SeaBIOS August 2016

seabios@seabios.org

14 participants
25 discussions

[PATCH v3 0/3] support booting more than 255 CPUs with QEMU
by Igor Mammedov 11 Aug '16

11 Aug '16

Changelog since: v2: * rebase on top of current master /smp_scan() changes/ v1: * s/count_cpu/apic_id_init/ * merge handle_x2apic() into apic_id_init() RFC: * move out max-cpus check out of mptable_setup() * factor out CPU counting/apic ID detection in separate function * return back accidentially deleted debug message with APIC ID * drop unused code in smp_setup() According to SDM, if CPUs have APIC ID more than 254 firmware should pass control to OS in x2APIC mode. This series adds x2APIC bootstrap initialization. Rebased QEMU side of x2APIC support for the next merge window: https://www.mail-archive.com/qemu-devel@nongnu.org/msg390671.html Igor Mammedov (2): paravirt: disable legacy bios tables in case of more than 255 CPUs support booting with more than 255 CPUs Kevin O'Connor (1): smp: refactor present CPU APIC ID detection and counting src/x86.h | 1 + src/fw/paravirt.c | 6 ++++-- src/fw/smp.c | 52 +++++++++++++++++++++++++++++++++++++--------------- 3 files changed, 42 insertions(+), 17 deletions(-) -- 2.7.4

4 12

Re: [SeaBIOS] [QUESTION]stuck in SeaBIOS and vm cannot be reset any more
by Xulei (Stone) 11 Aug '16

11 Aug '16

> On 09/08/2016 10:04, Xulei (Stone) wrote: > > Following your suggestion, i'm now sure it is caused by missing SMI. > > I have tried adding dprintf() like this: > > > > --- a/roms/seabios/src/fw/smm.c > > +++ b/roms/seabios/src/fw/smm.c > > @@ -65,7 +65,8 @@ handle_smi(u16 cs) > > u8 cmd = inb(PORT_SMI_CMD); > > struct smm_layout *smm = MAKE_FLATPTR(cs, 0); > > u32 rev = smm->cpu.i32.smm_rev & SMM_REV_MASK; > > - dprintf(DEBUG_HDL_smi, "handle_smi cmd=%x smbase=%p\n", cmd, > smm); > > + if(cmd == 0x00) { > > + dprintf(1, "handle_smi cmd=%x smbase=%p\n", cmd, smm); > > + } > > > > if (smm == (void*)BUILD_SMM_INIT_ADDR) { > > // relocate SMBASE to 0xa0000 @@ -147,14 +148,14 @@ > > smm_relocate_and_restore(void) { > > /* init APM status port */ > > outb(0x01, PORT_SMI_STATUS); > > + dprintf(1,"before SMI====\n"); > > > > /* raise an SMI interrupt */ > > outb(0x00, PORT_SMI_CMD); > > + dprintf(1,"after SMI=====\n"); > > > > /* wait until SMM code executed */ > > while (inb(PORT_SMI_STATUS) != 0x00) > > ; > > + dprintf(1,"smm code executes complete====\n"); > > > > And the failed case log output like this: > > 2016-08-03 16:23:15PCI: Using 00:02.0 for primary VGA > > 2016-08-03 16:23:15smm_device_setup start > > 2016-08-03 16:23:15init smm > > 2016-08-03 16:23:15before SMI==== > > 2016-08-03 16:23:15after SMI===== > > > > So, it's obviously that after outb(0x01, PORT_SMI_STATUS), bios does > > not handle_smi, so PORT_SMI_STATUS is always 0x01. What's more, when > > this problem happens, rebooting vm cannot restore it any more. My vm > > is always stuck at the same place until i destroy it. > > > > And I have already tried kernel commit c43203cab1e which still can not > > solve this problem. > > Any idea, Kevin and Paolo? > > 0xb2 is handled within QEMU, so it may be useful to make sure that QEMU is > sending the KVM_SMI ioctl. From there the best tool is still KVM tracing and > printk. I suggest replacing the dprintf with a simple outb like outb(0x21, > 0x402) (before) and outb(0x23, 0x402) (after). They show as "!" and "#" in the > trace, and they are easy to spot in the KVM trace. Following your suggestion, I found this problem may be caused by the flag of HF_SMM_MASK. I'm now sure QEMU is sending the KVM_SMI ioctl, and kmod already handles this ioctl. I add printk in inject_pending_event(), like this: /* try to inject new event if pending */ + if(get_smi) + printk(KERN_INFO "is_smm:%d\n", is_smm(vcpu)); + if (vcpu->arch.smi_pending && !is_smm(vcpu)) { vcpu->arch.smi_pending = false; process_smi(vcpu); } Then I found that the normal output is (is_smm is 0): 2016-08-11T09:44:45.090078+08:00|info|kernel[-]|[1269634.151054get smi ioctl from qemu 2016-08-11T09:44:45.090097+08:00|info|kernel[-]|[1269634.151056in process_smi_request 2016-08-11T09:44:45.090114+08:00|info|kernel[-]|[1269634.151057is_smm:0 2016-08-11T09:44:45.090131+08:00|info|kernel[-]|[1269634.151058process smi 2016-08-11T09:44:45.090914+08:00|info|kernel[-]|[1269634.151575get smi ioctl from qemu 2016-08-11T09:44:45.090947+08:00|info|kernel[-]|[1269634.151578in process_smi_request 2016-08-11T09:44:45.090972+08:00|info|kernel[-]|[1269634.151579is_smm:0 When problem occurs the output is (is_smm is 1): 2016-08-11T10:07:11.755982+08:00|info|kernel[-]|[1270981.916056get smi ioctl from qemu 2016-08-11T10:07:11.756035+08:00|info|kernel[-]|[1270981.916062in process_smi_request 2016-08-11T10:07:11.756078+08:00|info|kernel[-]|[1270981.916064is_smm:1 2016-08-11T10:07:13.896977+08:00|info|kernel[-]|[1270984.058117get smi ioctl from qemu 2016-08-11T10:07:13.897063+08:00|info|kernel[-]|[1270984.058121in process_smi_request 2016-08-11T10:07:13.897091+08:00|info|kernel[-]|[1270984.058123is_smm:1 So, it seems kvm cannot set HF_SMM_MASK flag correctly when we continuously reset VM. Any futher suggestion? Paolo :) > Paolo

