SeaBIOS

seabios@seabios.org

2 participants
3801 discussions

[PATCH seabios 0/3] add kvmtool support
by Gerd Hoffmann 06 Nov '17

06 Nov '17

Hi, The 1.11 seabios release which is just around the corner brings serial console support to seabios. Which is useful for kvmtool too. So I decided to undust my old kvmtool support patch series and check out where we stand. Patch #1 adds the config option, wires up everything and adds ram detection for kvmtool. Patches #2 and #3 tweak the seabios virtio drivers so they work with kvmtool. With that in place seabios works nicely. Loads grub from hard disk, allows to edit menu entries, loads and runs the linux kernel. But there are also some known issues: (1) ram detection is clumpsy (see patch #1) due to kvmtool not passing ram information to the firmware (on x86, seems other platforms use the device tree for that). (2) kvmtool virtio implementation seems to have some bugs. specifically handing over the virtio-blk devices from seabios to the linux kernel doesn't work (see patch #3). (3) when kernels are booted via seabios+grub all the command line args kvmtool adds to the kernel command line on direct kernel boot will not be passed to the kernel. Not sure how much of a problem that is in practice, due to (2) I didn't yet manage to boot a linux guest to the login prompt. cheers, Gerd Gerd Hoffmann (3): kvmtool: initial support kvmtool: allow mmio for legacy bar 0 kvmtool: support larger virtio queues Makefile | 1 + src/fw/paravirt.h | 3 +++ src/hw/virtio-ring.h | 2 +- src/fw/paravirt.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++++ src/hw/virtio-pci.c | 16 ++++++++++++---- src/post.c | 4 +++- src/sercon.c | 2 ++ src/Kconfig | 23 +++++++++++++++++++++-- 8 files changed, 91 insertions(+), 8 deletions(-) -- 2.9.3

2 7

seabios serial console vs. sgabios
by Gerd Hoffmann 06 Nov '17

06 Nov '17

Hi, One problem we have with the serial console support in seabios and sgabios: It can happen that both are activated. We'll go fix that in qemu, but that'll work for new qemu versions only, not for the old already released ones. So I think it would be useful if seabios makes sure sgabios and serial console are not both active at the same time, i.e. in case we find "vgaroms/sgabios.bin" being present either turn off builtin serial console, or don't load sgabios (and possibly also enable builtin serial console). Comments? cheers, Gerd

5 7

[PATCH] Ignore sgabios rom in case sercon is enabled.
by Gerd Hoffmann 04 Nov '17

04 Nov '17

We don't want them both be active. Signed-off-by: Gerd Hoffmann <kraxel(a)redhat.com> --- src/optionroms.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/src/optionroms.c b/src/optionroms.c index 092393a56c..a5ecd4cd11 100644 --- a/src/optionroms.c +++ b/src/optionroms.c @@ -193,6 +193,11 @@ run_file_roms(const char *prefix, int isvga, u64 *sources) file = romfile_findprefix(prefix, file); if (!file) break; + if (strcmp(file->name, "vgaroms/sgabios.bin") == 0 && + CONFIG_SERCON && romfile_loadint("etc/sercon-port", 0)) { + dprintf(1, "sercon: is enabled, not loading sgabios rom.\n"); + continue; + } struct rom_header *rom = deploy_romfile(file); if (rom) { setRomSource(sources, rom, (u32)file); -- 2.9.3

3 2

[PATCH] Disable ScreenAndDebug in case both serial console and serial debug are active
by Gerd Hoffmann 03 Nov '17

03 Nov '17

Signed-off-by: Gerd Hoffmann <kraxel(a)redhat.com> --- src/sercon.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/src/sercon.c b/src/sercon.c index 0b7722ec08..72ebf82452 100644 --- a/src/sercon.c +++ b/src/sercon.c @@ -523,6 +523,10 @@ void sercon_setup(void) return; dprintf(1, "sercon: using ioport 0x%x\n", addr); + if (CONFIG_DEBUG_SERIAL) + if (addr == CONFIG_DEBUG_SERIAL_PORT) + ScreenAndDebug = 0; + vgabios = GET_IVT(0x10); seabios = FUNC16(entry_10); if (vgabios.seg != seabios.seg || -- 2.9.3

