~CEPH_CAP_PIN) &&
+ !(used & (CEPH_CAP_FILE_RD | CEPH_CAP_ANY_FILE_WR))) {
+ __cap_delay_requeue(mdsc, ci);
}
spin_unlock(&ci->i_ceph_lock);
@@ -2123,7 +2071,6 @@ static int try_flush_caps(struct inode *inode, u64 *ptid)
retry_locked:
if (ci->i_dirty_caps && ci->i_auth_cap) {
struct ceph_cap *cap = ci->i_auth_cap;
- int delayed;
if (session != cap->session) {
spin_unlock(&ci->i_ceph_lock);
@@ -2152,18 +2099,10 @@ static int try_flush_caps(struct inode *inode, u64 *ptid)
&oldest_flush_tid);
/* __send_cap drops i_ceph_lock */
- delayed = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH,
- CEPH_CLIENT_CAPS_SYNC,
- __ceph_caps_used(ci),
- __ceph_caps_wanted(ci),
- (cap->issued | cap->implemented),
- flushing, flush_tid, oldest_flush_tid);
-
- if (delayed) {
- spin_lock(&ci->i_ceph_lock);
- __cap_delay_requeue(mdsc, ci, true);
- spin_unlock(&ci->i_ceph_lock);
- }
+ __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH, CEPH_CLIENT_CAPS_SYNC,
+ __ceph_caps_used(ci), __ceph_caps_wanted(ci),
+ (cap->issued | cap->implemented),
+ flushing, flush_tid, oldest_flush_tid);
} else {
if (!list_empty(&ci->i_cap_flush_list)) {
struct ceph_cap_flush *cf =
@@ -2363,22 +2302,13 @@ static void __kick_flushing_caps(struct ceph_mds_client *mdsc,
if (cf->caps) {
dout("kick_flushing_caps %p cap %p tid %llu %s\n",
inode, cap, cf->tid, ceph_cap_string(cf->caps));
- ci->i_ceph_flags |= CEPH_I_NODELAY;
-
- ret = __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH,
+ __send_cap(mdsc, cap, CEPH_CAP_OP_FLUSH,
(cf->tid < last_snap_flush ?
CEPH_CLIENT_CAPS_PENDING_CAPSNAP : 0),
__ceph_caps_used(ci),
__ceph_caps_wanted(ci),
(cap->issued | cap->implemented),
cf->caps, cf->tid, oldest_flush_tid);
- if (ret) {
- pr_err("kick_flushing_caps: error sending "
- "cap flush, ino (%llx.%llx) "
- "tid %llu flushing %s\n",
- ceph_vinop(inode), cf->tid,
- ceph_cap_string(cf->caps));
- }
} else {
struct ceph_cap_snap *capsnap =
container_of(cf, struct ceph_cap_snap,
@@ -2999,7 +2929,7 @@ void ceph_put_cap_refs(struct ceph_inode_info *ci, int had)
dout("put_cap_refs %p had %s%s%s\n", inode, ceph_cap_string(had),
last ? " last" : "", put ? " put" : "");
- if (last && !flushsnaps)
+ if (last)
ceph_check_caps(ci, 0, NULL);
else if (flushsnaps)
ceph_flush_snaps(ci, NULL);
@@ -3417,10 +3347,10 @@ static void handle_cap_grant(struct inode *inode,
wake_up_all(&ci->i_cap_wq);
if (check_caps == 1)
- ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_AUTHONLY,
+ ceph_check_caps(ci, CHECK_CAPS_AUTHONLY | CHECK_CAPS_NOINVAL,
session);
else if (check_caps == 2)
- ceph_check_caps(ci, CHECK_CAPS_NODELAY, session);
+ ceph_check_caps(ci, CHECK_CAPS_NOINVAL, session);
else
mutex_unlock(&session->s_mutex);
}
@@ -4095,7 +4025,6 @@ void ceph_check_delayed_caps(struct ceph_mds_client *mdsc)
{
struct inode *inode;
struct ceph_inode_info *ci;
- int flags = CHECK_CAPS_NODELAY;
dout("check_delayed_caps\n");
while (1) {
@@ -4115,7 +4044,7 @@ void ceph_check_delayed_caps(struct ceph_mds_client *mdsc)
if (inode) {
dout("check_delayed_caps on %p\n", inode);
- ceph_check_caps(ci, flags, NULL);
+ ceph_check_caps(ci, 0, NULL);
/* avoid calling iput_final() in tick thread */
ceph_async_iput(inode);
}
@@ -4140,7 +4069,7 @@ void ceph_flush_dirty_caps(struct ceph_mds_client *mdsc)
ihold(inode);
dout("flush_dirty_caps %p\n", inode);
spin_unlock(&mdsc->cap_dirty_lock);
- ceph_check_caps(ci, CHECK_CAPS_NODELAY|CHECK_CAPS_FLUSH, NULL);
+ ceph_check_caps(ci, CHECK_CAPS_FLUSH, NULL);
iput(inode);
spin_lock(&mdsc->cap_dirty_lock);
}
@@ -4161,7 +4090,7 @@ void __ceph_touch_fmode(struct ceph_inode_info *ci,
/* queue periodic check */
if (bits && list_empty(&ci->i_cap_delay_list))
- __cap_delay_requeue(mdsc, ci, true);
+ __cap_delay_requeue(mdsc, ci);
}
void ceph_get_fmode(struct ceph_inode_info *ci, int fmode, int count)
@@ -4210,7 +4139,6 @@ int ceph_drop_caps_for_unlink(struct inode *inode)
if (inode->i_nlink == 1) {
drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN);
- ci->i_ceph_flags |= CEPH_I_NODELAY;
if (__ceph_caps_dirty(ci)) {
struct ceph_mds_client *mdsc =
ceph_inode_to_client(inode)->mdsc;
@@ -4266,8 +4194,6 @@ int ceph_encode_inode_release(void **p, struct inode *inode,
if (force || (cap->issued & drop)) {
if (cap->issued & drop) {
int wanted = __ceph_caps_wanted(ci);
- if ((ci->i_ceph_flags & CEPH_I_NODELAY) == 0)
- wanted |= cap->mds_wanted;
dout("encode_inode_release %p cap %p "
"%s -> %s, wanted %s -> %s\n", inode, cap,
ceph_cap_string(cap->issued),
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 60a2dfa02ba2..ed70bb448568 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -2129,12 +2129,11 @@ static ssize_t __ceph_copy_file_range(struct file *src_file, loff_t src_off,
if (endoff > size) {
int caps_flags = 0;
-
/* Let the MDS know about dst file size change */
- if (ceph_quota_is_max_bytes_approaching(dst_inode, endoff))
- caps_flags |= CHECK_CAPS_NODELAY;
if (ceph_inode_set_size(dst_inode, endoff))
caps_flags |= CHECK_CAPS_AUTHONLY;
+ if (ceph_quota_is_max_bytes_approaching(dst_inode, endoff))
+ caps_flags |= CHECK_CAPS_AUTHONLY;
if (caps_flags)
ceph_check_caps(dst_ci, caps_flags, NULL);
}
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index bb73b0c8c4d9..5c8d3b01bc5d 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -471,7 +471,6 @@ struct inode *ceph_alloc_inode(struct super_block *sb)
ci->i_prealloc_cap_flush = NULL;
INIT_LIST_HEAD(&ci->i_cap_flush_list);
init_waitqueue_head(&ci->i_cap_wq);
- ci->i_hold_caps_min = 0;
ci->i_hold_caps_max = 0;
INIT_LIST_HEAD(&ci->i_cap_delay_list);
INIT_LIST_HEAD(&ci->i_cap_snaps);
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 1ea76466efcb..9ddaaeefc6f0 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -170,9 +170,9 @@ struct ceph_cap {
struct list_head caps_item;
};
-#define CHECK_CAPS_NODELAY 1 /* do not delay any further */
-#define CHECK_CAPS_AUTHONLY 2 /* only check auth cap */
-#define CHECK_CAPS_FLUSH 4 /* flush any dirty caps */
+#define CHECK_CAPS_AUTHONLY 1 /* only check auth cap */
+#define CHECK_CAPS_FLUSH 2 /* flush any dirty caps */
+#define CHECK_CAPS_NOINVAL 4 /* don't invalidate pagecache */
struct ceph_cap_flush {
u64 tid;
@@ -352,7 +352,6 @@ struct ceph_inode_info {
struct ceph_cap_flush *i_prealloc_cap_flush;
struct list_head i_cap_flush_list;
wait_queue_head_t i_cap_wq; /* threads waiting on a capability */
- unsigned long i_hold_caps_min; /* jiffies */
unsigned long i_hold_caps_max; /* jiffies */
struct list_head i_cap_delay_list; /* for delayed cap release to mds */
struct ceph_cap_reservation i_cap_migration_resv;
@@ -514,7 +513,6 @@ static inline struct inode *ceph_find_inode(struct super_block *sb,
* Ceph inode.
