SeaBIOS May 2012

seabios@seabios.org

23 participants
48 discussions

[PATCHv2 0/9] Update low-memory allocations - avoid using 9-segment.
by Kevin O'Connor 20 May '12

20 May '12

This patch series updates the SeaBIOS' low-memory allocation scheme so that it no longer uses the EBDA or the 9-segment for internal low memory allocations. Instead, the e-segment is used where SeaBIOS requires read/writable "low" memory. There is now plenty of space in the e-segment (there has been ever since SeaBIOS was updated to relocate its initialization code), while using 9-segment space can reserve the memory from OS use (in particular on older operating systems). Also, using the 9-segment (and indirectly the EBDA) is more fragile as it requires handling of a relocatable EBDA. As a result of this change, SeaBIOS will be able to allocate larger amounts of low-memory. Indeed, the size of the "extra stack" for 16bit code is increased from 512 bytes to 2K in this series. Changes in v2: * Fixed ehci error resulting from changes to stack alignment * Introduce new GET/SET_LOW and GET/SET_LOWFLAT macros (which mirror GET_GLOBAL and GET_GLOBALFLAT) which produce better 16bit code. -Kevin Kevin O'Connor (9): Use the e-segment instead of the 9-segment for bios "low mem". Add mechanism to declare variables as "low mem" and use for extra stack. Convert timer code EBDA variables to VARLOW variables. Convert boot code EBDA variables to VARLOW variables. Convert ps2 code EBDA variables to VARLOW variables. Convert USB keyboard code EBDA variables to VARLOW variables. Convert disk code EBDA variables to VARLOW variables. EBDA cleanups. Convert GET/SET_FLATPTR() accesses to "low mem" to GET/SET_LOWFLAT(). src/acpi.c | 2 +- src/ahci.c | 2 +- src/asm-offsets.c | 8 -- src/ata.c | 7 +- src/biosvar.h | 100 ++++------------ src/block.c | 5 +- src/blockcmd.h | 2 + src/boot.c | 19 ++-- src/bootsplash.c | 1 - src/cdrom.c | 84 +++++++------- src/clock.c | 32 +++--- src/config.h | 2 + src/coreboot.c | 3 +- src/disk.c | 75 +++++------- src/disk.h | 38 ++++++- src/memmap.c | 2 +- src/mouse.c | 31 +++--- src/mtrr.c | 2 +- src/optionroms.c | 14 +-- src/pcibios.c | 2 +- src/pciinit.c | 3 +- src/pirtable.c | 2 +- src/pmm.c | 116 ++++------------- src/post.c | 50 ++++++-- src/ps2port.c | 19 ++-- src/romlayout.S | 8 +- src/shadow.c | 4 +- src/smbios.c | 1 - src/stacks.c | 15 ++- src/types.h | 4 + src/usb-ehci.c | 40 +++--- src/usb-hid.c | 17 ++- src/usb-ohci.c | 26 ++-- src/usb-uhci.c | 43 +++---- src/util.h | 3 +- src/xen.h | 3 +- tools/layoutrom.py | 349 +++++++++++++++++++++++++++++----------------------- 37 files changed, 551 insertions(+), 583 deletions(-) -- 1.7.6.5

1 9

Re: [SeaBIOS] [PATCH 1/6] remove iasl-generated hex files from git repo
by Kevin O'Connor 18 May '12

18 May '12

On Wed, May 16, 2012 at 02:53:24PM -0600, Marc Jones wrote: > I think that this is a good idea. I think that I am hitting a problem > with incompatible .hex and iasl versions. > I am having problem with commit 15952a : Revert "Fix 64bit PCI issues > on Windows" and I am using the coreboot crosstools iasl. > > > Intel ACPI Component Architecture > ASL Optimizing Compiler version 20110922-64 [Mar 2 2012] > Copyright (c) 2000 - 2011 Intel Corporation > > ASL Input: /home/marc/devel/coreboot/build/seabios/out/ssdt-proc.dsl.i > - 36 lines, 765 bytes, 8 keywords > AML Output: /home/marc/devel/coreboot/build/seabios/out/ssdt-proc.aml > - 120 bytes, 6 named objects, 2 executable opcodes > Listing File: /home/marc/devel/coreboot/build/seabios/out/ssdt-proc.lst > - 2956 bytes > Hex Dump: /home/marc/devel/coreboot/build/seabios/out/ssdt-proc.hex > - 1492 bytes > > Compilation complete. 0 Errors, 0 Warnings, 0 Remarks, 0 Optimizations > /home/marc/devel/coreboot/build/seabios/out/ssdt-pcihp.dsl.i 21: > Method (_EJ0, 1) { Return(PCEJ(0x01)) } Name (_SUN, 0x01) } > Warning 1099 - Reserved > method should not return a value ^ (_EJ0) > > /home/marc/devel/coreboot/build/seabios/out/ssdt-pcihp.dsl.i 26: > Method (_EJ0, 1) { Return(PCEJ(0x02)) } Name (_SUN, 0x02) } > Warning 1099 - Reserved > method should not return a value ^ (_EJ0) > > ... > > The build passes with my distro toolchain version: ASL Optimizing > Compiler version 20100528 [Oct 15 2010] > > It seems that this might be a problem with the .hex files being > generated by a different version of IASL. Just building the .hex file > seems to fix this. The hex file is just C code that defines an array, so it can't be the cause of any iasl errors. I'm not sure what would cause these errors, but can you confirm the error is still there after a "make clean". If so, you can email me a tar of your tree (including the out/ directory) and I'll try to take a look. -Kevin

