Hi.
I encountered problems reloading L4Linux. After shutdown loaded L4Linux, I tried reloading the L4Linux (same configuration) using 'l' command in l4con but it didn't work(network device probe failed, lost ide disk interrupt). During booting L4linux showed "irq_thread: RMGR denied IRQ 14: code 0x1". Because I use l4con, I can see the system memory on bottom right. After shutdown loaded L4Linux, it still shows the same value before. Using 'a' command in l4con memory region used by previous L4Linux seem to be still allocated. IMHO, this is the source to fail to reloading L4Linux. Because memory region by used L4Linux includes the DMA region, reloading L4Linux cannot control the system device. Is this a bug in L4Linux or other L4Env server? Also, see the attached log.
Best regards,
Hi Jugwan,
I encountered problems reloading L4Linux. After shutdown loaded L4Linux, I tried reloading the L4Linux (same configuration) using 'l' command in l4con but it didn't work(network device probe failed, lost ide disk interrupt). During booting L4linux showed "irq_thread: RMGR denied IRQ 14: code 0x1". Because I use l4con, I can see the system memory on bottom right. After shutdown loaded L4Linux, it still shows the same value before. Using 'a' command in l4con memory region used by previous L4Linux seem to be still allocated. IMHO, this is the source to fail to reloading L4Linux. Because memory region by used L4Linux includes the DMA region, reloading L4Linux cannot control the system device. Is this a bug in L4Linux or other L4Env server? Also, see the attached log.
IIRC, for restarting and resource freeing to work your need to start the event server and tell all participating parties to use it (most often the --events switch).
Maybe others can provide more detail on this ...
Cheers, Martin
2006-03-12 (일), 19:19 +0100, Martin Pohlack 쓰시길:
Hi Jugwan,
I encountered problems reloading L4Linux. After shutdown loaded L4Linux, I tried reloading the L4Linux (same configuration) using 'l' command in l4con but it didn't work(network device probe failed, lost ide disk interrupt). During booting L4linux showed "irq_thread: RMGR denied IRQ 14: code 0x1". Because I use l4con, I can see the system memory on bottom right. After shutdown loaded L4Linux, it still shows the same value before. Using 'a' command in l4con memory region used by previous L4Linux seem to be still allocated. IMHO, this is the source to fail to reloading L4Linux. Because memory region by used L4Linux includes the DMA region, reloading L4Linux cannot control the system device. Is this a bug in L4Linux or other L4Env server? Also, see the attached log.
IIRC, for restarting and resource freeing to work your need to start the event server and tell all participating parties to use it (most often the --events switch).
Maybe others can provide more detail on this .
Thanks martin for advice. I tried to reload L4Linux with event server but still failed though some incorrect behavior disappeared. After shutdown loaded L4Linux, l4con showed the memory used by L4Linux was returned. But using 'a' command in run (not l4con, it's my mistake in previous mail) memory region is still allocated to the L4Linux. And reloading L4Linux still failed. Booting reloaded L4Linux hung in acquiring IRQ 11 (network card) and CPU utilization became 100%.
And I couldn't find using levent library or event related functions (levents_*) in either L4Linux Makefile (l4linux-2.6/arch/l4/Makefile) or any L4Linux sources. This is needed to use the event server, isn't it?
Please see the attached my grub configuration and log.
Hi,
On Mon Mar 13, 2006 at 16:18:14 +0900, Jugwan Eom wrote:
Thanks martin for advice. I tried to reload L4Linux with event server but still failed though some incorrect behavior disappeared. After shutdown loaded L4Linux, l4con showed the memory used by L4Linux was returned. But using 'a' command in run (not l4con, it's my mistake in previous mail) memory region is still allocated to the L4Linux. And
That's probably because you didn't specify --events for the loader.
reloading L4Linux still failed. Booting reloaded L4Linux hung in acquiring IRQ 11 (network card) and CPU utilization became 100%.
I'll try to investigate this.
And I couldn't find using levent library or event related functions (levents_*) in either L4Linux Makefile (l4linux-2.6/arch/l4/Makefile) or any L4Linux sources. This is needed to use the event server, isn't it?