*/
#define CEPH_I_DIR_ORDERED (1 << 0) /* dentries in dir are ordered */
-#define CEPH_I_NODELAY (1 << 1) /* do not delay cap release */
#define CEPH_I_FLUSH (1 << 2) /* do not delay flush of dirty metadata */
#define CEPH_I_POOL_PERM (1 << 3) /* pool rd/wr bits are valid */
#define CEPH_I_POOL_RD (1 << 4) /* can read from pool */
--
2.21.1
^ permalink raw reply related
* [PATCH v4 6/9] drm/vmwgfx: Support huge page faults
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: Thomas Hellstrom, Michal Hocko, pv-drivers, Roland Scheidegger,
Dan Williams, Ralph Campbell, Matthew Wilcox (Oracle),
Jérôme Glisse, linux-graphics-maintainer, Andrew Morton,
Christian König, Kirill A. Shutemov
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
With vmwgfx dirty-tracking we need a specialized huge_fault
callback. Implement and hook it up.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 4 ++
drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c | 74 +++++++++++++++++++++-
drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c | 5 +-
3 files changed, 81 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index a31e726d6d71..82d86f2d2569 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -1428,6 +1428,10 @@ void vmw_bo_dirty_unmap(struct vmw_buffer_object *vbo,
pgoff_t start, pgoff_t end);
vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf);
vm_fault_t vmw_bo_vm_mkwrite(struct vm_fault *vmf);
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
+ enum page_entry_size pe_size);
+#endif
/**
* VMW_DEBUG_KMS - Debug output for kernel mode-setting
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
index 17a5dca7b921..cde3e07ebaf7 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
@@ -473,7 +473,7 @@ vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf)
* a lot of unnecessary write faults.
*/
if (vbo->dirty && vbo->dirty->method == VMW_BO_DIRTY_MKWRITE)
- prot = vma->vm_page_prot;
+ prot = vm_get_page_prot(vma->vm_flags & ~VM_SHARED);
else
prot = vm_get_page_prot(vma->vm_flags);
@@ -486,3 +486,75 @@ vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf)
return ret;
}
+
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
+ enum page_entry_size pe_size)
+{
+ struct vm_area_struct *vma = vmf->vma;
+ struct ttm_buffer_object *bo = (struct ttm_buffer_object *)
+ vma->vm_private_data;
+ struct vmw_buffer_object *vbo =
+ container_of(bo, struct vmw_buffer_object, base);
+ pgprot_t prot;
+ vm_fault_t ret;
+ pgoff_t fault_page_size;
+ bool write = vmf->flags & FAULT_FLAG_WRITE;
+ bool is_cow_mapping =
+ (vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE;
+
+ switch (pe_size) {
+ case PE_SIZE_PMD:
+ fault_page_size = HPAGE_PMD_SIZE >> PAGE_SHIFT;
+ break;
+#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
+ case PE_SIZE_PUD:
+ fault_page_size = HPAGE_PUD_SIZE >> PAGE_SHIFT;
+ break;
+#endif
+ default:
+ WARN_ON_ONCE(1);
+ return VM_FAULT_FALLBACK;
+ }
+
+ /* Always do write dirty-tracking and COW on PTE level. */
+ if (write && (READ_ONCE(vbo->dirty) || is_cow_mapping))
+ return VM_FAULT_FALLBACK;
+
+ ret = ttm_bo_vm_reserve(bo, vmf);
+ if (ret)
+ return ret;
+
+ if (vbo->dirty) {
+ pgoff_t allowed_prefault;
+ unsigned long page_offset;
+
+ page_offset = vmf->pgoff -
+ drm_vma_node_start(&bo->base.vma_node);
+ if (page_offset >= bo->num_pages ||
+ vmw_resources_clean(vbo, page_offset,
+ page_offset + PAGE_SIZE,
+ &allowed_prefault)) {
+ ret = VM_FAULT_SIGBUS;
+ goto out_unlock;
+ }
+
+ /*
+ * Write protect, so we get a new fault on write, and can
+ * split.
+ */
+ prot = vm_get_page_prot(vma->vm_flags & ~VM_SHARED);
+ } else {
+ prot = vm_get_page_prot(vma->vm_flags);
+ }
+
+ ret = ttm_bo_vm_fault_reserved(vmf, prot, 1, fault_page_size);
+ if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
+ return ret;
+
+out_unlock:
+ dma_resv_unlock(bo->base.resv);
+
+ return ret;
+}
+#endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
index ce288756531b..34100d1f5a9d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
@@ -34,7 +34,10 @@ int vmw_mmap(struct file *filp, struct vm_area_struct *vma)
.page_mkwrite = vmw_bo_vm_mkwrite,
.fault = vmw_bo_vm_fault,
.open = ttm_bo_vm_open,
- .close = ttm_bo_vm_close
+ .close = ttm_bo_vm_close,
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+ .huge_fault = vmw_bo_vm_huge_fault,
+#endif
};
struct drm_file *file_priv = filp->private_data;
struct vmw_private *dev_priv = vmw_priv(file_priv->minor->dev);
--
2.21.1
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related
* [PATCH v4 1/9] fs: Constify vma argument to vma_is_dax
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: pv-drivers, linux-graphics-maintainer, Thomas Hellstrom,
Andrew Morton, Michal Hocko, Matthew Wilcox (Oracle),
Kirill A. Shutemov, Ralph Campbell, Jérôme Glisse,
Christian König, Dan Williams, Roland Scheidegger
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
The vma argument is only dereferenced for reading.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
include/linux/fs.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 98e0349adb52..4f41fdbf402f 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3367,7 +3367,7 @@ static inline bool io_is_direct(struct file *filp)
return (filp->f_flags & O_DIRECT) || IS_DAX(filp->f_mapping->host);
}
-static inline bool vma_is_dax(struct vm_area_struct *vma)
+static inline bool vma_is_dax(const struct vm_area_struct *vma)
{
return vma->vm_file && IS_DAX(vma->vm_file->f_mapping->host);
}
--
2.21.1
^ permalink raw reply related
* [PATCH v4 2/9] mm: Introduce vma_is_special_huge
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: pv-drivers, linux-graphics-maintainer, Thomas Hellstrom,
Andrew Morton, Michal Hocko, Matthew Wilcox (Oracle),
Kirill A. Shutemov, Ralph Campbell, Jérôme Glisse,
Christian König, Dan Williams
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
For VM_PFNMAP and VM_MIXEDMAP vmas that want to support transhuge pages
and -page table entries, introduce vma_is_special_huge() that takes the
same codepaths as vma_is_dax().