1 0

[PATCH 1/2] Add ACPI_EXTRACT_PKG_START macro parsing
by Gleb Natapov 17 May '12

17 May '12

It allows to extract the beginning of a Package object content. Signed-off-by: Gleb Natapov <gleb(a)redhat.com> --- tools/acpi_extract.py | 12 ++++++++++++ 1 files changed, 12 insertions(+), 0 deletions(-) diff --git a/tools/acpi_extract.py b/tools/acpi_extract.py index 5f613e4..14c1dfe 100755 --- a/tools/acpi_extract.py +++ b/tools/acpi_extract.py @@ -29,6 +29,7 @@ # ACPI_EXTRACT_PROCESSOR_START - start of Processor() block # ACPI_EXTRACT_PROCESSOR_STRING - extract a NameString from Processor() # ACPI_EXTRACT_PROCESSOR_END - offset at last byte of Processor() + 1 +# ACPI_EXTRACT_PKG_START - start of Package block # # ACPI_EXTRACT_ALL_CODE - create an array storing the generated AML bytecode # @@ -181,6 +182,15 @@ def aml_processor_end(offset): pkglen = aml_pkglen(offset) return offset + pkglen +def aml_package_start(offset): + offset = aml_name_string(offset) + 4 + # 0x12 PkgLength NumElements PackageElementList + if (aml[offset] != 0x12): + die( "Name offset 0x%x: expected 0x12 actual 0x%x" % + (offset, aml[offset])); + offset += 1 + return offset + aml_pkglen_bytes(offset) + 1 + lineno = 0 for line in fileinput.input(): # Strip trailing newline @@ -263,6 +273,8 @@ for i in range(len(asl)): offset = aml_processor_string(offset) elif (directive == "ACPI_EXTRACT_PROCESSOR_END"): offset = aml_processor_end(offset) + elif (directive == "ACPI_EXTRACT_PKG_START"): + offset = aml_package_start(offset) else: die("Unsupported directive %s" % directive) -- 1.7.7.3

3 12

[PATCH v3] pci: clean all funcs when hot-removing multifunc device
by Amos Kong 16 May '12