No, an application itself doesn't need to know of this service, just the servers that allocate resources for a client.
title Fiasco / l4linux root (hd0,0) kernel /fiasco/l4/v2/bootstrap modaddr 0x06000000 module /fiasco/l4/v2/fiasco -nokdb -loadcnt -nowait -serial_esc -comspeed 115200 -comport 1 module /fiasco/l4/v2/sigma0 module /fiasco/l4/v2/roottask task modname "bmodfs" module module module module module module module /fiasco/l4/v2/names --events module /fiasco/l4/v2/log module /fiasco/l4/v2/dm_phys --isa=0x00800000 -v --events module /fiasco/l4/v2/simple_ts -t 380 --events module /fiasco/l4/v2/rtc --events module /fiasco/l4/v2/l4io --noirq --events module /fiasco/l4/v2/events module /fiasco/l4/v2/bmodfs --events module /fiasco/l4/v2/libloader.s.so module /l4/cfg/run.bmodfs.cfg module /fiasco/l4/v2/run module /fiasco/l4/lx/vmlinuz26 module /l4/cfg/lx/l4linux26_rd_tty.cfg module /l4/rd/tty.rd module /fiasco/l4/v2/l4exec --events module /fiasco/l4/v2/con --l4io --cpuload --events module /fiasco/l4/v2/loader --fprov=BMODFS run.bmodfs.cfg l4linux26_rd_tty.cfg vbeset 0x117
My config looks similar to the following:
... module .../events module .../names --events module .../log module .../dm_phys --events module .../simple_ts -t 400 --events module .../rtc module .../l4io --noirq module .../l4exec module .../con --events --l4io --cpuload module .../bmodfs modules ... module .../loader --events apps...
That are all '--events' switches that need to appear.
Adam
2006-03-13 (월), 19:29 +0100, Adam Lackorzynski 쓰시길:
reloading L4Linux still failed. Booting reloaded L4Linux hung in acquiring IRQ 11 (network card) and CPU utilization became 100%.
I'll try to investigate this.
With event server, reloading L4Linux is now in good working. But it fails to use a network device (E1000) while other devices like IDE and SCSI don't have any problems. Reloaded L4Linux kernel prints this message during booting. "e1000: probe of 0000:03:02.0 failed with error -5"
And there is one bothering message in log. .... con | vc_open_out(): vc[2] 1024x768@16, gmode:0x35 l4lx | l4fbcon: 1024x750 2bypp, size: 1536000 l4lx | l4lx_thread_create: Created thread 11.06 (CON refresher) l4lx | atexit(function = 0x004fa4d0) called
--l4io_request_mem_region error--------------------------IP: 0040d768 (11.03) jdb: .... Using 'g' command, I can reload L4Linux anyway.
Best regards,
On Fri Mar 17, 2006 at 09:26:46 +0900, Jugwan Eom wrote:
2006-03-13 (월), 19:29 +0100, Adam Lackorzynski 쓰시길:
reloading L4Linux still failed. Booting reloaded L4Linux hung in acquiring IRQ 11 (network card) and CPU utilization became 100%.
I'll try to investigate this.
With event server, reloading L4Linux is now in good working. But it fails to use a network device (E1000) while other devices like IDE and SCSI don't have any problems. Reloaded L4Linux kernel prints this message during booting. "e1000: probe of 0000:03:02.0 failed with error -5"
And there is one bothering message in log. .... con | vc_open_out(): vc[2] 1024x768@16, gmode:0x35 l4lx | l4fbcon: 1024x750 2bypp, size: 1536000 l4lx | l4lx_thread_create: Created thread 11.06 (CON refresher) l4lx | atexit(function = 0x004fa4d0) called
--l4io_request_mem_region error--------------------------IP:
0040d768 (11.03) jdb: .... Using 'g' command, I can reload L4Linux anyway.
Could be that the memory regions aren't freed properly. We'll investigate.
Adam
On Friday 17 March 2006 15:56, Adam Lackorzynski wrote:
On Fri Mar 17, 2006 at 09:26:46 +0900, Jugwan Eom wrote:
2006-03-13 (월), 19:29 +0100, Adam Lackorzynski 쓰시길:
reloading L4Linux still failed. Booting reloaded L4Linux hung in acquiring IRQ 11 (network card) and CPU utilization became 100%.
I'll try to investigate this.
With event server, reloading L4Linux is now in good working. But it fails to use a network device (E1000) while other devices like IDE and SCSI don't have any problems. Reloaded L4Linux kernel prints this message during booting. "e1000: probe of 0000:03:02.0 failed with error -5"
And there is one bothering message in log. .... con | vc_open_out(): vc[2] 1024x768@16, gmode:0x35 l4lx | l4fbcon: 1024x750 2bypp, size: 1536000 l4lx | l4lx_thread_create: Created thread 11.06 (CON refresher) l4lx | atexit(function = 0x004fa4d0) called
--l4io_request_mem_region error--------------------------IP:
0040d768 (11.03) jdb: .... Using 'g' command, I can reload L4Linux anyway.
Could be that the memory regions aren't freed properly. We'll investigate.
Or use the ore network server?
Frank
l4-hackers@os.inf.tu-dresden.de