[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH v2 4/7] s390x/css: add missing css state conditi
From: |
Juan Quintela |
Subject: |
Re: [Qemu-devel] [PATCH v2 4/7] s390x/css: add missing css state conditionally |
Date: |
Wed, 31 May 2017 20:52:29 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/25.2 (gnu/linux) |
Halil Pasic <address@hidden> wrote:
> Although we have recently vmstatified the migration of some css
> infrastructure, for some css entities there is still state to be
> migrated left, because the focus was keeping migration stream
> compatibility (that is basically everything as-is).
>
> Let us add vmstate helpers and extend existing vmstate descriptions so
> that we have everything we need. Let us guard the added state with via
> css_migration_enabled, so we keep the compatible behavior if css
> migration is disabled.
>
> Let's also annotate the bits which do not need to be migrated for better
> readability.
>
> Signed-off-by: Halil Pasic <address@hidden>
> ---
> hw/intc/s390_flic.c | 20 +++++++++++++++
> hw/s390x/css.c | 74
> +++++++++++++++++++++++++++++++++++++++++++++++++++++
> 2 files changed, 94 insertions(+)
Reviewed-by: Juan Quintela <address@hidden>
For the vmstate bits. I have exactly zero clue about s390
> +static const VMStateDescription vmstate_chp_info = {
> + .name = "s390_chp_info",
> + .version_id = 1,
> + .minimum_version_id = 1,
> + .fields = (VMStateField[]) {
> + VMSTATE_UINT8(in_use, ChpInfo),
> + VMSTATE_UINT8(type, ChpInfo),
> + VMSTATE_UINT8(is_virtual, ChpInfo),
> + VMSTATE_END_OF_LIST()
> + }
> +};
> +
> typedef struct SubchSet {
> SubchDev *sch[MAX_SCHID + 1];
> unsigned long schids_used[BITS_TO_LONGS(MAX_SCHID + 1)];
> @@ -215,6 +248,19 @@ typedef struct CssImage {
> ChpInfo chpids[MAX_CHPID + 1];
> } CssImage;
>
> +static const VMStateDescription vmstate_css_img = {
> + .name = "s390_css_img",
> + .version_id = 1,
> + .minimum_version_id = 1,
> + .fields = (VMStateField[]) {
> + /* Subchannel sets have no relevant state. */
> + VMSTATE_STRUCT_ARRAY(chpids, CssImage, MAX_CHPID + 1, 0,
> + vmstate_chp_info, ChpInfo),
> + VMSTATE_END_OF_LIST()
> + }
> +
> +};
> +static const VMStateDescription vmstate_css = {
> + .name = "s390_css",
> + .version_id = 1,
> + .minimum_version_id = 1,
> + .fields = (VMStateField[]) {
> + VMSTATE_QTAILQ_V(pending_crws, ChannelSubSys, 1,
> vmstate_crw_container,
> + CrwContainer, sibling),
> + VMSTATE_BOOL(sei_pending, ChannelSubSys),
> + VMSTATE_BOOL(do_crw_mchk, ChannelSubSys),
> + VMSTATE_BOOL(crws_lost, ChannelSubSys),
> + /* These were kind of migrated by virtio */
> + VMSTATE_UINT8(max_cssid, ChannelSubSys),
> + VMSTATE_UINT8(max_ssid, ChannelSubSys),
> + VMSTATE_BOOL(chnmon_active, ChannelSubSys),
> + VMSTATE_UINT64(chnmon_area, ChannelSubSys),
> + VMSTATE_ARRAY_OF_POINTER_TO_STRUCT(css, ChannelSubSys, MAX_CSSID + 1,
> + 0, vmstate_css_img, CssImage),
I was about to suggest to move css from pointer to an embedded struct,
and then noticed that MAX_CSSID is .... 65535. I guess that this is
going to be sparse, very sparse. Perhaps there is an easier way to
transmit it?
You are transmiting:
65000 structs each of size MAX_CHPID (255) and each element is byte +
byte +byte = 65000 * 255 * 3 = 47 MB.
If it is really sparse, I think that sending something like a list
of <index in array> <contents> could make sense, no?
Or I am missunderstanding something?
Later, Juan.
- [Qemu-devel] [PATCH v2 0/7] migration: s390x css migration, Halil Pasic, 2017/05/29
- [Qemu-devel] [PATCH v2 2/7] s390x: add helper get_machine_class, Halil Pasic, 2017/05/29
- [Qemu-devel] [PATCH v2 4/7] s390x/css: add missing css state conditionally, Halil Pasic, 2017/05/29
- Re: [Qemu-devel] [PATCH v2 4/7] s390x/css: add missing css state conditionally,
Juan Quintela <=
- [Qemu-devel] [PATCH v2 1/7] s390x: vmstatify config migration for virtio-ccw, Halil Pasic, 2017/05/29
- [Qemu-devel] [PATCH v2 3/7] s390x: add css_migration_enabled to machine class, Halil Pasic, 2017/05/29
- [Qemu-devel] [PATCH v2 6/7] s390x/css: activate ChannelSubSys migration, Halil Pasic, 2017/05/29
- [Qemu-devel] [PATCH v2 5/7] s390x/css: add ORB to SubchDev, Halil Pasic, 2017/05/29
- [Qemu-devel] [PATCH v2 7/7] s390x/css: use SubchDev.orb, Halil Pasic, 2017/05/29
- Re: [Qemu-devel] [PATCH v2 0/7] migration: s390x css migration, Dr. David Alan Gilbert, 2017/05/31