16 May '12

Hotplug CallTrace: int acpiphp_enable_slot(struct acpiphp_slot *slot) \_enable_device(slot); \_pci_bus_add_devices(bus); # un-added new devs(all funcs in slot) will be added list_for_each_entry(dev, &bus->devices, bus_list) { if (dev->is_added) continue; pci_bus_add_device(dev); device_add(&dev->dev); dev->is_added = 1; 'dev->is_added' is used to trace if pci dev is added to bus, all funcs in same slot would be added to bus in enable_device(slot). So we need to clean all funcs of same slot in disable_device(slot). But hot-remove exists bug: https://bugzilla.kernel.org/show_bug.cgi?id=43219 (dmesg and DSDT were attached in bz), detail: . Boot up a Linux VM with 8 pci block devices which are the 8 functions in one pci slot. | # qemu-kvm ... | -drive file=images/u0,if=none,id=drv0,format=qcow2,cache=none \ | -device virtio-blk-pci,drive=drv0,id=v0,multifunction=on,addr=0x03.0 \ | .... | -drive file=images/u7,if=none,id=drv7,format=qcow2,cache=none \ | -device virtio-blk-pci,drive=drv7,id=v7,multifunction=on,addr=0x03.7 \ . Check devices in guest. | vm)# ls /dev/vd* | vda vdb vdc vde vdf vdg vdh | vm)# lspci |grep block | 00:03.0 SCSI storage controller: Red Hat, Inc Virtio block device | ... | 00:03.7 SCSI storage controller: Red Hat, Inc Virtio block device | . Func1~7 still exist in guest after hot-removing the whole slot by qemu monitor cmd. | vm)# lspci |grep block (00:03.0 disappeared) | 00:03.1 SCSI storage controller: Red Hat, Inc Virtio block device (rev ff) | ... | 00:03.7 SCSI storage controller: Red Hat, Inc Virtio block device (rev ff) | vm)# ls /dev/vd* (vda disappeared) | vdb vdc vde vdf vdg vdh | vm)# mkfs /dev/vdb | INFO: task mkfs.ext2:1784 blocked for more than 120 seconds. We process pci slot as a whole device in pciphp spec, seabios only defines one device for a slot in ACPI DSDT table. In acpiphp_glue.c:register_slot(), only one entry is added (for func#0) into 'slot->funcs' list. When we release the whole slot, only the entry in 'slot->funcs' will be cleaned, so func#1~7 could not be cleaned from system. | drivers/pci/hotplug/acpiphp_glue.c: | static int disable_device(struct acpiphp_slot *slot) { | list_for_each_entry(func, &slot->funcs, sibling) { | pdev = pci_get_slot(slot->bridge->pci_bus, | PCI_DEVFN(slot->device, func->function)); | ..clean code.. // those code is only executed 1 time(for func#0) | __pci_remove_bus_device(pdev); | pci_dev_put(pdev); Hotpluging multifunc of guests(WinXp/Win7) is fine. --- v1 thread: http://marc.info/?t=131597601700003&r=1&w=2 Changes from v1: - rebase patch to latest linux.git - remove unnecessary multiplefunction check - rename 'i' to meaningful 'fn' - fix coding style Changes from v2: - update detail reason(calltrace) to commitlog - remove hardcode 8, find funcs in pci devlist Signed-off-by: Amos Kong <akong(a)redhat.com> --- drivers/pci/hotplug/acpiphp_glue.c | 9 +++++---- 1 files changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/pci/hotplug/acpiphp_glue.c b/drivers/pci/hotplug/acpiphp_glue.c index 806c44f..a7442d9 100644 --- a/drivers/pci/hotplug/acpiphp_glue.c +++ b/drivers/pci/hotplug/acpiphp_glue.c @@ -885,7 +885,7 @@ static void disable_bridges(struct pci_bus *bus) static int disable_device(struct acpiphp_slot *slot) { struct acpiphp_func *func; - struct pci_dev *pdev; + struct pci_dev *pdev, *tmp; struct pci_bus *bus = slot->bridge->pci_bus; /* The slot will be enabled when func 0 is added, so check @@ -902,9 +902,10 @@ static int disable_device(struct acpiphp_slot *slot) func->bridge = NULL; } - pdev = pci_get_slot(slot->bridge->pci_bus, - PCI_DEVFN(slot->device, func->function)); - if (pdev) { + list_for_each_entry_safe(pdev, tmp, &bus->devices, bus_list) { + if (PCI_SLOT(pdev->devfn) != slot->device) + continue; + pci_stop_bus_device(pdev); if (pdev->subordinate) { disable_bridges(pdev->subordinate);

4 6

Re: [SeaBIOS] [coreboot] seabios feature request : reboot VM or retry boot devices when no valid boot disk is found
by Kevin O'Connor 14 May '12

14 May '12

On Fri, May 11, 2012 at 09:07:25AM -0700, ron minnich wrote: > On Fri, May 11, 2012 at 6:28 AM, Maurits van de Lande > <M.vandeLande(a)vdl-fittings.com> wrote: > > > I would prefer to automatically reboot the VM (or retry the boot devices) > > when no bootable device is found, then when the iSCSI storage is available > > again the VM can boot normally. > > > > (when I manually sent CTRL+ALT+DEL to the VM the system booted normaly > > again) > > > It would be an interesting option. Sometimes, for hardware, you really > want it to wait with the error message on the screen for a while so > you can tell what happened. At the same time, there's nothing more > annoying than having it wait forever when you know the hardware > finally is ready for a boot ... Perhaps it would be possible to set a > reasonable optional timeout and then have it reboot or, better, just > redo the hardware scan and try again. Makes sense - something like the below should work. -Kevin diff --git a/src/boot.c b/src/boot.c index 8cc94b0..832347a 100644 --- a/src/boot.c +++ b/src/boot.c @@ -617,6 +617,25 @@ boot_rom(u32 vector) call_boot_entry(so, 0); } +// Unable to find bootable device - warn user and eventually retry. +static void +boot_fail(void) +{ + printf("No bootable device.\n"); + // Wait for 60 seconds and then reboot. + u32 end = calc_future_timer(60*1000); + for (;;) { + if (check_timer(end)) + break; + wait_irq(); + } + printf("Rebooting.\n"); + struct bregs br; + memset(&br, 0, sizeof(br)); + br.code = SEGOFF(SEG_BIOS, (u32)reset_vector); + call16big(&br); +} + // Determine next boot method and attempt a boot using it. static void do_boot(u16 seq_nr) @@ -625,10 +644,7 @@ do_boot(u16 seq_nr) panic("Boot support not compiled in.\n"); if (seq_nr >= BEVCount) { - printf("No bootable device.\n"); - // Loop with irqs enabled - this allows ctrl+alt+delete to work. - for (;;) - wait_irq(); + boot_fail(); } // Boot the given BEV type.