2 1

[PATCH] Increase precision of usec timer calculation
by Stefan Berger 02 Nov '17

02 Nov '17

When timer_calc_usec() is used with large timeout falues, such as 60s, the function lacks precision and produces different results than when using timer_calc(time / 1000) for the same timeout. This patch fixes the precision issue by falling back to timer_calc(time / 1000) for usec values over one second. Signed-off-by: Stefan Berger <stefanb(a)linux.vnet.ibm.com> --- src/hw/timer.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/src/hw/timer.c b/src/hw/timer.c index 03d22b2..2f0c864 100644 --- a/src/hw/timer.c +++ b/src/hw/timer.c @@ -213,6 +213,8 @@ timer_calc(u32 msecs) u32 timer_calc_usec(u32 usecs) { + if (usecs > 1000000) + return timer_calc(usecs / 1000); return timer_read() + DIV_ROUND_UP(GET_GLOBAL(TimerKHz) * usecs, 1000); } -- 2.5.5

2 4

[PATCH] serial console conflicts with serial debugging
by Gerd Hoffmann 02 Nov '17

02 Nov '17

Signed-off-by: Gerd Hoffmann <kraxel(a)redhat.com> --- src/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/src/Kconfig b/src/Kconfig index 4870a3e7b7..985594c51b 100644 --- a/src/Kconfig +++ b/src/Kconfig @@ -553,6 +553,7 @@ menu "Debugging" config DEBUG_SERIAL depends on DEBUG_LEVEL != 0 bool "Serial port debugging" + depends on !SERCON default n help Send debugging information to serial port. -- 2.9.3

3 3

USB Device is not stable, Cannot use it with long time
by Lying 31 Oct '17

31 Oct '17

First, I can running my qemu guest normally, And i can use keyboard and mouse in my VM, But i cannot use those after one day. And cannot find any useful log in my system. those message is my VM state( with qemu-monitor-command ): Device 0.1, Port 1, Speed 1.5 Mb/s, Product G100s Gaming Keyboard, ID: hostdev0 Device 0.2, Port 2, Speed 1.5 Mb/s, Product USB Optical Mouse, ID: hostdev1 it show me qemu add device normally,But i cannot use it

1 0

Re: [SeaBIOS] Failure to detect high-capacity SD card
by Kevin O'Connor 28 Oct '17

