From: "Michael S. Tsirkin" <mst@redhat.com>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>,
lersek@redhat.com, kraxel@redhat.com, pbonzini@redhat.com
Subject: [Qemu-devel] [PATCH v4 0/2] future proof rom loading for cross versiom migration
Date: Mon, 19 Aug 2013 17:26:49 +0300 [thread overview]
Message-ID: <1376922370-5681-1-git-send-email-mst@redhat.com> (raw)
Changes from v3:
- drop new API, all changes are now internal to memory subsystem
Patch 1/2 is new. Patch 2/2 had minor changes.
At laszlo's request (offline), here's the diff of patch 2/2
from v3 to v4:
diff --git a/hw/core/loader.c b/hw/core/loader.c
index 32d807a..7b3d3ee 100644
--- a/hw/core/loader.c
+++ b/hw/core/loader.c
@@ -610,15 +610,10 @@ static void rom_insert(Rom *rom)
static void *rom_set_mr(Rom *rom, Object *owner, const char *name)
{
- /*
- * Migration code expects that all RAM blocks are full pages.
- * Round MR size up to satisfy this condition.
- */
- unsigned size = ROUND_UP(rom->datasize, qemu_migration_page_size());
void *data;
rom->mr = g_malloc(sizeof(*rom->mr));
- memory_region_init_ram(rom->mr, owner, name, size);
+ memory_region_init_ram(rom->mr, owner, name, rom->datasize);
memory_region_set_readonly(rom->mr, true);
vmstate_register_ram_global(rom->mr);
Changes from v2: address comments on v2 by Peter Maydell
- switch from global constant to function
- use memory_region_init_ram instead of _ram_ptr
- disable for 1.6
Changes from v1: address comments by Peter Maydell
- drop useless data=data line
- rename target_page_size to migration_page_size to make use clear
Peter, you also suggested somehow hiding this within memory core.
I don't see a clean way to do this without lots of code
changes, I think what I propose here is acceptable for now
and we can always rework APIs without wire format changes.
Please review, and consider for merging.
Original cover letter below.
ROM files that are put in FW CFG are copied to guest ram, by BIOS, but
they are not backed by RAM so they don't get migrated.
Each time we'll change at least two bytes in such a ROM this will break
cross-version migration: since we can migrate after BIOS has read the first
byte but before it has read the second one, getting an inconsistent state.
This patchset makes QEMU future-proof against such changes.
Naturally, this only helps for -M 1.7 and up, older machine types
will still have the cross-version migration bug.
I think this should be applied for 1.7, this way we won't
have this problem from 1.8 and on.
Note: compatibility code depends on patch
'pc: cleanup 1.4 compat support' to work correctly
for machine types 1.5 and back.
Michael S. Tsirkin (2):
arch_init: align MR size to target page size
loader: store FW CFG ROM files in RAM
arch_init.c | 3 ++-
hw/core/loader.c | 49 ++++++++++++++++++++++++++++++++++++++++++++++---
hw/i386/pc_piix.c | 2 ++
hw/i386/pc_q35.c | 2 ++
include/hw/loader.h | 1 +
5 files changed, 53 insertions(+), 4 deletions(-)
--
MST
next reply other threads:[~2013-08-19 14:25 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-08-19 14:26 Michael S. Tsirkin [this message]
2013-08-19 14:26 ` [Qemu-devel] [PATCH v4 1/2] arch_init: align MR size to target page size Michael S. Tsirkin
2013-08-19 17:37 ` Laszlo Ersek
2013-08-19 17:45 ` Peter Maydell
2013-08-19 17:51 ` Michael S. Tsirkin
2013-08-19 17:48 ` Michael S. Tsirkin
2013-08-19 18:04 ` Laszlo Ersek
2013-08-19 14:26 ` [Qemu-devel] [PATCH v4 2/2] loader: store FW CFG ROM files in RAM Michael S. Tsirkin
2013-08-19 17:28 ` Paolo Bonzini
2013-08-19 17:40 ` Michael S. Tsirkin
2013-08-19 17:50 ` Paolo Bonzini
2013-08-19 18:05 ` Laszlo Ersek
2013-08-19 16:35 ` [Qemu-devel] [PATCH v4 0/2] future proof rom loading for cross versiom migration Peter Maydell
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=1376922370-5681-1-git-send-email-mst@redhat.com \
--to=mst@redhat.com \
--cc=kraxel@redhat.com \
--cc=lersek@redhat.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.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 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).