1 0

[PATCH] checkstack: Handle conditional checks at start of functions
by Kevin O'Connor 10 Aug '16

10 Aug '16

Recent versions of gcc will sometimes place conditional checks in the code prior to setting up the function's stack frame. Handle this case correctly. Signed-off-by: Kevin O'Connor <kevin(a)koconnor.net> --- scripts/checkstack.py | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/scripts/checkstack.py b/scripts/checkstack.py index 5d9b0bf..255768a 100755 --- a/scripts/checkstack.py +++ b/scripts/checkstack.py @@ -153,6 +153,11 @@ def main(): if '%esp' in insn or insn.startswith('leal'): # Still part of initial header continue + if not stackusage and ( + insn.startswith('test') or insn.startswith('cmp') + or insn.startswith('j')): + # There may be conditional checks prior to stack frame + continue cur.basic_stack_usage = stackusage atstart = 0 -- 2.5.5

1 0

Re: [SeaBIOS] enable hvm in SeaBios
by Kevin O'Connor 10 Aug '16

10 Aug '16

On Wed, Aug 10, 2016 at 12:24:25PM +0000, pieter van puymbroeck wrote: > Hello, > > Currently I'm building a new homelab based on omnios. > > There a kvm enabled zone is created and inside the zone I have 3 kvm-guests running. These guests are running oracle vm (xen based). All is running fine except if I want to start an vm inside oracle vm, the system complains about: > > > Error: HVM guest support is unavailable: is VT/AMD-V supported by your CPU and enabled in your BIOS? > [...] > So it looks to me that the bios the kvm is using ( which appears to be SeaBIOS (version 0.6.1.2-20110201_165504-titi) ) is not enabling the nested virtualisation by default. Is this correct? > > > How can I fix this? > SeaBIOS doesn't control this. Instead, this is something that QEMU/KVM does the setup for. You can look through the QEMU documentation for the appropriate flags (and minimum software versions) or ask on the QEMU/KVM mailing lists. -Kevin