The use of "special" follows the definition in memory.c, vm_normal_page():
"Special" mappings do not wish to be associated with a "struct page"
(either it doesn't exist, or it exists but they don't want to touch it)
For PAGE_SIZE pages, "special" is determined per page table entry to be
able to deal with COW pages. But since we don't have huge COW pages,
we can classify a vma as either "special huge" or "normal huge".
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
include/linux/mm.h | 17 +++++++++++++++++
mm/huge_memory.c | 6 +++---
2 files changed, 20 insertions(+), 3 deletions(-)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 0157d293935f..d370ce2932a1 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2822,6 +2822,23 @@ extern long copy_huge_page_from_user(struct page *dst_page,
const void __user *usr_src,
unsigned int pages_per_huge_page,
bool allow_pagefault);
+
+/**
+ * vma_is_special_huge - Are transhuge page-table entries considered special?
+ * @vma: Pointer to the struct vm_area_struct to consider
+ *
+ * Whether transhuge page-table entries are considered "special" following
+ * the definition in vm_normal_page().
+ *
+ * Return: true if transhuge page-table entries should be considered special,
+ * false otherwise.
+ */
+static inline bool vma_is_special_huge(const struct vm_area_struct *vma)
+{
+ return vma_is_dax(vma) || (vma->vm_file &&
+ (vma->vm_flags & (VM_PFNMAP | VM_MIXEDMAP)));
+}
+
#endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLBFS */
#ifdef CONFIG_DEBUG_PAGEALLOC
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 41a0fbddc96b..f8d24fc3f4df 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1789,7 +1789,7 @@ int zap_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
orig_pmd = pmdp_huge_get_and_clear_full(tlb->mm, addr, pmd,
tlb->fullmm);
tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
- if (vma_is_dax(vma)) {
+ if (vma_is_special_huge(vma)) {
if (arch_needs_pgtable_deposit())
zap_deposited_table(tlb->mm, pmd);
spin_unlock(ptl);
@@ -2053,7 +2053,7 @@ int zap_huge_pud(struct mmu_gather *tlb, struct vm_area_struct *vma,
*/
pudp_huge_get_and_clear_full(tlb->mm, addr, pud, tlb->fullmm);
tlb_remove_pud_tlb_entry(tlb, pud, addr);
- if (vma_is_dax(vma)) {
+ if (vma_is_special_huge(vma)) {
spin_unlock(ptl);
/* No zero page support yet */
} else {
@@ -2162,7 +2162,7 @@ static void __split_huge_pmd_locked(struct vm_area_struct *vma, pmd_t *pmd,
*/
if (arch_needs_pgtable_deposit())
zap_deposited_table(mm, pmd);
- if (vma_is_dax(vma))
+ if (vma_is_special_huge(vma))
return;
page = pmd_page(_pmd);
if (!PageDirty(page) && pmd_dirty(_pmd))
--
2.21.1
^ permalink raw reply related
* [PATCH v4 3/9] mm: Split huge pages on write-notify or COW
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: Thomas Hellstrom, Michal Hocko, pv-drivers, Dan Williams,
Ralph Campbell, Matthew Wilcox (Oracle), Jérôme Glisse,
linux-graphics-maintainer, Andrew Morton, Christian König,
Kirill A. Shutemov
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
We currently only do COW and write-notify on the PTE level, so if the
huge_fault() handler returns VM_FAULT_FALLBACK on wp faults,
split the huge pages and page-table entries. Also do this for huge PUDs
if there is no huge_fault() handler and the vma is not anonymous, similar
to how it's done for PMDs.
Note that fs/dax.c does the splitting in the huge_fault() handler, but as
huge_fault() is implemented by modules we need to consider whether to
export the splitting functions for use in the modules or whether to try
to keep calls in the core. Opt for the latter. A follow-up patch can
remove the dax.c split_huge_pmd() if needed.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
mm/memory.c | 27 +++++++++++++++++++--------
1 file changed, 19 insertions(+), 8 deletions(-)
diff --git a/mm/memory.c b/mm/memory.c
index 17aadc751e5c..4c49fe963e5c 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3880,11 +3880,14 @@ static inline vm_fault_t wp_huge_pmd(struct vm_fault *vmf, pmd_t orig_pmd)
{
if (vma_is_anonymous(vmf->vma))
return do_huge_pmd_wp_page(vmf, orig_pmd);
- if (vmf->vma->vm_ops->huge_fault)
- return vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PMD);
+ if (vmf->vma->vm_ops->huge_fault) {
+ vm_fault_t ret = vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PMD);
- /* COW handled on pte level: split pmd */
- VM_BUG_ON_VMA(vmf->vma->vm_flags & VM_SHARED, vmf->vma);
+ if (!(ret & VM_FAULT_FALLBACK))
+ return ret;
+ }
+
+ /* COW or write-notify handled on pte level: split pmd. */
__split_huge_pmd(vmf->vma, vmf->pmd, vmf->address, false, NULL);
return VM_FAULT_FALLBACK;
@@ -3897,12 +3900,20 @@ static inline bool vma_is_accessible(struct vm_area_struct *vma)
static vm_fault_t create_huge_pud(struct vm_fault *vmf)
{
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && \
+ defined(CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD)
/* No support for anonymous transparent PUD pages yet */
if (vma_is_anonymous(vmf->vma))
- return VM_FAULT_FALLBACK;
- if (vmf->vma->vm_ops->huge_fault)
- return vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PUD);
+ goto split;
+ if (vmf->vma->vm_ops->huge_fault) {
+ vm_fault_t ret = vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PUD);
+
+ if (!(ret & VM_FAULT_FALLBACK))
+ return ret;
+ }
+split:
+ /* COW or write-notify not handled on PUD level: split pud.*/
+ __split_huge_pud(vmf->vma, vmf->pud, vmf->address);
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
return VM_FAULT_FALLBACK;
}
--
2.21.1
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related
* [PATCH v4 9/9] drm/vmwgfx: Hook up the helpers to align buffer objects
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: pv-drivers, linux-graphics-maintainer, Thomas Hellstrom,
Andrew Morton, Michal Hocko, Matthew Wilcox (Oracle),
Kirill A. Shutemov, Ralph Campbell, Jérôme Glisse,
Christian König, Dan Williams, Roland Scheidegger
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
Start using the helpers that align buffer object user-space addresses and
buffer object vram addresses to huge page boundaries.