4 8

[PATCH 0/8] Update low-memory allocations - avoid using 9-segment.
by Kevin O'Connor 14 May '12

14 May '12

This patch series updates the SeaBIOS' low-memory allocation scheme so that it no longer uses the EBDA or the 9-segment for internal low memory allocations. Instead, the e-segment is used where SeaBIOS requires read/writable "low" memory. There is now plenty of space in the e-segment (there has been ever since SeaBIOS was updated to relocate its initialization code), while using 9-segment space can reserve the memory from OS use (in particular on older operating systems). Also, using the 9-segment (and indirectly the EBDA) is more fragile as it requires handling of a relocatable EBDA. As a result of this change, SeaBIOS will be able to allocate larger amounts of low-memory. Indeed, the size of the "extra stack" for 16bit code is increased from 512 bytes to 2K in this series. -Kevin Kevin O'Connor (8): Use the e-segment instead of the 9-segment for bios "low mem". Add mechanism to declare variables as "low mem" and use for extra stack. Convert timer code EBDA variables to VARLOW variables. Convert boot code EBDA variables to VARLOW variables. Convert ps2 code EBDA variables to VARLOW variables. Convert USB keyboard code EBDA variables to VARLOW variables. Convert disk code EBDA variables to VARLOW variables. EBDA cleanups. src/acpi.c | 2 +- src/ahci.c | 2 +- src/asm-offsets.c | 8 ---- src/ata.c | 6 +-- src/biosvar.h | 80 +---------------------------------- src/block.c | 5 +- src/blockcmd.h | 2 + src/boot.c | 19 ++++----- src/bootsplash.c | 1 - src/cdrom.c | 80 ++++++++++++++++++------------------ src/clock.c | 32 ++++++++------- src/config.h | 2 + src/coreboot.c | 3 +- src/disk.c | 73 ++++++++++++++------------------ src/disk.h | 38 ++++++++++++++++- src/memmap.c | 2 +- src/mouse.c | 31 +++++++------- src/mtrr.c | 2 +- src/optionroms.c | 14 +----- src/pcibios.c | 2 +- src/pciinit.c | 3 +- src/pirtable.c | 2 +- src/pmm.c | 116 ++++++++++++---------------------------------------- src/post.c | 50 ++++++++++++++++------ src/ps2port.c | 18 ++++---- src/romlayout.S | 9 ++-- src/shadow.c | 4 +- src/smbios.c | 1 - src/stacks.c | 15 ++++--- src/types.h | 4 ++ src/usb-hid.c | 17 ++++++- src/util.h | 3 +- src/xen.h | 3 +- tools/layoutrom.py | 81 ++++++++++++++++++++++++------------ 34 files changed, 337 insertions(+), 393 deletions(-) -- 1.7.6.5