1 0

enable hvm in SeaBios
by pieter van puymbroeck 10 Aug '16

10 Aug '16

Hello, Currently I'm building a new homelab based on omnios. There a kvm enabled zone is created and inside the zone I have 3 kvm-guests running. These guests are running oracle vm (xen based). All is running fine except if I want to start an vm inside oracle vm, the system complains about: Error: HVM guest support is unavailable: is VT/AMD-V supported by your CPU and enabled in your BIOS? Inside the kvm enabled zone, kvm is enabled: root@ovm1:/root# ls -l /dev/kvm crw------- 1 root sys 125, 0 Aug 10 11:26 /dev/kvm root@ovm1:/root# root@ovm1:/root# modinfo |grep -i kvm 224 fffffffff8408000 3a030 125 1 kvm (kvm driver v0.1) root@ovm1:/root# For sure I checked the processor itself from the global zone if it is supported: root@sea:/root# isainfo -v 64-bit amd64 applications vmx pclmulqdq aes sse4.2 sse4.1 ssse3 popcnt tscp cx16 sse3 sse2 sse fxsr mmx cmov amd_sysc cx8 tsc fpu 32-bit i386 applications vmx pclmulqdq aes sse4.2 sse4.1 ssse3 popcnt tscp ahf cx16 sse3 sse2 sse fxsr mmx cmov sep cx8 tsc fpu root@sea:/root# and to me it looks I have the required flags available. inside the lvm-guest the flags are indeed not given to the guest: [root@ovm2 ~]# uname -s -r -i -p Linux 2.6.39-400.279.1.el5uek x86_64 x86_64 [root@ovm2 ~]# cat /etc/oracle-release Oracle VM Server release 5.7 [root@ovm2 ~]# grep flags /proc/cpuinfo flags : fpu de tsc msr pae mce cx8 apic sep mca cmov pat clflush mmx fxsr sse sse2 ss syscall nx lm constant_tsc up nopl pni ssse3 cx16 sse4_1 sse4_2 popcnt aes hypervisor lahf_lm [root@ovm2 ~]# it does not matter if I start the kvm guest using -cpu host \ or using -cpu qemu64,+vmx,+aes,+sse4.2,+sse4.1,+ssse3 \ # --> this one is faster So it looks to me that the bios the kvm is using ( which appears to be SeaBIOS (version 0.6.1.2-20110201_165504-titi) ) is not enabling the nested virtualisation by default. Is this correct? How can I fix this? Thanks and best regards, Pieter

1 0

[PATCH 1/2] pvscsi: Fix incorrect arguments order in call to memalign_low
by Dana Rubin 09 Aug '16

09 Aug '16

From: Dana Rubin <dana.rubin(a)ravellosystems.com> First argument should be 'align' and second 'size'. Signed-off-by: <dana.rubin(a)ravellosystems.com> Signed-off-by: <shmulik.ladkani(a)ravellosystems.com> --- src/hw/pvscsi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/hw/pvscsi.c b/src/hw/pvscsi.c index 05cbd91..7535493 100644 --- a/src/hw/pvscsi.c +++ b/src/hw/pvscsi.c @@ -167,7 +167,7 @@ pvscsi_init_rings(void *iobase, struct pvscsi_ring_dsc_s **ring_dsc) { struct PVSCSICmdDescSetupRings cmd = {0,}; - struct pvscsi_ring_dsc_s *dsc = memalign_low(sizeof(*dsc), PAGE_SIZE); + struct pvscsi_ring_dsc_s *dsc = memalign_low(PAGE_SIZE, sizeof(*dsc)); if (!dsc) { warn_noalloc(); return; -- 1.9.1

2 2

[PATCH 0/5] Simplify SeaVGABIOS headers
by Kevin O'Connor 09 Aug '16

09 Aug '16

This series reorganizes the vgabios headers so that there is just one header for each C file, except for vgautil.h which is used for C files that only need simple function and variable extern definitions. -Kevin Kevin O'Connor (5): vgafb: Move header definitions from vgabios.h to new file vgafb.h vgainit: Move video param setup to stdvga_build_video_param() vgautil: Add new header file with misc function and variable definitions vgautil: Move generic definitions from stdvga.h to vgautil.h vgautil: Move definitions from cbvga.h and clext.h to vgautil.h vgasrc/bochsvga.c | 5 +-- vgasrc/cbvga.c | 5 +-- vgasrc/cbvga.h | 20 ------------ vgasrc/clext.c | 4 +-- vgasrc/clext.h | 20 ------------ vgasrc/geodevga.c | 3 +- vgasrc/stdvga.c | 1 + vgasrc/stdvga.h | 32 +------------------ vgasrc/stdvgaio.c | 3 +- vgasrc/stdvgamodes.c | 13 +++++++- vgasrc/swcursor.c | 4 ++- vgasrc/vbe.c | 8 ++++- vgasrc/vgabios.c | 4 ++- vgasrc/vgabios.h | 72 +++-------------------------------------- vgasrc/vgafb.c | 4 ++- vgasrc/vgafb.h | 42 ++++++++++++++++++++++++ vgasrc/vgafonts.c | 2 +- vgasrc/vgahw.h | 3 +- vgasrc/vgainit.c | 11 ++----- vgasrc/vgautil.h | 90 ++++++++++++++++++++++++++++++++++++++++++++++++++++ 20 files changed, 182 insertions(+), 164 deletions(-) delete mode 100644 vgasrc/cbvga.h delete mode 100644 vgasrc/clext.h create mode 100644 vgasrc/vgafb.h create mode 100644 vgasrc/vgautil.h -- 2.5.5

