From: "Michael R. Hines" <mrhines-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org>
To: "Michael S. Tsirkin" <mst-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: Roland Dreier <roland-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
Sean Hefty <sean.hefty-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>,
Hal Rosenstock
<hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>,
Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
Christoph Lameter <cl-vYTEC60ixJUAvxtiuMwx3w@public.gmane.org>,
linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
qemu-devel-qX2TKyscuCcdnm+yROfE0A@public.gmane.org
Subject: Re: [PATCH] rdma: don't make pages writeable if not requiested
Date: Thu, 21 Mar 2013 08:23:48 -0400 [thread overview]
Message-ID: <514AFBD4.2050201@linux.vnet.ibm.com> (raw)
In-Reply-To: <20130321061838.GA28319-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Yes, I'd be happy to try the patch.
Got meetings all day...... but will dive in soon.
On 03/21/2013 02:18 AM, Michael S. Tsirkin wrote:
> core/umem.c seems to get the arguments to get_user_pages
> in the reverse order: it sets writeable flag and
> breaks COW for MAP_SHARED if and only if hardware needs to
> write the page.
>
> This breaks memory overcommit for users such as KVM:
> each time we try to register a page to send it to remote, this
> breaks COW. It seems that for applications that only have
> REMOTE_READ permission, there is no reason to break COW at all.
>
> If the page that is COW has lots of copies, this makes the user process
> quickly exceed the cgroups memory limit. This makes RDMA mostly useless
> for virtualization, thus the stable tag.
>
> Reported-by: "Michael R. Hines" <mrhines-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org>
> Cc: stable-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> Signed-off-by: Michael S. Tsirkin <mst-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
> ---
>
> Note: compile-tested only, I don't have RDMA hardware at the moment.
> Michael, could you please try this patch (also fixing your
> usespace code not to request write access) and report?
>
> Note2: grep for get_user_pages in infiniband drivers turns up
> lots of users who set write to 1 unconditionally.
> These might be bugs too, should be checked.
>
> drivers/infiniband/core/umem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
> index a841123..5929598 100644
> --- a/drivers/infiniband/core/umem.c
> +++ b/drivers/infiniband/core/umem.c
> @@ -152,7 +152,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
> ret = get_user_pages(current, current->mm, cur_base,
> min_t(unsigned long, npages,
> PAGE_SIZE / sizeof (struct page *)),
> - 1, !umem->writable, page_list, vma_list);
> + !umem->writable, 1, page_list, vma_list);
>
> if (ret < 0)
> goto out;
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
WARNING: multiple messages have this Message-ID (diff)
From: "Michael R. Hines" <mrhines@linux.vnet.ibm.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Roland Dreier <roland@kernel.org>,
Sean Hefty <sean.hefty@intel.com>,
Hal Rosenstock <hal.rosenstock@gmail.com>,
Yishai Hadas <yishaih@mellanox.com>,
Christoph Lameter <cl@linux.com>,
linux-rdma@vger.kernel.org, linux-kernel@vger.kernel.org,
qemu-devel@nongnu.org
Subject: Re: [PATCH] rdma: don't make pages writeable if not requiested
Date: Thu, 21 Mar 2013 08:23:48 -0400 [thread overview]
Message-ID: <514AFBD4.2050201@linux.vnet.ibm.com> (raw)
In-Reply-To: <20130321061838.GA28319@redhat.com>
Yes, I'd be happy to try the patch.
Got meetings all day...... but will dive in soon.
On 03/21/2013 02:18 AM, Michael S. Tsirkin wrote:
> core/umem.c seems to get the arguments to get_user_pages
> in the reverse order: it sets writeable flag and
> breaks COW for MAP_SHARED if and only if hardware needs to
> write the page.
>
> This breaks memory overcommit for users such as KVM:
> each time we try to register a page to send it to remote, this
> breaks COW. It seems that for applications that only have
> REMOTE_READ permission, there is no reason to break COW at all.
>
> If the page that is COW has lots of copies, this makes the user process
> quickly exceed the cgroups memory limit. This makes RDMA mostly useless
> for virtualization, thus the stable tag.
>
> Reported-by: "Michael R. Hines" <mrhines@linux.vnet.ibm.com>
> Cc: stable@vger.kernel.org
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
>
> Note: compile-tested only, I don't have RDMA hardware at the moment.