This is to improve the chances of allowing huge page-table entries.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
drivers/gpu/drm/drm_file.c | 1 +
drivers/gpu/drm/vmwgfx/vmwgfx_drv.c | 13 +++++++++++++
drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 1 +
drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c | 2 +-
4 files changed, 16 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/drm_file.c b/drivers/gpu/drm/drm_file.c
index 40fae356d202..1df2fca608c3 100644
--- a/drivers/gpu/drm/drm_file.c
+++ b/drivers/gpu/drm/drm_file.c
@@ -932,3 +932,4 @@ unsigned long drm_get_unmapped_area(struct file *file,
return current->mm->get_unmapped_area(file, uaddr, len, pgoff, flags);
}
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+EXPORT_SYMBOL_GPL(drm_get_unmapped_area);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index e962048f65d2..5452cabb4a2e 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -1215,6 +1215,18 @@ static void vmw_remove(struct pci_dev *pdev)
drm_put_dev(dev);
}
+static unsigned long
+vmw_get_unmapped_area(struct file *file, unsigned long uaddr,
+ unsigned long len, unsigned long pgoff,
+ unsigned long flags)
+{
+ struct drm_file *file_priv = file->private_data;
+ struct vmw_private *dev_priv = vmw_priv(file_priv->minor->dev);
+
+ return drm_get_unmapped_area(file, uaddr, len, pgoff, flags,
+ &dev_priv->vma_manager);
+}
+
static int vmwgfx_pm_notifier(struct notifier_block *nb, unsigned long val,
void *ptr)
{
@@ -1386,6 +1398,7 @@ static const struct file_operations vmwgfx_driver_fops = {
.compat_ioctl = vmw_compat_ioctl,
#endif
.llseek = noop_llseek,
+ .get_unmapped_area = vmw_get_unmapped_area,
};
static struct drm_driver driver = {
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 06267184aa0a..9ea145cffa3d 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -929,6 +929,7 @@ extern int vmw_mmap(struct file *filp, struct vm_area_struct *vma);
extern void vmw_validation_mem_init_ttm(struct vmw_private *dev_priv,
size_t gran);
+
/**
* TTM buffer object driver - vmwgfx_ttm_buffer.c
*/
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
index d8ea3dd10af0..34c721ab3ff3 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_buffer.c
@@ -754,7 +754,7 @@ static int vmw_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
break;
case TTM_PL_VRAM:
/* "On-card" video ram */
- man->func = &ttm_bo_manager_func;
+ man->func = &vmw_thp_func;
man->gpu_offset = 0;
man->flags = TTM_MEMTYPE_FLAG_FIXED | TTM_MEMTYPE_FLAG_MAPPABLE;
man->available_caching = TTM_PL_FLAG_CACHED;
--
2.21.1
^ permalink raw reply related
* [PATCH v4 3/9] mm: Split huge pages on write-notify or COW
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: pv-drivers, linux-graphics-maintainer, Thomas Hellstrom,
Andrew Morton, Michal Hocko, Matthew Wilcox (Oracle),
Kirill A. Shutemov, Ralph Campbell, Jérôme Glisse,
Christian König, Dan Williams
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
We currently only do COW and write-notify on the PTE level, so if the
huge_fault() handler returns VM_FAULT_FALLBACK on wp faults,
split the huge pages and page-table entries. Also do this for huge PUDs
if there is no huge_fault() handler and the vma is not anonymous, similar
to how it's done for PMDs.
Note that fs/dax.c does the splitting in the huge_fault() handler, but as
huge_fault() is implemented by modules we need to consider whether to
export the splitting functions for use in the modules or whether to try
to keep calls in the core. Opt for the latter. A follow-up patch can
remove the dax.c split_huge_pmd() if needed.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
mm/memory.c | 27 +++++++++++++++++++--------
1 file changed, 19 insertions(+), 8 deletions(-)
diff --git a/mm/memory.c b/mm/memory.c
index 17aadc751e5c..4c49fe963e5c 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -3880,11 +3880,14 @@ static inline vm_fault_t wp_huge_pmd(struct vm_fault *vmf, pmd_t orig_pmd)
{
if (vma_is_anonymous(vmf->vma))
return do_huge_pmd_wp_page(vmf, orig_pmd);
- if (vmf->vma->vm_ops->huge_fault)
- return vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PMD);
+ if (vmf->vma->vm_ops->huge_fault) {
+ vm_fault_t ret = vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PMD);
- /* COW handled on pte level: split pmd */
- VM_BUG_ON_VMA(vmf->vma->vm_flags & VM_SHARED, vmf->vma);
+ if (!(ret & VM_FAULT_FALLBACK))
+ return ret;
+ }
+
+ /* COW or write-notify handled on pte level: split pmd. */
__split_huge_pmd(vmf->vma, vmf->pmd, vmf->address, false, NULL);
return VM_FAULT_FALLBACK;
@@ -3897,12 +3900,20 @@ static inline bool vma_is_accessible(struct vm_area_struct *vma)
static vm_fault_t create_huge_pud(struct vm_fault *vmf)
{
-#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+#if defined(CONFIG_TRANSPARENT_HUGEPAGE) && \
+ defined(CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD)
/* No support for anonymous transparent PUD pages yet */
if (vma_is_anonymous(vmf->vma))
- return VM_FAULT_FALLBACK;
- if (vmf->vma->vm_ops->huge_fault)
- return vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PUD);
+ goto split;
+ if (vmf->vma->vm_ops->huge_fault) {
+ vm_fault_t ret = vmf->vma->vm_ops->huge_fault(vmf, PE_SIZE_PUD);
+
+ if (!(ret & VM_FAULT_FALLBACK))
+ return ret;
+ }
+split:
+ /* COW or write-notify not handled on PUD level: split pud.*/
+ __split_huge_pud(vmf->vma, vmf->pud, vmf->address);
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
return VM_FAULT_FALLBACK;
}
--
2.21.1
^ permalink raw reply related
* [PATCH v4 4/9] mm: Add vmf_insert_pfn_xxx_prot() for huge page-table entries
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: Thomas Hellstrom, Michal Hocko, pv-drivers, Dan Williams,
Ralph Campbell, Matthew Wilcox (Oracle), Jérôme Glisse,
linux-graphics-maintainer, Andrew Morton, Christian König,
Kirill A. Shutemov
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
For graphics drivers needing to modify the page-protection, add
huge page-table entries counterparts to vmf_insert_pfn_prot().
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
include/linux/huge_mm.h | 41 +++++++++++++++++++++++++++++++++++++++--
mm/huge_memory.c | 38 ++++++++++++++++++++++++++++++++------
2 files changed, 71 insertions(+), 8 deletions(-)
diff --git a/include/linux/huge_mm.h b/include/linux/huge_mm.h
index 0b84e13e88e2..a95d1bc8ffe8 100644
--- a/include/linux/huge_mm.h
+++ b/include/linux/huge_mm.h
@@ -47,8 +47,45 @@ extern bool move_huge_pmd(struct vm_area_struct *vma, unsigned long old_addr,
extern int change_huge_pmd(struct vm_area_struct *vma, pmd_t *pmd,
unsigned long addr, pgprot_t newprot,
int prot_numa);
-vm_fault_t vmf_insert_pfn_pmd(struct vm_fault *vmf, pfn_t pfn, bool write);
-vm_fault_t vmf_insert_pfn_pud(struct vm_fault *vmf, pfn_t pfn, bool write);
+vm_fault_t vmf_insert_pfn_pmd_prot(struct vm_fault *vmf, pfn_t pfn,
+ pgprot_t pgprot, bool write);
+
+/**
+ * vmf_insert_pfn_pmd - insert a pmd size pfn
+ * @vmf: Structure describing the fault
+ * @pfn: pfn to insert
+ * @pgprot: page protection to use
+ * @write: whether it's a write fault
+ *
+ * Insert a pmd size pfn. See vmf_insert_pfn() for additional info.
+ *
+ * Return: vm_fault_t value.
+ */
+static inline vm_fault_t vmf_insert_pfn_pmd(struct vm_fault *vmf, pfn_t pfn,
+ bool write)
+{
+ return vmf_insert_pfn_pmd_prot(vmf, pfn, vmf->vma->vm_page_prot, write);
+}
+vm_fault_t vmf_insert_pfn_pud_prot(struct vm_fault *vmf, pfn_t pfn,
+ pgprot_t pgprot, bool write);
+
+/**
+ * vmf_insert_pfn_pud - insert a pud size pfn
+ * @vmf: Structure describing the fault
+ * @pfn: pfn to insert
+ * @pgprot: page protection to use
+ * @write: whether it's a write fault
+ *
+ * Insert a pud size pfn. See vmf_insert_pfn() for additional info.
+ *
+ * Return: vm_fault_t value.