1 6

[PATCH] blockcmd: CMD_SCSI op is only used in 32bit mode
by Kevin O'Connor 09 Aug '16

09 Aug '16

Reduce the size of the 16bit code slightly by recognizing that CMD_SCSI is only used in 32bit mode. Signed-off-by: Kevin O'Connor <kevin(a)koconnor.net> --- src/hw/blockcmd.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/src/hw/blockcmd.c b/src/hw/blockcmd.c index 093c5d7..f147100 100644 --- a/src/hw/blockcmd.c +++ b/src/hw/blockcmd.c @@ -118,6 +118,8 @@ scsi_fill_cmd(struct disk_op_s *op, void *cdbcmd, int maxcdb) cmd->count = cpu_to_be16(op->count); return GET_GLOBALFLAT(op->drive_gf->blksize); case CMD_SCSI: + if (MODESEGMENT) + return -1; memcpy(cdbcmd, op->cdbcmd, maxcdb); return op->blocksize; default: @@ -129,13 +131,15 @@ scsi_fill_cmd(struct disk_op_s *op, void *cdbcmd, int maxcdb) int scsi_is_read(struct disk_op_s *op) { - return op->command == CMD_READ || (op->command == CMD_SCSI && op->blocksize); + return op->command == CMD_READ || ( + !MODESEGMENT && op->command == CMD_SCSI && op->blocksize); } // Check if a SCSI device is ready to receive commands int scsi_is_ready(struct disk_op_s *op) { + ASSERT32FLAT(); dprintf(6, "scsi_is_ready (drive=%p)\n", op->drive_gf); /* Retry TEST UNIT READY for 5 seconds unless MEDIUM NOT PRESENT is @@ -181,6 +185,7 @@ scsi_is_ready(struct disk_op_s *op) int scsi_drive_setup(struct drive_s *drive, const char *s, int prio) { + ASSERT32FLAT(); struct disk_op_s dop; memset(&dop, 0, sizeof(dop)); dop.drive_gf = drive; -- 2.5.5

1 1

Re: [SeaBIOS] [QUESTION]stuck in SeaBIOS and vm cannot be reset any more
by Paolo Bonzini 09 Aug '16

09 Aug '16

On 09/08/2016 10:04, Xulei (Stone) wrote: > Following your suggestion, i'm now sure it is caused by missing SMI. > I have tried adding dprintf() like this: > > --- a/roms/seabios/src/fw/smm.c > +++ b/roms/seabios/src/fw/smm.c > @@ -65,7 +65,8 @@ handle_smi(u16 cs) > u8 cmd = inb(PORT_SMI_CMD); > struct smm_layout *smm = MAKE_FLATPTR(cs, 0); > u32 rev = smm->cpu.i32.smm_rev & SMM_REV_MASK; > - dprintf(DEBUG_HDL_smi, "handle_smi cmd=%x smbase=%p\n", cmd, smm); > + if(cmd == 0x00) { > + dprintf(1, "handle_smi cmd=%x smbase=%p\n", cmd, smm); > + } > > if (smm == (void*)BUILD_SMM_INIT_ADDR) { > // relocate SMBASE to 0xa0000 > @@ -147,14 +148,14 @@ smm_relocate_and_restore(void) { > /* init APM status port */ > outb(0x01, PORT_SMI_STATUS); > + dprintf(1,"before SMI====\n"); > > /* raise an SMI interrupt */ > outb(0x00, PORT_SMI_CMD); > + dprintf(1,"after SMI=====\n"); > > /* wait until SMM code executed */ > while (inb(PORT_SMI_STATUS) != 0x00) > ; > + dprintf(1,"smm code executes complete====\n"); > > And the failed case log output like this: > 2016-08-03 16:23:15PCI: Using 00:02.0 for primary VGA > 2016-08-03 16:23:15smm_device_setup start > 2016-08-03 16:23:15init smm > 2016-08-03 16:23:15before SMI==== > 2016-08-03 16:23:15after SMI===== > > So, it's obviously that after outb(0x01, PORT_SMI_STATUS), bios does not > handle_smi, so PORT_SMI_STATUS is always 0x01. What's more, when this > problem happens, rebooting vm cannot restore it any more. My vm is always > stuck at the same place until i destroy it. > > And I have already tried kernel commit c43203cab1e which still can not > solve this problem. > Any idea, Kevin and Paolo? 0xb2 is handled within QEMU, so it may be useful to make sure that QEMU is sending the KVM_SMI ioctl. From there the best tool is still KVM tracing and printk. I suggest replacing the dprintf with a simple outb like outb(0x21, 0x402) (before) and outb(0x23, 0x402) (after). They show as "!" and "#" in the trace, and they are easy to spot in the KVM trace. Paolo