> Michael, could you please try this patch (also fixing your
> usespace code not to request write access) and report?
>
> Note2: grep for get_user_pages in infiniband drivers turns up
> lots of users who set write to 1 unconditionally.
> These might be bugs too, should be checked.
>
> drivers/infiniband/core/umem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
> index a841123..5929598 100644
> --- a/drivers/infiniband/core/umem.c
> +++ b/drivers/infiniband/core/umem.c
> @@ -152,7 +152,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
> ret = get_user_pages(current, current->mm, cur_base,
> min_t(unsigned long, npages,
> PAGE_SIZE / sizeof (struct page *)),
> - 1, !umem->writable, page_list, vma_list);
> + !umem->writable, 1, page_list, vma_list);
>
> if (ret < 0)
> goto out;
WARNING: multiple messages have this Message-ID (diff)
From: "Michael R. Hines" <mrhines@linux.vnet.ibm.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Roland Dreier <roland@kernel.org>,
qemu-devel@nongnu.org, linux-rdma@vger.kernel.org,
Yishai Hadas <yishaih@mellanox.com>,
linux-kernel@vger.kernel.org,
Hal Rosenstock <hal.rosenstock@gmail.com>,
Sean Hefty <sean.hefty@intel.com>,
Christoph Lameter <cl@linux.com>
Subject: Re: [Qemu-devel] [PATCH] rdma: don't make pages writeable if not requiested
Date: Thu, 21 Mar 2013 08:23:48 -0400 [thread overview]
Message-ID: <514AFBD4.2050201@linux.vnet.ibm.com> (raw)
In-Reply-To: <20130321061838.GA28319@redhat.com>
Yes, I'd be happy to try the patch.
Got meetings all day...... but will dive in soon.
On 03/21/2013 02:18 AM, Michael S. Tsirkin wrote:
> core/umem.c seems to get the arguments to get_user_pages
> in the reverse order: it sets writeable flag and
> breaks COW for MAP_SHARED if and only if hardware needs to
> write the page.
>
> This breaks memory overcommit for users such as KVM:
> each time we try to register a page to send it to remote, this
> breaks COW. It seems that for applications that only have
> REMOTE_READ permission, there is no reason to break COW at all.
>
> If the page that is COW has lots of copies, this makes the user process
> quickly exceed the cgroups memory limit. This makes RDMA mostly useless
> for virtualization, thus the stable tag.
>
> Reported-by: "Michael R. Hines" <mrhines@linux.vnet.ibm.com>
> Cc: stable@vger.kernel.org
> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> ---
>
> Note: compile-tested only, I don't have RDMA hardware at the moment.
> Michael, could you please try this patch (also fixing your
> usespace code not to request write access) and report?
>
> Note2: grep for get_user_pages in infiniband drivers turns up
> lots of users who set write to 1 unconditionally.
> These might be bugs too, should be checked.
>
> drivers/infiniband/core/umem.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/core/umem.c b/drivers/infiniband/core/umem.c
> index a841123..5929598 100644
> --- a/drivers/infiniband/core/umem.c
> +++ b/drivers/infiniband/core/umem.c
> @@ -152,7 +152,7 @@ struct ib_umem *ib_umem_get(struct ib_ucontext *context, unsigned long addr,
> ret = get_user_pages(current, current->mm, cur_base,
> min_t(unsigned long, npages,
> PAGE_SIZE / sizeof (struct page *)),
> - 1, !umem->writable, page_list, vma_list);
> + !umem->writable, 1, page_list, vma_list);
>
> if (ret < 0)
> goto out;
next prev parent reply other threads:[~2013-03-21 12:23 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-21 6:18 [PATCH] rdma: don't make pages writeable if not requiested Michael S. Tsirkin
2013-03-21 6:18 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 6:18 ` Michael S. Tsirkin
[not found] ` <20130321061838.GA28319-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 6:55 ` Roland Dreier
2013-03-21 6:55 ` [Qemu-devel] " Roland Dreier
2013-03-21 6:55 ` Roland Dreier
[not found] ` <CAL1RGDUcMj9QVsuQgK+ozw64L6-cGehL7YBUJ1_ckni6TD=Kcg-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2013-03-21 7:03 ` Michael S. Tsirkin
2013-03-21 7:03 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 7:03 ` Michael S. Tsirkin
[not found] ` <20130321070357.GD28328-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 7:15 ` Roland Dreier