28 Oct '17

On Fri, Oct 27, 2017 at 08:26:35PM -0500, Matt DeVillier wrote: > On Fri, Oct 27, 2017 at 7:08 PM, Kevin O'Connor <kevin(a)koconnor.net> wrote: > > > Can you try the two patches below? (Go back to master, apply the > > first, gather the log, go back to master, apply the second, gather the > > log.) > > > > -Kevin > > > > w/patch 1 only: https://paste.ubuntu.com/25833503/ > > w/patch 2 only: https://paste.ubuntu.com/25833542/ Okay. One more thing to try is a slightly different voltage range. -Kevin --- a/src/hw/sdcard.c +++ b/src/hw/sdcard.c @@ -123,6 +123,17 @@ struct sdhci_s { #define SRF_DATA 0x04 // SDHCI result flags +#define SR_OCR_VDD_165_195 (1<<7) +#define SR_OCR_VDD_27_28 (1<<15) +#define SR_OCR_VDD_28_29 (1<<16) +#define SR_OCR_VDD_29_30 (1<<17) +#define SR_OCR_VDD_30_31 (1<<18) +#define SR_OCR_VDD_31_32 (1<<19) +#define SR_OCR_VDD_32_33 (1<<20) +#define SR_OCR_VDD_33_34 (1<<21) +#define SR_OCR_VDD_34_35 (1<<22) +#define SR_OCR_VDD_35_36 (1<<23) +#define SR_OCR_S18R (1<<24) #define SR_OCR_CCS (1<<30) #define SR_OCR_NOTBUSY (1<<31) @@ -181,7 +192,7 @@ static int sdcard_pio(struct sdhci_s *regs, int cmd, u32 *param) { u32 state = readl(&regs->present_state); - dprintf(9, "sdcard_pio cmd %x %x %x\n", cmd, *param, state); + dprintf(1, "sdcard_pio cmd %x %x %x\n", cmd, *param, state); if ((state & SP_CMD_INHIBIT) || ((cmd & 0x03) == 0x03 && state & SP_DAT_INHIBIT)) { dprintf(1, "sdcard_pio not ready %x\n", state); @@ -203,7 +214,7 @@ sdcard_pio(struct sdhci_s *regs, int cmd, u32 *param) writew(&regs->irq_status, SI_CMD_COMPLETE); // Read response memcpy(param, regs->response, sizeof(regs->response)); - dprintf(9, "sdcard cmd %x response %x %x %x %x\n" + dprintf(1, "sdcard cmd %x response %x %x %x %x\n" , cmd, param[0], param[1], param[2], param[3]); return 0; } @@ -302,13 +313,13 @@ sdcard_set_power(struct sdhci_s *regs) u32 cap = readl(&regs->cap_lo); u32 volt, vbits; if (cap & SD_CAPLO_V33) { - volt = 1<<20; + volt = SR_OCR_VDD_32_33 | SR_OCR_VDD_33_34; vbits = SPC_V33; } else if (cap & SD_CAPLO_V30) { - volt = 1<<18; + volt = SR_OCR_VDD_29_30 | SR_OCR_VDD_30_31; vbits = SPC_V30; } else if (cap & SD_CAPLO_V18) { - volt = 1<<7; + volt = SR_OCR_VDD_165_195; vbits = SPC_V18; } else { dprintf(1, "SD controller unsupported volt range (%x)\n", cap); @@ -399,11 +410,11 @@ sdcard_card_setup(struct sddrive_s *drive, int volt, int prio) if (ret) return ret; // Let card know SDHC/SDXC is supported and confirm voltage - u32 hcs = 0, vrange = (volt >= (1<<15) ? 0x100 : 0x200) | 0xaa; + u32 hcs = 0, vrange = (volt >= SR_OCR_VDD_27_28 ? 0x100 : 0x200) | 0xaa; param[0] = vrange; ret = sdcard_pio(regs, SC_SEND_IF_COND, param); if (!ret && param[0] == vrange) - hcs = (1<<30); + hcs = SR_OCR_CCS | (1<<28); // Verify SD card (instead of MMC or SDIO) param[0] = 0x00; ret = sdcard_pio_app(regs, SC_APP_SEND_OP_COND, param); @@ -414,8 +425,12 @@ sdcard_card_setup(struct sddrive_s *drive, int volt, int prio) if (ret) return ret; drive->card_type |= SF_MMC; - hcs = (1<<30); + hcs = SR_OCR_CCS; } + // XXX + if (volt == (SR_OCR_VDD_32_33 | SR_OCR_VDD_33_34) + && hcs && readl(&regs->cap_lo) & SD_CAPLO_V18) + hcs |= SR_OCR_S18R; // Init card u32 end = timer_calc(SDHCI_POWERUP_TIMEOUT); for (;;) { @@ -434,6 +449,10 @@ sdcard_card_setup(struct sddrive_s *drive, int volt, int prio) } msleep(5); // Avoid flooding log when debugging } + if (param[0] & SR_OCR_S18R) { + dprintf(1, "Card requests 1.8 volt\n"); + return -1; + } drive->card_type |= (param[0] & SR_OCR_CCS) ? SF_HIGHCAPACITY : 0; // Select card (get cid, set rca, get csd, select card) param[0] = 0x00;

2 1

Re: [SeaBIOS] Failure to detect high-capacity SD card
by Kevin O'Connor 28 Oct '17