+ */
+static inline vm_fault_t vmf_insert_pfn_pud(struct vm_fault *vmf, pfn_t pfn,
+ bool write)
+{
+ return vmf_insert_pfn_pud_prot(vmf, pfn, vmf->vma->vm_page_prot, write);
+}
+
enum transparent_hugepage_flag {
TRANSPARENT_HUGEPAGE_FLAG,
TRANSPARENT_HUGEPAGE_REQ_MADV_FLAG,
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index f8d24fc3f4df..b2ec62cca3ae 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -811,11 +811,24 @@ static void insert_pfn_pmd(struct vm_area_struct *vma, unsigned long addr,
pte_free(mm, pgtable);
}
-vm_fault_t vmf_insert_pfn_pmd(struct vm_fault *vmf, pfn_t pfn, bool write)
+/**
+ * vmf_insert_pfn_pmd_prot - insert a pmd size pfn
+ * @vmf: Structure describing the fault
+ * @pfn: pfn to insert
+ * @pgprot: page protection to use
+ * @write: whether it's a write fault
+ *
+ * Insert a pmd size pfn. See vmf_insert_pfn() for additional info and
+ * also consult the vmf_insert_mixed_prot() documentation when
+ * @pgprot != @vmf->vma->vm_page_prot.
+ *
+ * Return: vm_fault_t value.
+ */
+vm_fault_t vmf_insert_pfn_pmd_prot(struct vm_fault *vmf, pfn_t pfn,
+ pgprot_t pgprot, bool write)
{
unsigned long addr = vmf->address & PMD_MASK;
struct vm_area_struct *vma = vmf->vma;
- pgprot_t pgprot = vma->vm_page_prot;
pgtable_t pgtable = NULL;
/*
@@ -843,7 +856,7 @@ vm_fault_t vmf_insert_pfn_pmd(struct vm_fault *vmf, pfn_t pfn, bool write)
insert_pfn_pmd(vma, addr, vmf->pmd, pfn, pgprot, write, pgtable);
return VM_FAULT_NOPAGE;
}
-EXPORT_SYMBOL_GPL(vmf_insert_pfn_pmd);
+EXPORT_SYMBOL_GPL(vmf_insert_pfn_pmd_prot);
#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
static pud_t maybe_pud_mkwrite(pud_t pud, struct vm_area_struct *vma)
@@ -889,11 +902,24 @@ static void insert_pfn_pud(struct vm_area_struct *vma, unsigned long addr,
spin_unlock(ptl);
}
-vm_fault_t vmf_insert_pfn_pud(struct vm_fault *vmf, pfn_t pfn, bool write)
+/**
+ * vmf_insert_pfn_pud_prot - insert a pud size pfn
+ * @vmf: Structure describing the fault
+ * @pfn: pfn to insert
+ * @pgprot: page protection to use
+ * @write: whether it's a write fault
+ *
+ * Insert a pud size pfn. See vmf_insert_pfn() for additional info and
+ * also consult the vmf_insert_mixed_prot() documentation when
+ * @pgprot != @vmf->vma->vm_page_prot.
+ *
+ * Return: vm_fault_t value.
+ */
+vm_fault_t vmf_insert_pfn_pud_prot(struct vm_fault *vmf, pfn_t pfn,
+ pgprot_t pgprot, bool write)
{
unsigned long addr = vmf->address & PUD_MASK;
struct vm_area_struct *vma = vmf->vma;
- pgprot_t pgprot = vma->vm_page_prot;
/*
* If we had pud_special, we could avoid all these restrictions,
@@ -914,7 +940,7 @@ vm_fault_t vmf_insert_pfn_pud(struct vm_fault *vmf, pfn_t pfn, bool write)
insert_pfn_pud(vma, addr, vmf->pud, pfn, pgprot, write);
return VM_FAULT_NOPAGE;
}
-EXPORT_SYMBOL_GPL(vmf_insert_pfn_pud);
+EXPORT_SYMBOL_GPL(vmf_insert_pfn_pud_prot);
#endif /* CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD */
static void touch_pmd(struct vm_area_struct *vma, unsigned long addr,
--
2.21.1
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related
* [PATCH v4 5/9] drm/ttm, drm/vmwgfx: Support huge TTM pagefaults
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: pv-drivers, linux-graphics-maintainer, Thomas Hellstrom,
Andrew Morton, Michal Hocko, Matthew Wilcox (Oracle),
Kirill A. Shutemov, Ralph Campbell, Jérôme Glisse,
Christian König, Dan Williams, Roland Scheidegger
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
Support huge (PMD-size and PUD-size) page-table entries by providing a
huge_fault() callback.
We still support private mappings and write-notify by splitting the huge
page-table entries on write-access.
Note that for huge page-faults to occur, either the kernel needs to be
compiled with trans-huge-pages always enabled, or the kernel needs to be
compiled with trans-huge-pages enabled using madvise, and the user-space
app needs to call madvise() to enable trans-huge pages on a per-mapping
basis.
Furthermore huge page-faults will not succeed unless buffer objects and
user-space addresses are aligned on huge page size boundaries.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
---
drivers/gpu/drm/ttm/ttm_bo_vm.c | 161 ++++++++++++++++++++-
drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c | 2 +-
include/drm/ttm/ttm_bo_api.h | 3 +-
3 files changed, 161 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index 389128b8c4dd..0af14835504c 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -156,6 +156,89 @@ vm_fault_t ttm_bo_vm_reserve(struct ttm_buffer_object *bo,
}
EXPORT_SYMBOL(ttm_bo_vm_reserve);
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+/**
+ * ttm_bo_vm_insert_huge - Insert a pfn for PUD or PMD faults
+ * @vmf: Fault data
+ * @bo: The buffer object
+ * @page_offset: Page offset from bo start
+ * @fault_page_size: The size of the fault in pages.
+ * @pgprot: The page protections.
+ * Does additional checking whether it's possible to insert a PUD or PMD
+ * pfn and performs the insertion.
+ *
+ * Return: VM_FAULT_NOPAGE on successful insertion, VM_FAULT_FALLBACK if
+ * a huge fault was not possible, or on insertion error.
+ */
+static vm_fault_t ttm_bo_vm_insert_huge(struct vm_fault *vmf,
+ struct ttm_buffer_object *bo,
+ pgoff_t page_offset,
+ pgoff_t fault_page_size,
+ pgprot_t pgprot)
+{
+ pgoff_t i;
+ vm_fault_t ret;
+ unsigned long pfn;
+ pfn_t pfnt;
+ struct ttm_tt *ttm = bo->ttm;
+ bool write = vmf->flags & FAULT_FLAG_WRITE;
+
+ /* Fault should not cross bo boundary. */
+ page_offset &= ~(fault_page_size - 1);
+ if (page_offset + fault_page_size > bo->num_pages)
+ goto out_fallback;
+
+ if (bo->mem.bus.is_iomem)
+ pfn = ttm_bo_io_mem_pfn(bo, page_offset);
+ else
+ pfn = page_to_pfn(ttm->pages[page_offset]);
+
+ /* pfn must be fault_page_size aligned. */
+ if ((pfn & (fault_page_size - 1)) != 0)
+ goto out_fallback;
+
+ /* Check that memory is contiguous. */
+ if (!bo->mem.bus.is_iomem) {
+ for (i = 1; i < fault_page_size; ++i) {
+ if (page_to_pfn(ttm->pages[page_offset + i]) != pfn + i)
+ goto out_fallback;
+ }
+ } else if (bo->bdev->driver->io_mem_pfn) {
+ for (i = 1; i < fault_page_size; ++i) {
+ if (ttm_bo_io_mem_pfn(bo, page_offset + i) != pfn + i)
+ goto out_fallback;
+ }
+ }
+
+ pfnt = __pfn_to_pfn_t(pfn, PFN_DEV);
+ if (fault_page_size == (HPAGE_PMD_SIZE >> PAGE_SHIFT))
+ ret = vmf_insert_pfn_pmd_prot(vmf, pfnt, pgprot, write);
+#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
+ else if (fault_page_size == (HPAGE_PUD_SIZE >> PAGE_SHIFT))
+ ret = vmf_insert_pfn_pud_prot(vmf, pfnt, pgprot, write);
+#endif
+ else
+ WARN_ON_ONCE(ret = VM_FAULT_FALLBACK);
+
+ if (ret != VM_FAULT_NOPAGE)
+ goto out_fallback;
+
+ return VM_FAULT_NOPAGE;
+out_fallback:
+ count_vm_event(THP_FAULT_FALLBACK);
+ return VM_FAULT_FALLBACK;
+}
+#else
+static vm_fault_t ttm_bo_vm_insert_huge(struct vm_fault *vmf,
+ struct ttm_buffer_object *bo,
+ pgoff_t page_offset,
+ pgoff_t fault_page_size,
+ pgprot_t pgprot)
+{
+ return VM_FAULT_FALLBACK;
+}
+#endif
+
/**
* ttm_bo_vm_fault_reserved - TTM fault helper
* @vmf: The struct vm_fault given as argument to the fault callback
@@ -163,6 +246,7 @@ EXPORT_SYMBOL(ttm_bo_vm_reserve);
* @num_prefault: Maximum number of prefault pages. The caller may want to
* specify this based on madvice settings and the size of the GPU object
* backed by the memory.