1 0

Re: [SeaBIOS] [QUESTION]stuck in SeaBIOS and vm cannot be reset any more
by Xulei (Stone) 09 Aug '16

09 Aug '16

>On Tue, Aug 02, 2016 at 04:18:30AM +0000, Xulei (Stone) wrote: >> >On Fri, Jul 29, 2016 at 04:04:59AM +0000, Xulei (Stone) wrote: >> >> After one day, the vm is stuck. Looking from the following seabios >> >> log, it seems seabios stops at "PCI: Using 00:02.0 for primary >> >> VGA", and can not execute handle_smp() any more. >> >> What may be the reason? >> > >> >More debugging info would be necessary to find this problem. You >> >could try reproducing and attaching gdb ( >> >http://www.seabios.org/Debugging#Debugging_with_gdb_on_QEMU ). >> >Alternatively, a kvm trace log may help. >> > >> kvm trace (seems useful) indicates that cpu 0 keeps always to access 0x00b3 ioport. >> 0x00b3 is PORT_SMI_STATUS, so i guess my bios is stuck in the >> function smm_relocate_and_restore { >> ... >> /* wait until SMM code executed */ >> while (inb(PORT_SMI_STATUS) != 0x00) >> ... >> } > >I'd try adding dprintf() statements around all the code at the top of >smm_relocate_and_restore() and enable the dprintf() at the top of >handle_smi(). > >It would also be useful if you can extract the log from the last two >working reboots to compare it to the failed case. Following your suggestion, i'm now sure it is caused by missing SMI. I have tried adding dprintf() like this: --- a/roms/seabios/src/fw/smm.c +++ b/roms/seabios/src/fw/smm.c @@ -65,7 +65,8 @@ handle_smi(u16 cs) u8 cmd = inb(PORT_SMI_CMD); struct smm_layout *smm = MAKE_FLATPTR(cs, 0); u32 rev = smm->cpu.i32.smm_rev & SMM_REV_MASK; - dprintf(DEBUG_HDL_smi, "handle_smi cmd=%x smbase=%p\n", cmd, smm); + if(cmd == 0x00) { + dprintf(1, "handle_smi cmd=%x smbase=%p\n", cmd, smm); + } if (smm == (void*)BUILD_SMM_INIT_ADDR) { // relocate SMBASE to 0xa0000 @@ -147,14 +148,14 @@ smm_relocate_and_restore(void) { /* init APM status port */ outb(0x01, PORT_SMI_STATUS); + dprintf(1,"before SMI====\n"); /* raise an SMI interrupt */ outb(0x00, PORT_SMI_CMD); + dprintf(1,"after SMI=====\n"); /* wait until SMM code executed */ while (inb(PORT_SMI_STATUS) != 0x00) ; + dprintf(1,"smm code executes complete====\n"); And the failed case log output like this: 2016-08-03 16:23:15PCI: Using 00:02.0 for primary VGA 2016-08-03 16:23:15smm_device_setup start 2016-08-03 16:23:15init smm 2016-08-03 16:23:15before SMI==== 2016-08-03 16:23:15after SMI===== So, it's obviously that after outb(0x01, PORT_SMI_STATUS), bios does not handle_smi, so PORT_SMI_STATUS is always 0x01. What's more, when this problem happens, rebooting vm cannot restore it any more. My vm is always stuck at the same place until i destroy it. And I have already tried kernel commit c43203cab1e which still can not solve this problem. Any idea, Kevin and Paolo? > > > >-Kevin

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS August 2016