Patchwork Re: [PATCH 0/6] option rom patches, next round

login
register
mail settings
Submitter Gerd Hoffmann
Date Dec. 17, 2009, 2:46 p.m.
Message ID <4B2A443B.6090407@redhat.com>
Download mbox | patch
Permalink /patch/41321/
State New
Headers show

Comments

Gerd Hoffmann - Dec. 17, 2009, 2:46 p.m.
On 12/17/09 13:32, Gerd Hoffmann wrote:
>    Hi,
>
> All in one patch series for pci rom bar support and option rom loading
> via fw_cfg for the non-pci roms.  This time the fw_cfg interface is
> actually tested with a little linux userspace app (attached below).

Now some seabios bits.

I got the naming of *both* vga and optionrom directories wrong, so there 
is first an incremental fix for this patch series.

Also attached are two seabios patches (go on top of the two ones from 
anthony) which implement the seabios side of the file interface.

'-kernel' and '-vga std' are working again ;)

cheers,
   Gerd
From 517fd48fdddfe93b5ef4081c11e87b250892cd49 Mon Sep 17 00:00:00 2001
From: Gerd Hoffmann <kraxel@redhat.com>
Date: Thu, 17 Dec 2009 15:32:03 +0100
Subject: [PATCH] loader: fix dirname tyops


Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 hw/loader.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
Anthony Liguori - Dec. 17, 2009, 8:22 p.m.
Gerd Hoffmann wrote:
> On 12/17/09 13:32, Gerd Hoffmann wrote:
>>    Hi,
>>
>> All in one patch series for pci rom bar support and option rom loading
>> via fw_cfg for the non-pci roms.  This time the fw_cfg interface is
>> actually tested with a little linux userspace app (attached below).
>
> Now some seabios bits.
>
> I got the naming of *both* vga and optionrom directories wrong, so 
> there is first an incremental fix for this patch series.
>
> Also attached are two seabios patches (go on top of the two ones from 
> anthony) which implement the seabios side of the file interface.
>
> '-kernel' and '-vga std' are working again ;)

Very nice!  No reason not to convert std vga and vmware vga to use the 
pci based rom loading interface.

Care to fold everything into one series that we can bring into the tree?

> cheers,
>   Gerd
Gerd Hoffmann - Dec. 18, 2009, 9:30 a.m.
On 12/17/09 21:22, Anthony Liguori wrote:
> Very nice! No reason not to convert std vga and vmware vga to use the
> pci based rom loading interface.

seabios thinks about vgabios.bin:

Scan for VGA option rom
Attempting to init PCI bdf 00:02.0 (dev/ven 11111234)
Attempting to map option rom on dev 00:02.0
Option rom sizing returned f0800000 ffff0000
Inspecting possible rom at 0xf0800000 (dv=11111234 bdf=10)
No valid pci signature found
^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Adding a pci rom header (with vendor and product id which seabios will 
check) doesn't look that useful for a generic vga bios though.  If we go 
down that route we'll need at least two roms then (one for std, one for 
vmware-svga) where the only difference is the pci header ...

cheers,
   Gerd

Patch

diff --git a/hw/loader.c b/hw/loader.c
index 451ee54..ccc0ccc 100644
--- a/hw/loader.c
+++ b/hw/loader.c
@@ -628,7 +628,7 @@  int rom_add_vga(const char *file)
 {
     if (!rom_enable_driver_roms)
         return 0;
-    return rom_add_file(file, "vgabios", file,
+    return rom_add_file(file, "vgaroms", file,
                         PC_ROM_MIN_VGA, PC_ROM_MAX, PC_ROM_ALIGN);
 }
 
@@ -636,7 +636,7 @@  int rom_add_option(const char *file)
 {
     if (!rom_enable_driver_roms)
         return 0;
-    return rom_add_file(file, "genrom", file,
+    return rom_add_file(file, "genroms", file,
                         PC_ROM_MIN_OPTION, PC_ROM_MAX, PC_ROM_ALIGN);
 }