+ * @fault_page_size: The size of the fault in pages.
*
* This function inserts one or more page table entries pointing to the
* memory backing the buffer object, and then returns a return code
@@ -176,7 +260,8 @@ EXPORT_SYMBOL(ttm_bo_vm_reserve);
*/
vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
pgprot_t prot,
- pgoff_t num_prefault)
+ pgoff_t num_prefault,
+ pgoff_t fault_page_size)
{
struct vm_area_struct *vma = vmf->vma;
struct ttm_buffer_object *bo = vma->vm_private_data;
@@ -268,6 +353,13 @@ vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
prot = pgprot_decrypted(prot);
}
+ /* We don't prefault on huge faults. Yet. */
+ if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) && fault_page_size != 1) {
+ ret = ttm_bo_vm_insert_huge(vmf, bo, page_offset,
+ fault_page_size, prot);
+ goto out_io_unlock;
+ }
+
/*
* Speculatively prefault a number of pages. Only error on
* first page.
@@ -334,7 +426,7 @@ vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf)
return ret;
prot = vma->vm_page_prot;
- ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT);
+ ret = ttm_bo_vm_fault_reserved(vmf, prot, TTM_BO_VM_NUM_PREFAULT, 1);
if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
return ret;
@@ -344,6 +436,66 @@ vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf)
}
EXPORT_SYMBOL(ttm_bo_vm_fault);
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+/**
+ * ttm_pgprot_is_wrprotecting - Is a page protection value write-protecting?
+ * @prot: The page protection value
+ *
+ * Return: true if @prot is write-protecting. false otherwise.
+ */
+static bool ttm_pgprot_is_wrprotecting(pgprot_t prot)
+{
+ /*
+ * This is meant to say "pgprot_wrprotect(prot) == prot" in a generic
+ * way. Unfortunately there is no generic pgprot_wrprotect.
+ */
+ return pte_val(pte_wrprotect(__pte(pgprot_val(prot)))) ==
+ pgprot_val(prot);
+}
+
+static vm_fault_t ttm_bo_vm_huge_fault(struct vm_fault *vmf,
+ enum page_entry_size pe_size)
+{
+ struct vm_area_struct *vma = vmf->vma;
+ pgprot_t prot;
+ struct ttm_buffer_object *bo = vma->vm_private_data;
+ vm_fault_t ret;
+ pgoff_t fault_page_size = 0;
+ bool write = vmf->flags & FAULT_FLAG_WRITE;
+
+ switch (pe_size) {
+ case PE_SIZE_PMD:
+ fault_page_size = HPAGE_PMD_SIZE >> PAGE_SHIFT;
+ break;
+#ifdef CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
+ case PE_SIZE_PUD:
+ fault_page_size = HPAGE_PUD_SIZE >> PAGE_SHIFT;
+ break;
+#endif
+ default:
+ WARN_ON_ONCE(1);
+ return VM_FAULT_FALLBACK;
+ }
+
+ /* Fallback on write dirty-tracking or COW */
+ if (write && ttm_pgprot_is_wrprotecting(vma->vm_page_prot))
+ return VM_FAULT_FALLBACK;
+
+ ret = ttm_bo_vm_reserve(bo, vmf);
+ if (ret)
+ return ret;
+
+ prot = vm_get_page_prot(vma->vm_flags);
+ ret = ttm_bo_vm_fault_reserved(vmf, prot, 1, fault_page_size);
+ if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
+ return ret;
+
+ dma_resv_unlock(bo->base.resv);
+
+ return ret;
+}
+#endif
+
void ttm_bo_vm_open(struct vm_area_struct *vma)
{
struct ttm_buffer_object *bo = vma->vm_private_data;
@@ -445,7 +597,10 @@ static const struct vm_operations_struct ttm_bo_vm_ops = {
.fault = ttm_bo_vm_fault,
.open = ttm_bo_vm_open,
.close = ttm_bo_vm_close,
- .access = ttm_bo_vm_access
+ .access = ttm_bo_vm_access,
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+ .huge_fault = ttm_bo_vm_huge_fault,
+#endif
};
static struct ttm_buffer_object *ttm_bo_vm_lookup(struct ttm_bo_device *bdev,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
index f07aa857587c..17a5dca7b921 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_page_dirty.c
@@ -477,7 +477,7 @@ vm_fault_t vmw_bo_vm_fault(struct vm_fault *vmf)
else
prot = vm_get_page_prot(vma->vm_flags);
- ret = ttm_bo_vm_fault_reserved(vmf, prot, num_prefault);
+ ret = ttm_bo_vm_fault_reserved(vmf, prot, num_prefault, 1);
if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
return ret;
diff --git a/include/drm/ttm/ttm_bo_api.h b/include/drm/ttm/ttm_bo_api.h
index 66ca49db9633..4fc90d53aa15 100644
--- a/include/drm/ttm/ttm_bo_api.h
+++ b/include/drm/ttm/ttm_bo_api.h
@@ -732,7 +732,8 @@ vm_fault_t ttm_bo_vm_reserve(struct ttm_buffer_object *bo,
vm_fault_t ttm_bo_vm_fault_reserved(struct vm_fault *vmf,
pgprot_t prot,
- pgoff_t num_prefault);
+ pgoff_t num_prefault,
+ pgoff_t fault_page_size);
vm_fault_t ttm_bo_vm_fault(struct vm_fault *vmf);
--
2.21.1
^ permalink raw reply related
* [PATCH v4 8/9] drm/vmwgfx: Introduce a huge page aligning TTM range manager
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: Thomas Hellstrom, Michal Hocko, pv-drivers, Roland Scheidegger,
Dan Williams, Ralph Campbell, Matthew Wilcox (Oracle),
Jérôme Glisse, linux-graphics-maintainer, Andrew Morton,
Christian König, Kirill A. Shutemov
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
Using huge page-table entries requires that the physical address of the
start of a buffer object is huge page size aligned.
Make a special version of the TTM range manager that accomplishes this,
but falls back to a smaller page size alignment (PUD->PMD, PMD->NORMAL)
to avoid eviction.