2013-03-21 7:15 ` [Qemu-devel] " Roland Dreier
2013-03-21 7:15 ` Roland Dreier
[not found] ` <CAG4TOxPkhOhGmzeA1K4a0Zw8HxS-QkOr-PCx7mJgA+KkuH3ZiQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2013-03-21 8:51 ` Michael S. Tsirkin
2013-03-21 8:51 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 8:51 ` Michael S. Tsirkin
[not found] ` <20130321085107.GE28328-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 9:13 ` Roland Dreier
2013-03-21 9:13 ` [Qemu-devel] " Roland Dreier
2013-03-21 9:13 ` Roland Dreier
[not found] ` <CAL1RGDVnkLZU2Vge4o3BwDxnAfGv7TQRMqE6ha3MUt39CVp5NQ-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2013-03-21 9:39 ` Michael S. Tsirkin
2013-03-21 9:39 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 9:39 ` Michael S. Tsirkin
[not found] ` <20130321093946.GG28328-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 17:11 ` Jason Gunthorpe
2013-03-21 17:11 ` [Qemu-devel] " Jason Gunthorpe
2013-03-21 17:11 ` Jason Gunthorpe
[not found] ` <20130321171115.GA653-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2013-03-21 17:15 ` Michael S. Tsirkin
2013-03-21 17:15 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 17:15 ` Michael S. Tsirkin
[not found] ` <20130321171525.GE2994-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 17:21 ` Jason Gunthorpe
2013-03-21 17:21 ` [Qemu-devel] " Jason Gunthorpe
2013-03-21 17:21 ` Jason Gunthorpe
[not found] ` <20130321172150.GA3118-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2013-03-21 17:42 ` Michael S. Tsirkin
2013-03-21 17:42 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 17:42 ` Michael S. Tsirkin
[not found] ` <20130321174237.GA4060-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 17:57 ` Jason Gunthorpe
2013-03-21 17:57 ` [Qemu-devel] " Jason Gunthorpe
2013-03-21 17:57 ` Jason Gunthorpe
[not found] ` <20130321175732.GA3263-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2013-03-21 18:03 ` Michael S. Tsirkin
2013-03-21 18:03 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 18:03 ` Michael S. Tsirkin
2013-03-21 18:16 ` Michael S. Tsirkin
2013-03-21 18:16 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 18:16 ` Michael S. Tsirkin
[not found] ` <20130321181633.GC4366-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 18:41 ` Jason Gunthorpe
2013-03-21 18:41 ` [Qemu-devel] " Jason Gunthorpe
2013-03-21 18:41 ` Jason Gunthorpe
[not found] ` <20130321184135.GA8044-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>
2013-03-21 19:15 ` Michael S. Tsirkin
2013-03-21 19:15 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 19:15 ` Michael S. Tsirkin
[not found] ` <20130321191541.GB5272-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 20:09 ` Jason Gunthorpe
2013-03-21 20:09 ` [Qemu-devel] " Jason Gunthorpe
2013-03-21 20:09 ` Jason Gunthorpe
2013-03-21 9:32 ` Michael S. Tsirkin
2013-03-21 9:32 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 9:32 ` Michael S. Tsirkin
[not found] ` <20130321093230.GF28328-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2013-03-21 11:30 ` Michael S. Tsirkin
2013-03-21 11:30 ` [Qemu-devel] " Michael S. Tsirkin
2013-03-21 11:30 ` Michael S. Tsirkin
2013-03-21 12:23 ` Michael R. Hines [this message]
2013-03-21 12:23 ` [Qemu-devel] " Michael R. Hines
2013-03-21 12:23 ` Michael R. Hines
2013-03-21 12:32 ` Michael S. Tsirkin
2013-03-21 12:32 ` [Qemu-devel] " Michael S. Tsirkin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=514AFBD4.2050201@linux.vnet.ibm.com \
--to=mrhines-23vcf4htsmix0ybbhkvfkdbpr1lh4cv8@public.gmane.org \
--cc=cl-vYTEC60ixJUAvxtiuMwx3w@public.gmane.org \
--cc=hal.rosenstock-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
--cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
--cc=mst-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
--cc=qemu-devel-qX2TKyscuCcdnm+yROfE0A@public.gmane.org \
--cc=roland-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org \
--cc=sean.hefty-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
--cc=yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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.