Ray, On 03/29/16 12:32, Ruiyu Ni wrote: > The existing code already supports HTTP boot to RAM Disk. > The patch 1/3 is to follow the ACPI spec to use reserved memory instead > of BS memory. > Patch 2/3 is to fix the memory leak issue when Http Boot finishes or fails. > Patch 3/3 is to update memory type information check algorithm to support > Ram Disk case. > > Ruiyu Ni (3): > MdeModulePkg/Bds: Allocate reserved memory for RAM Disk boot media > MdeModulePkg/Bds: Free resources after ram disk boot finishes > MdeModulePkg/Bds: Memory Bins don't count the memory used by RAM Disk > > MdeModulePkg/Library/UefiBootManagerLib/BmBoot.c | 303 > +++++++++++++++------ > MdeModulePkg/Library/UefiBootManagerLib/BmMisc.c | 19 +- > .../Library/UefiBootManagerLib/InternalBm.h | 48 ++-- > .../UefiBootManagerLib/UefiBootManagerLib.inf | 1 + > 4 files changed, 270 insertions(+), 101 deletions(-) >
do you think the same problems exist with RAM Disks if a platform still uses the IntelFrameworkModulePkg BDS? I'm asking because Paulo Alcantara (CC'd) recently submitted a patch for OVMF that would enable RAM Disks in OVMF: http://thread.gmane.org/gmane.comp.bios.edk2.devel/9476 Paulo's initial idea was to control the inclusion of RamDiskDxe thru a new build flag called RAMDISK_SUPPORT_ENABLE. I suggested to include the driver unconditionally instead, seeing no downsides to it. However, if the combination of RamDiskDxe *plus* IntelFrameworkModulePkg/BdsDxe leads to errors -- see especially patch #1 in your series, which fixes an ACPI spec violation --, then I'll change my mind. In that case RamDiskDxe should be included in OVMF conditionally at best. (And I think the flag should be called simply RAMDISK_ENABLE then.) Thanks Laszlo _______________________________________________ edk2-devel mailing list edk2-devel@lists.01.org https://lists.01.org/mailman/listinfo/edk2-devel