If other drivers want to use it in the future, it can be made a
TTM generic helper. Note that drivers can force eviction for a certain
alignment by assigning the TTM GPU alignment correspondingly.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
drivers/gpu/drm/vmwgfx/Makefile | 1 +
drivers/gpu/drm/vmwgfx/vmwgfx_drv.h | 7 ++
drivers/gpu/drm/vmwgfx/vmwgfx_thp.c | 166 ++++++++++++++++++++++++++++
3 files changed, 174 insertions(+)
create mode 100644 drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
diff --git a/drivers/gpu/drm/vmwgfx/Makefile b/drivers/gpu/drm/vmwgfx/Makefile
index c877a21a0739..421dd2a497a5 100644
--- a/drivers/gpu/drm/vmwgfx/Makefile
+++ b/drivers/gpu/drm/vmwgfx/Makefile
@@ -11,4 +11,5 @@ vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \
vmwgfx_validation.o vmwgfx_page_dirty.o \
ttm_object.o ttm_lock.o
+vmwgfx-$(CONFIG_TRANSPARENT_HUGEPAGE) += vmwgfx_thp.o
obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index 82d86f2d2569..06267184aa0a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -1433,6 +1433,13 @@ vm_fault_t vmw_bo_vm_huge_fault(struct vm_fault *vmf,
enum page_entry_size pe_size);
#endif
+/* Transparent hugepage support - vmwgfx_thp.c */
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+extern const struct ttm_mem_type_manager_func vmw_thp_func;
+#else
+#define vmw_thp_func ttm_bo_manager_func
+#endif
+
/**
* VMW_DEBUG_KMS - Debug output for kernel mode-setting
*
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c b/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
new file mode 100644
index 000000000000..b7c816ba7166
--- /dev/null
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_thp.c
@@ -0,0 +1,166 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Huge page-table-entry support for IO memory.
+ *
+ * Copyright (C) 2007-2019 Vmware, Inc. All rights reservedd.
+ */
+#include "vmwgfx_drv.h"
+#include <drm/ttm/ttm_module.h>
+#include <drm/ttm/ttm_bo_driver.h>
+#include <drm/ttm/ttm_placement.h>
+
+/**
+ * struct vmw_thp_manager - Range manager implementing huge page alignment
+ *
+ * @mm: The underlying range manager. Protected by @lock.
+ * @lock: Manager lock.
+ */
+struct vmw_thp_manager {
+ struct drm_mm mm;
+ spinlock_t lock;
+};
+
+static int vmw_thp_insert_aligned(struct drm_mm *mm, struct drm_mm_node *node,
+ unsigned long align_pages,
+ const struct ttm_place *place,
+ struct ttm_mem_reg *mem,
+ unsigned long lpfn,
+ enum drm_mm_insert_mode mode)
+{
+ if (align_pages >= mem->page_alignment &&
+ (!mem->page_alignment || align_pages % mem->page_alignment == 0)) {
+ return drm_mm_insert_node_in_range(mm, node,
+ mem->num_pages,
+ align_pages, 0,
+ place->fpfn, lpfn, mode);
+ }
+
+ return -ENOSPC;
+}
+
+static int vmw_thp_get_node(struct ttm_mem_type_manager *man,
+ struct ttm_buffer_object *bo,
+ const struct ttm_place *place,
+ struct ttm_mem_reg *mem)
+{
+ struct vmw_thp_manager *rman = (struct vmw_thp_manager *) man->priv;
+ struct drm_mm *mm = &rman->mm;
+ struct drm_mm_node *node;
+ unsigned long align_pages;
+ unsigned long lpfn;
+ enum drm_mm_insert_mode mode = DRM_MM_INSERT_BEST;
+ int ret;
+
+ node = kzalloc(sizeof(*node), GFP_KERNEL);
+ if (!node)
+ return -ENOMEM;
+
+ lpfn = place->lpfn;
+ if (!lpfn)
+ lpfn = man->size;
+
+ mode = DRM_MM_INSERT_BEST;
+ if (place->flags & TTM_PL_FLAG_TOPDOWN)
+ mode = DRM_MM_INSERT_HIGH;
+
+ spin_lock(&rman->lock);
+ if (IS_ENABLED(CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD)) {
+ align_pages = (HPAGE_PUD_SIZE >> PAGE_SHIFT);
+ if (mem->num_pages >= align_pages) {
+ ret = vmw_thp_insert_aligned(mm, node, align_pages,
+ place, mem, lpfn, mode);
+ if (!ret)
+ goto found_unlock;
+ }
+ }
+
+ align_pages = (HPAGE_PMD_SIZE >> PAGE_SHIFT);
+ if (mem->num_pages >= align_pages) {
+ ret = vmw_thp_insert_aligned(mm, node, align_pages, place, mem,
+ lpfn, mode);
+ if (!ret)
+ goto found_unlock;
+ }
+
+ ret = drm_mm_insert_node_in_range(mm, node, mem->num_pages,
+ mem->page_alignment, 0,
+ place->fpfn, lpfn, mode);
+found_unlock:
+ spin_unlock(&rman->lock);
+
+ if (unlikely(ret)) {
+ kfree(node);
+ } else {
+ mem->mm_node = node;
+ mem->start = node->start;
+ }
+
+ return 0;
+}
+
+
+
+static void vmw_thp_put_node(struct ttm_mem_type_manager *man,
+ struct ttm_mem_reg *mem)
+{
+ struct vmw_thp_manager *rman = (struct vmw_thp_manager *) man->priv;
+
+ if (mem->mm_node) {
+ spin_lock(&rman->lock);
+ drm_mm_remove_node(mem->mm_node);
+ spin_unlock(&rman->lock);
+
+ kfree(mem->mm_node);
+ mem->mm_node = NULL;
+ }
+}
+
+static int vmw_thp_init(struct ttm_mem_type_manager *man,
+ unsigned long p_size)
+{
+ struct vmw_thp_manager *rman;
+
+ rman = kzalloc(sizeof(*rman), GFP_KERNEL);
+ if (!rman)
+ return -ENOMEM;
+
+ drm_mm_init(&rman->mm, 0, p_size);
+ spin_lock_init(&rman->lock);
+ man->priv = rman;
+ return 0;
+}
+
+static int vmw_thp_takedown(struct ttm_mem_type_manager *man)
+{
+ struct vmw_thp_manager *rman = (struct vmw_thp_manager *) man->priv;
+ struct drm_mm *mm = &rman->mm;
+
+ spin_lock(&rman->lock);
+ if (drm_mm_clean(mm)) {
+ drm_mm_takedown(mm);
+ spin_unlock(&rman->lock);
+ kfree(rman);
+ man->priv = NULL;
+ return 0;
+ }
+ spin_unlock(&rman->lock);
+ return -EBUSY;
+}
+
+static void vmw_thp_debug(struct ttm_mem_type_manager *man,
+ struct drm_printer *printer)
+{
+ struct vmw_thp_manager *rman = (struct vmw_thp_manager *) man->priv;
+
+ spin_lock(&rman->lock);
+ drm_mm_print(&rman->mm, printer);
+ spin_unlock(&rman->lock);
+}
+
+const struct ttm_mem_type_manager_func vmw_thp_func = {
+ .init = vmw_thp_init,
+ .takedown = vmw_thp_takedown,
+ .get_node = vmw_thp_get_node,
+ .put_node = vmw_thp_put_node,
+ .debug = vmw_thp_debug
+};
--
2.21.1
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related
* [PATCH v4 1/9] fs: Constify vma argument to vma_is_dax
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: Thomas Hellstrom, Michal Hocko, pv-drivers, Roland Scheidegger,
Dan Williams, Ralph Campbell, Matthew Wilcox (Oracle),
Jérôme Glisse, linux-graphics-maintainer, Andrew Morton,
Christian König, Kirill A. Shutemov
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
The vma argument is only dereferenced for reading.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
include/linux/fs.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 98e0349adb52..4f41fdbf402f 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3367,7 +3367,7 @@ static inline bool io_is_direct(struct file *filp)
return (filp->f_flags & O_DIRECT) || IS_DAX(filp->f_mapping->host);
}
-static inline bool vma_is_dax(struct vm_area_struct *vma)
+static inline bool vma_is_dax(const struct vm_area_struct *vma)
{
return vma->vm_file && IS_DAX(vma->vm_file->f_mapping->host);
}
--
2.21.1
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related
* [PATCH v4 0/9] Huge page-table entries for TTM
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: pv-drivers, linux-graphics-maintainer, Thomas Hellström,
Andrew Morton, Michal Hocko, Matthew Wilcox (Oracle),
Kirill A. Shutemov, Ralph Campbell, Jérôme Glisse,
Christian König, Dan Williams
In order to reduce TLB misses and CPU usage this patchset enables huge-
and giant page-table entries for TTM and TTM-enabled graphics drivers.
