qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Cornelia Huck <cohuck@redhat.com>
To: Gavin Shan <gshan@redhat.com>, qemu-arm@nongnu.org
Cc: qemu-devel@nongnu.org, maz@kernel.org, eric.auger@redhat.com,
	zhenyzha@redhat.com, richard.henderson@linaro.org,
	peter.maydell@linaro.org, shan.gavin@gmail.com
Subject: Re: [PATCH v6 7/7] hw/arm/virt: Add properties to disable high memory regions
Date: Wed, 26 Oct 2022 13:10:16 +0200	[thread overview]
Message-ID: <87r0yusuw7.fsf@redhat.com> (raw)
In-Reply-To: <9472b6fc-00dd-bf92-ed23-0f5a9328c45f@redhat.com>

On Wed, Oct 26 2022, Gavin Shan <gshan@redhat.com> wrote:

> Hi Connie,
>
> On 10/25/22 6:54 PM, Cornelia Huck wrote:
>> On Mon, Oct 24 2022, Gavin Shan <gshan@redhat.com> wrote:
>> 
>>> These 3 high memory regions are usually enabled by default, but
>> 
>> s/These 3/The/ ?
>> 
>
> Ok.
>
>>> they may be not used. For example, VIRT_HIGH_GIC_REDIST2 isn't
>>> needed by GICv2. This leads to waste in the PA space.
>> 
>> When building the command line, do we have enough information on when
>> the regions provide something useful, and when they just waste space?
>> 
>
> I think the help messages are already indicative. For example, the help
> messages for 'highmem-redist2' indicate the region is only needed by
> GICv3 or GICv4. 'highmem-ecam' and 'highmem-mmio' are needed by PCI ECAM
> and MMIO and the key words 'high' indicates they're the corresponding
> second window.
>
> #./qemu-system-aarch64 -M virt,?
> highmem-ecam=<bool>    - Set on/off to enable/disable high memory region for PCI ECAM
> highmem-mmio=<bool>    - Set on/off to enable/disable high memory region for PCI MMIO
> highmem-redists=<bool> - Set on/off to enable/disable high memory region for GICv3 or GICv4 redistributor

OK, hopefully this is enough for anyone building a command line
directly. (Do we want to encourage management software like libvirt to
switch off regions that are not needed?)

>
>>>
>>> Add properties to allow users selectively disable them if needed:
>>> "highmem-redists", "highmem-ecam", "highmem-mmio".
>>>
>>> Suggested-by: Marc Zyngier <maz@kernel.org>
>>> Signed-off-by: Gavin Shan <gshan@redhat.com>
>>> ---
>>>   docs/system/arm/virt.rst | 12 ++++++++
>>>   hw/arm/virt.c            | 64 ++++++++++++++++++++++++++++++++++++++++
>>>   2 files changed, 76 insertions(+)
>>>
>>> diff --git a/docs/system/arm/virt.rst b/docs/system/arm/virt.rst
>>> index 4454706392..a1668a969d 100644
>>> --- a/docs/system/arm/virt.rst
>>> +++ b/docs/system/arm/virt.rst
>>> @@ -98,6 +98,18 @@ compact-highmem
>>>     Set ``on``/``off`` to enable/disable the compact layout for high memory regions.
>>>     The default is ``on`` for machine types later than ``virt-7.2``.
>>>   
>>> +highmem-redists
>>> +  Set ``on``/``off`` to enable/disable the high memry region for GICv3/4
>> 
>> s/memry/memory/
>> 
>
> Ok, copy-and-paste error. Will be fixed.
>
>>> +  redistributor. The default is ``on``.
>> 
>> Do we need to add a note about what effects setting this to "off" may
>> have, e.g. "Setting this to ``off`` may limit the maximum number of
>> cpus." or so? And/or "Setting this to ``off`` when using GICv2 will save
>> some space."?
>> 
>
> We may not mention GICv2 since GICv3/v4 are already mentioned. It's a
> good idea to mention that the maximum number of CPUs is reduced when
> it's turned off. I will have something like below in next respin if
> you agree.
>
> highmem-redists
>    Set ``on``/``off`` to enable/disable the high memroy region for GICv3 or
>    GICv4 redistributor. The default is ``on``. Setting this to ``off`` will
>    limit the maximum number of CPUs when GICv3 or GICv4 is used.

