> OK, staring at this packet dump, I see that your firmware isn't setting > a default value for the IP TTL field, but seeing as the UEFI spec > doesn't require this, that's a Syslinux bug. Interestingly, the firmware > on my test machine does fill out the TTL field with a value of 64. > > -- > Matt Fleming, Intel Open Source Technology Center >Some additional news. The system where the EFI boot fail is a Sony SVC1512. I now tested it with an IBM T430 and there it ... let me say 'boot'. I got the screen with Getting cached packets My IP is 192.168.1.187 but than nothing happen. I run a tcpdump and see that in the background my vmlinuz, initrd and also sqashfs.img was loaded via HTTP. This image run automatically a script which perform a curl against my HTTP server. I see the incomming curl request in my HTTP accesslog, so all run fine, but on my laptop screen only the two rows above are shown. When booting in BIOS mode, the screen change and I see the boot of the linux and also the curl command and reply from the HTTP server. This reply tells me if the system was successfully registered and is important to see Regards Michael
On Wed, 26 Jun, at 04:55:59PM, Michael Szerencsits wrote:> I run a tcpdump and see that in the background my vmlinuz, initrd and > also sqashfs.img was loaded via HTTP. > > This image run automatically a script which perform a curl against my > HTTP server. I see the incomming curl request in my HTTP accesslog, > so all run fine, but on my laptop screen only the two rows above are > shown.What you successfully booted an EFI kernel on this machine via any other means? E.g. from disk? This sounds to me like the kernel is failing to initialise your graphics adapter. Unless I've misunderstood something. -- Matt Fleming, Intel Open Source Technology Center
On Thu, 27 Jun, at 02:08:34PM, Michael Szerencsits wrote:> > > Oh, also, if possible please test 6.01-pre4, just to make sure it's not > > the TCP packets that are now causing problems on your Sony machine. > > > > -- > > Matt Fleming, Intel Open Source Technology Center > > > > I tested the pre4 on both systems. Same problems as before. > > I attached the trace from both system boot-efi4.cap is from sony and boot-efi5.cap from lenovo > > I've seen that sony load the syslinux64.efi 2 times, while lenovo doesn't.I can't see any HTTP traffic in either of these dumps. Which machine did you capture these logs on? Are you sure that the HTTP packets are reaching your server? -- Matt Fleming, Intel Open Source Technology Center
> I can't see any HTTP traffic in either of these dumps. Which machine did > you capture these logs on? Are you sure that the HTTP packets are > reaching your server? > > -- > Matt Fleming, Intel Open Source Technology Center >I excluded the HTTP packets from the lenovo trace, because of the size from the trace
> > I can't see any HTTP traffic in either of these dumps. Which machine did > you capture these logs on? Are you sure that the HTTP packets are > reaching your server? > > -- > Matt Fleming, Intel Open Source Technology Center >The boot-efi5.cap is from the booting lenovo system. I ran the trace again, but the file is 100MB