Patch 1 and 2 introduce a vma_is_special_huge() function to make the mm code
take the same path as DAX when splitting huge- and giant page table entries,
(which currently means zapping the page-table entry and rely on re-faulting).
Patch 3 makes the mm code split existing huge page-table entries
on huge_fault fallbacks. Typically on COW or on buffer-objects that want
write-notify. COW and write-notification is always done on the lowest
page-table level. See the patch log message for additional considerations.
Patch 4 introduces functions to allow the graphics drivers to manipulate
the caching- and encryption flags of huge page-table entries without ugly
hacks.
Patch 5 implements the huge_fault handler in TTM.
This enables huge page-table entries, provided that the kernel is configured
to support transhuge pages, either by default or using madvise().
However, they are unlikely to be inserted unless the kernel buffer object
pfns and user-space addresses align perfectly. There are various options
here, but since buffer objects that reside in system pages typically start
at huge page boundaries if they are backed by huge pages, we try to enforce
buffer object starting pfns and user-space addresses to be huge page-size
aligned if their size exceeds a huge page-size. If pud-size transhuge
("giant") pages are enabled by the arch, the same holds for those.
Patch 6 implements a specialized huge_fault handler for vmwgfx.
The vmwgfx driver may perform dirty-tracking and needs some special code
to handle that correctly.
Patch 7 implements a drm helper to align user-space addresses according
to the above scheme, if possible.
Patch 8 implements a TTM range manager for vmwgfx that does the same for
graphics IO memory. This may later be reused by other graphics drivers
if necessary.
Patch 9 finally hooks up the helpers of patch 7 and 8 to the vmwgfx driver.
A similar change is needed for graphics drivers that want a reasonable
likelyhood of actually using huge page-table entries.
If a buffer object size is not huge-page or giant-page aligned,
its size will NOT be inflated by this patchset. This means that the buffer
object tail will use smaller size page-table entries and thus no memory
overhead occurs. Drivers that want to pay the memory overhead price need to
implement their own scheme to inflate buffer-object sizes.
PMD size huge page-table-entries have been tested with vmwgfx and found to
work well both with system memory backed and IO memory backed buffer objects.
PUD size giant page-table-entries have seen limited (fault and COW) testing
using a modified kernel (to support 1GB page allocations) and a fake vmwgfx
TTM memory type. The vmwgfx driver does otherwise not support 1GB-size IO
memory resources.
Comments and suggestions welcome.
Thomas
Changes since RFC:
* Check for buffer objects present in contigous IO Memory (Christian König)
* Rebased on the vmwgfx emulated coherent memory functionality. That rebase
adds patch 5.
Changes since v1:
* Make the new TTM range manager vmwgfx-specific. (Christian König)
* Minor fixes for configs that don't support or only partially support
transhuge pages.
Changes since v2:
* Minor coding style and doc fixes in patch 5/9 (Christian König)
* Patch 5/9 doesn't touch mm. Remove from the patch title.
Changes since v3:
* Added reviews and acks
* Implemented ugly but generic ttm_pgprot_is_wrprotecting() instead of arch
specific code.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
^ permalink raw reply
* [PATCH v4 2/9] mm: Introduce vma_is_special_huge
From: Thomas Hellström (VMware) @ 2020-02-20 12:27 UTC (permalink / raw)
To: linux-mm, dri-devel, linux-kernel
Cc: Thomas Hellstrom, Michal Hocko, pv-drivers, Dan Williams,
Ralph Campbell, Matthew Wilcox (Oracle), Jérôme Glisse,
linux-graphics-maintainer, Andrew Morton, Christian König,
Kirill A. Shutemov
In-Reply-To: <20200220122719.4302-1-thomas_os@shipmail.org>
From: Thomas Hellstrom <thellstrom@vmware.com>
For VM_PFNMAP and VM_MIXEDMAP vmas that want to support transhuge pages
and -page table entries, introduce vma_is_special_huge() that takes the
same codepaths as vma_is_dax().
The use of "special" follows the definition in memory.c, vm_normal_page():
"Special" mappings do not wish to be associated with a "struct page"
(either it doesn't exist, or it exists but they don't want to touch it)
For PAGE_SIZE pages, "special" is determined per page table entry to be
able to deal with COW pages. But since we don't have huge COW pages,
we can classify a vma as either "special huge" or "normal huge".
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Acked-by: Christian König <christian.koenig@amd.com>
---
include/linux/mm.h | 17 +++++++++++++++++
mm/huge_memory.c | 6 +++---
2 files changed, 20 insertions(+), 3 deletions(-)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 0157d293935f..d370ce2932a1 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2822,6 +2822,23 @@ extern long copy_huge_page_from_user(struct page *dst_page,
const void __user *usr_src,
unsigned int pages_per_huge_page,
bool allow_pagefault);
+
+/**
+ * vma_is_special_huge - Are transhuge page-table entries considered special?
+ * @vma: Pointer to the struct vm_area_struct to consider
+ *
+ * Whether transhuge page-table entries are considered "special" following
+ * the definition in vm_normal_page().
+ *
+ * Return: true if transhuge page-table entries should be considered special,
+ * false otherwise.
+ */
+static inline bool vma_is_special_huge(const struct vm_area_struct *vma)
+{
+ return vma_is_dax(vma) || (vma->vm_file &&
+ (vma->vm_flags & (VM_PFNMAP | VM_MIXEDMAP)));
+}
+
#endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HUGETLBFS */
#ifdef CONFIG_DEBUG_PAGEALLOC
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index 41a0fbddc96b..f8d24fc3f4df 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1789,7 +1789,7 @@ int zap_huge_pmd(struct mmu_gather *tlb, struct vm_area_struct *vma,
orig_pmd = pmdp_huge_get_and_clear_full(tlb->mm, addr, pmd,
tlb->fullmm);
tlb_remove_pmd_tlb_entry(tlb, pmd, addr);
- if (vma_is_dax(vma)) {
+ if (vma_is_special_huge(vma)) {
if (arch_needs_pgtable_deposit())
zap_deposited_table(tlb->mm, pmd);
spin_unlock(ptl);
@@ -2053,7 +2053,7 @@ int zap_huge_pud(struct mmu_gather *tlb, struct vm_area_struct *vma,
*/
pudp_huge_get_and_clear_full(tlb->mm, addr, pud, tlb->fullmm);
tlb_remove_pud_tlb_entry(tlb, pud, addr);
- if (vma_is_dax(vma)) {
+ if (vma_is_special_huge(vma)) {
spin_unlock(ptl);
/* No zero page support yet */
} else {
@@ -2162,7 +2162,7 @@ static void __split_huge_pmd_locked(struct vm_area_struct *vma, pmd_t *pmd,
*/
if (arch_needs_pgtable_deposit())
zap_deposited_table(mm, pmd);
- if (vma_is_dax(vma))
+ if (vma_is_special_huge(vma))
return;
page = pmd_page(_pmd);
if (!PageDirty(page) && pmd_dirty(_pmd))
--
2.21.1
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel
^ permalink raw reply related
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.