OK, sounds reasonable to me.

>
> Since 'vms->highmem_redists' is changeable, the 'virt_max_cpus' in
> machvirt_init() needs to be recalculated based on that. The code change
> will be included into next respin. Besides, the follow-up error message
> will be improved to something like below.
>
>    error_report("Number of SMP CPUs requested (%d) exceeds max CPUs "
>                 "supported by machine 'mach-virt' (%d). The high memory "
>                 "region for GICv3 or GICv4 redistributor has been %s",
>                 max_cpus, virt_max_cpus,
>                 vms->highmem_redists ? "enabled" : "disabled");

Hm, the doc for error_report() states that "The resulting message should
be a single phrase, with no newline or trailing punctuation." Maybe

    if (max_cpus > virt_max_cpus) {
        error_report("Number of SMP CPUs requested (%d) exceeds max CPUs "
                     "supported by machine 'mach-virt' (%d)",
                     max_cpus, virt_max_cpus);
        if (vms->gic_version != VIRT_GIC_VERSION_2 &&
            !vms->higmem_redists) {
            error_printf("Try 'highmem-redists=on' for more CPUs\n");
        }
        exit(1);
    }



  reply	other threads:[~2022-10-26 11:13 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-24  3:54 [PATCH v6 0/7] hw/arm/virt: Improve address assignment for high memory regions Gavin Shan
2022-10-24  3:54 ` [PATCH v6 1/7] hw/arm/virt: Introduce virt_set_high_memmap() helper Gavin Shan
2022-10-24  3:54 ` [PATCH v6 2/7] hw/arm/virt: Rename variable size to region_size in virt_set_high_memmap() Gavin Shan
2022-10-24  3:54 ` [PATCH v6 3/7] hw/arm/virt: Introduce variable region_base " Gavin Shan
2022-10-24  3:54 ` [PATCH v6 4/7] hw/arm/virt: Introduce virt_get_high_memmap_enabled() helper Gavin Shan
2022-10-24  3:54 ` [PATCH v6 5/7] hw/arm/virt: Improve high memory region address assignment Gavin Shan
2022-10-25 16:29   ` Eric Auger
2022-10-26  0:33     ` Gavin Shan
2022-10-26 10:43       ` Cornelia Huck
2022-10-28  6:45         ` Gavin Shan
2022-10-24  3:54 ` [PATCH v6 6/7] hw/arm/virt: Add 'compact-highmem' property Gavin Shan
2022-10-25 10:30   ` Cornelia Huck
2022-10-25 16:33     ` Eric Auger
2022-10-26  3:16     ` Gavin Shan
2022-10-24  3:54 ` [PATCH v6 7/7] hw/arm/virt: Add properties to disable high memory regions Gavin Shan
2022-10-25 10:54   ` Cornelia Huck
2022-10-26  3:55     ` Gavin Shan
2022-10-26 11:10       ` Cornelia Huck [this message]
2022-10-28  6:53         ` Gavin Shan
2022-10-26  0:29 ` [PATCH v6 0/7] hw/arm/virt: Improve address assignment for " Gavin Shan
2022-10-28 18:06   ` Peter Maydell
2022-10-29 22:53     ` Gavin Shan
2022-10-29 11:29   ` Marc Zyngier
2022-10-29 22:49     ` Gavin Shan

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=87r0yusuw7.fsf@redhat.com \
    --to=cohuck@redhat.com \
    --cc=eric.auger@redhat.com \
    --cc=gshan@redhat.com \
    --cc=maz@kernel.org \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=richard.henderson@linaro.org \
    --cc=shan.gavin@gmail.com \
    --cc=zhenyzha@redhat.com \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).