1 8

[Patch 0/12] Pciinit redesign and 64bit PCI support
by Alexey Korolev 12 May '12

12 May '12

This patch series redesigns and adds new features to pciinit.c code. [Patches 1-4] There are no more arrays of bases and counts in new implementation. The new implementation is based on dynamic allocation of pci_region_entry structures which are grouped into lists. Each pci_region_entry structure could be a PCI bar or a downstream PCI region (bridge). Each entry has a set of attributes: type (IO, MEM,PREFMEM), is64bit, size. [Patch 5] Bridge regions are no longer rounded up to the next highest size - instead track alignment explicitly. This should improve the memory layout for bridge regions. Also, unused bridge regions will no longer be allocated any space. [Patch 6] Patch takes into account PCI bar and ROM regions of PCI bridges [Patches 7-12] Make pciinit.c 64bit aware. Support discovery and configuration of 64bit bars, with non-zero upper32 bits. Code allocates the 64bit PCI bars in high address range if they don't fir in 32bit range. Kevin O'Konor (6): 0001-pciinit-Introduction-of-pci_region_entry-structure.patch 0002-pciinit-Move-bus-bar-asignment.patch 0004-pciinit-Use-sorted-order-allocation.patch 0005-pciinit-Track-region-alignment-explicitly.patch 0006-pciinit-bridges-can-have-two-regions-too.patch 0007-pciinit-Switch-to-64bit-variable-types.patch Alexey Korolev (6): 0003-pciinit-Remove-size-element-from-pci_bus-r-structure.patch 0008-pciinit-Add-pci_region-structure.patch 0009-pciinit-64bit-capability-discovery-for-pci-bridges.patch 0010-Do-not-store-pci-region-stats-instead-calulate-the-s.patch 0011-Migrate-64bit-entries-to-64bit-pci-regions.patch 0012-Fix-64bit-PCI-issues-on-Windows.patch src/acpi-dsdt.dsl | 7 + src/acpi-dsdt.hex | 64 +++++++- src/config.h | 2 + src/pci.h | 6 +- src/pciinit.c | 464 ++++++++++++++++++++++++++++++----------------------- 5 files changed, 325 insertions(+), 218 deletions(-)

3 29

[seabios patch 0/5] dynamic pci i/o windows
by Gerd Hoffmann 12 May '12

12 May '12

Hi, This patch series makes the PCI I/O windows runtime-configurable via qemu firmware config interface. Main advantage is that we can size and shuffle around the PCI i/O windows according to the amount of memory the virtual machine has. We don't need a hole for 64bit PCI bars, we can just map them above the main memory. The hole for 32bit PCI bars can be enlarged for guests with less than 3.5 GB of memory. Oh, and the pci device initialization fix is there too ;) cheers, Gerd Gerd Hoffmann (5): pci: init all devices acpi: add qemu fwcfg driver acpi: update pci io windows according to fw_cfg info pciinit: make pci ressources configurable update src/acpi-dsdt.hex src/acpi-dsdt.dsl | 81 ++++++++++- src/acpi-dsdt.hex | 420 +++++++++++++++++++++++++++++++++++++++++++++++++++-- src/paravirt.c | 8 + src/paravirt.h | 2 + src/pciinit.c | 32 +++- 5 files changed, 522 insertions(+), 21 deletions(-)

5 18

Re: [SeaBIOS] patches from Google
by Kevin O'Connor 12 May '12

12 May '12

On Fri, May 11, 2012 at 09:09:22AM -0700, ron minnich wrote: > Let me know if this format is ok, this is the first in a series :-) The format is okay. It's preferable if you can mail the output of 'git format-patch' directly, but it's also okay as an attachment. I use the following to send the mails via mutt: mutt -e 'set abort_unmodified=no' -H 0001-my-patch.patch The patch itself, though, has some errors in it. [...] --- a/tools/test-gcc.sh +++ b/tools/test-gcc.sh [...] @@ -69,9 +74,7 @@ EOF $CC -Os -c -fwhole-program $TMPFILE2 -o $TMPFILE2o > /dev/null 2>&1 $CC -nostdlib -Os $TMPFILE1o $TMPFILE2o -o $TMPFILE3o > /dev/null 2>&1 if [ $? -ne 0 ]; then - echo " Working around non-functional -fwhole-program" >&2 - echo 2 - exit 0 + report_error " Working around non-functional -fwhole-program" That removes the exit and changes the behavior of the script. [...] @@ -104,9 +107,7 @@ $CC -O -g -fwhole-program -combine -c $TMPFILE1 $TMPFILE2 -o $TMPFILE1o > /dev/n if [ $? -eq 0 ]; then echo 0 else - echo " Working around non-functional -combine" >&2 - echo 1 -fi + report_error " Working around non-functional -combine" This removes the trailing "fi" and will break the script. Also, are you sure the "report_error" stuff will work in a non-bash environment? A number of reports have been made about bash-isms in the build - I don't want to add any back in. -Kevin

2 1

Re: [SeaBIOS] Website broken
by Kevin O'Connor 12 May '12

12 May '12

On Fri, May 11, 2012 at 01:10:26PM +0200, Fred . wrote: > Database error > Jump to: navigation, search > A database query syntax error has occurred. This may indicate a bug in > the software. The last attempted database query was: > > (SQL query hidden) > > from within function "Revision::fetchFromConds". Database returned > error "1: no such column: rev_sha1". Thanks for reporting. It should be fixed now. -Kevin

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS May 2012