28 Oct '17

On Thu, Oct 26, 2017 at 09:04:57PM -0500, Matt DeVillier wrote: > On Thu, Oct 26, 2017 at 11:26 AM, Kevin O'Connor <kevin(a)koconnor.net> wrote: > > > On Thu, Oct 26, 2017 at 11:14:53AM -0400, Matt DeVillier wrote: > > > On Thu, Oct 26, 2017 at 10:59 AM, Kevin O'Connor <kevin(a)koconnor.net> > > wrote: > > > > > > with the patch, USB devices are not available from the boot menu either > > (it > > > appears to hang?), and therefore device cannot be booted to get said log > > > > That sounds like a bad compile then - nothing in the patch should have > > impacted usb. > > > > In any case, the patch below tests the same thing and it should work > > on eMMC also. > > still timing out on detection: https://paste.ubuntu.com/25826387/ Can you try the two patches below? (Go back to master, apply the first, gather the log, go back to master, apply the second, gather the log.) -Kevin ====================== patch 1 ========================= --- a/src/hw/sdcard.c +++ b/src/hw/sdcard.c @@ -181,7 +181,7 @@ static int sdcard_pio(struct sdhci_s *regs, int cmd, u32 *param) { u32 state = readl(&regs->present_state); - dprintf(9, "sdcard_pio cmd %x %x %x\n", cmd, *param, state); + dprintf(1, "sdcard_pio cmd %x %x %x\n", cmd, *param, state); if ((state & SP_CMD_INHIBIT) || ((cmd & 0x03) == 0x03 && state & SP_DAT_INHIBIT)) { dprintf(1, "sdcard_pio not ready %x\n", state); @@ -203,7 +203,7 @@ sdcard_pio(struct sdhci_s *regs, int cmd, u32 *param) writew(&regs->irq_status, SI_CMD_COMPLETE); // Read response memcpy(param, regs->response, sizeof(regs->response)); - dprintf(9, "sdcard cmd %x response %x %x %x %x\n" + dprintf(1, "sdcard cmd %x response %x %x %x %x\n" , cmd, param[0], param[1], param[2], param[3]); return 0; } @@ -403,7 +403,7 @@ sdcard_card_setup(struct sddrive_s *drive, int volt, int prio) param[0] = vrange; ret = sdcard_pio(regs, SC_SEND_IF_COND, param); if (!ret && param[0] == vrange) - hcs = (1<<30); + hcs = (1<<30) | (1<<28); // Verify SD card (instead of MMC or SDIO) param[0] = 0x00; ret = sdcard_pio_app(regs, SC_APP_SEND_OP_COND, param); ====================== patch 2 ========================= --- a/src/hw/sdcard.c +++ b/src/hw/sdcard.c @@ -123,6 +123,7 @@ struct sdhci_s { #define SRF_DATA 0x04 // SDHCI result flags +#define SR_OCR_S18R (1<<24) #define SR_OCR_CCS (1<<30) #define SR_OCR_NOTBUSY (1<<31) @@ -181,7 +182,7 @@ static int sdcard_pio(struct sdhci_s *regs, int cmd, u32 *param) { u32 state = readl(&regs->present_state); - dprintf(9, "sdcard_pio cmd %x %x %x\n", cmd, *param, state); + dprintf(1, "sdcard_pio cmd %x %x %x\n", cmd, *param, state); if ((state & SP_CMD_INHIBIT) || ((cmd & 0x03) == 0x03 && state & SP_DAT_INHIBIT)) { dprintf(1, "sdcard_pio not ready %x\n", state); @@ -203,7 +204,7 @@ sdcard_pio(struct sdhci_s *regs, int cmd, u32 *param) writew(&regs->irq_status, SI_CMD_COMPLETE); // Read response memcpy(param, regs->response, sizeof(regs->response)); - dprintf(9, "sdcard cmd %x response %x %x %x %x\n" + dprintf(1, "sdcard cmd %x response %x %x %x %x\n" , cmd, param[0], param[1], param[2], param[3]); return 0; } @@ -416,6 +417,9 @@ sdcard_card_setup(struct sddrive_s *drive, int volt, int prio) drive->card_type |= SF_MMC; hcs = (1<<30); } + // XXX + if (volt == (1<<20) && hcs && readl(&regs->cap_lo) & SD_CAPLO_V18) + hcs |= SR_OCR_S18R; // Init card u32 end = timer_calc(SDHCI_POWERUP_TIMEOUT); for (;;) { @@ -434,6 +438,10 @@ sdcard_card_setup(struct sddrive_s *drive, int volt, int prio) } msleep(5); // Avoid flooding log when debugging } + if (param[0] & SR_OCR_S18R) { + dprintf(1, "Card requests 1.8 volt\n"); + return -1; + } drive->card_type |= (param[0] & SR_OCR_CCS) ? SF_HIGHCAPACITY : 0; // Select card (get cid, set rca, get csd, select card) param[0] = 0x00;

2 1

Time for a seabios 1.11 release?
by Gerd Hoffmann 27 Oct '17

27 Oct '17

Hi, Quite a few changes accumulated in master. Time to plan a new release I think, so we can pick up the improvements in qemu 2.11. Comments? Anything pending for qemu which needs firmware support? cheers, Gerd

2 3

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

SeaBIOS