Re: CFT: ndis(4) testers needed! - on 11.0-RELESE
Hi Gleb, Are you still accepting a bug report on this? I have a laptop that I need to use NDIS driver. This change impacted and causes kernel panic. I haven't used this laptop for a quite while and didn't realize it had stopped working. When I setup a wlan0 device with NDIS in wepmode, "ifconfig wlan0 up" causes kernel memory currupption and causes next ioctl/sysctl access programs such as "ps" and "sysctl -a" crash the kernel. Regards, Hiro On Fri, 31 Jul 2015 15:12:26 +0300 Gleb Smirnoff wrote: > Hi! > > I need to make couple of non-functional but rather large changes > to the if_ndis driver and will appreciate if anyone signs up to > test my changes. Please contact me if you can provide help. > > -- > Totus tuus, Glebius. > ___ > freebsd-current@freebsd.org mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-current > To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org" ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> > set hint.rsu.0.disabled="1" > This isn't sufficient. This leaves the device named rsu0 but disables it > without > letting other drivers attach to it. You need to remove the rsu driver > entirely > or use devctl to force the driver to ndis. > John Baldwin I can try on the slight chance that using devctl will make any difference regarding ndis. I have something else to do on that FreeBSD installation, like installing the missing devel/readline that should have been installed as a dependency of lang/gawk; I don't know why it wasn't. This error showed when I failed trying to build Haiku cross-tools. Tom ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
On Thursday, August 13, 2015 08:21:44 AM Thomas Mueller wrote: > > T> Thinking I might have used the wrong syntax with ifconfig, I tried again > > with rsu0 instead of ndis0 or other failed attempts, and was successful > > connecting to Internet. > > > AFAIU, if you got rsu(4) driver in your kernel it will attach to the > > hardware, > > and the ndis(4) driver won't be able to. > > > Have you tried to used ndis(4) before on an unpatched vanilla FreeBSD? > > > Totus tuus, Glebius. > > I have tried to use ndis before, but never successfully. I used custom > kernel, but src tree was not patched. > > Do I have to build a separate kernel without rsu to test ndis, or is > disabling at loader good enough? > > set hint.rsu.0.disabled="1" This isn't sufficient. This leaves the device named rsu0 but disables it without letting other drivers attach to it. You need to remove the rsu driver entirely or use devctl to force the driver to ndis. -- John Baldwin ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> If you haven't ever successfully run ndis before on this hardware, then > testing > the patch is quite useless :( Since we don't have a working reference point. > Sorry for waisting your time. > Totus tuus, Glebius. I remember trying ndis from a USB-stick installation of FreeBSD 9.2-STABLE, but it failed. I could try again, since I already have the setup, with set hint.rsu.0.disabled="1" from loader prompt. UPDATE: I tried, rsu did not show on "ifconfig", but it seems not viable. Last 15 lines of var/log/messages are: Aug 14 06:34:34 amelia2 root: /etc/rc: INFO: RSA1 host key exists. Aug 14 06:34:34 amelia2 root: /etc/rc: INFO: RSA host key exists. Aug 14 06:34:34 amelia2 root: /etc/rc: INFO: DSA host key exists. Aug 14 06:34:34 amelia2 root: /etc/rc: INFO: ECDSA host key exists. Aug 14 06:34:34 amelia2 root: /etc/rc: INFO: ED25519 host key exists. Aug 14 06:34:38 amelia2 login: ROOT LOGIN (root) ON ttyv0 Aug 14 06:36:19 amelia2 kernel: ndis0: on usbus1 Aug 14 06:36:19 amelia2 kernel: ndis0: NDIS API version: 5.1 Aug 14 06:36:19 amelia2 kernel: ndis0: ioctl 0xf800 isn't supported Aug 14 06:36:19 amelia2 last message repeated 20 times Aug 14 06:36:19 amelia2 kernel: ndis0: init handler failed Aug 14 06:36:19 amelia2 kernel: device_attach: ndis0 attach returned 6 Aug 14 06:36:19 amelia2 root: Unknown USB device: vendor 0x0cf3 product 0x3000 bus uhub4 Aug 14 06:36:19 amelia2 root: Unknown USB device: vendor 0x0cf3 product 0x9271 bus uhub4 Aug 14 06:37:26 amelia2 login: ROOT LOGIN (root) ON ttyv1 I feel like I was justified in this last try, having gone so far already. I browsed en.wikipedia.org/wiki/NDISwrapper and saw various reasons why ndis would likely fail. Maybe it's really not workable for Hiro H50191. Now should I svn revert the src tree and update normally, without patches? When possible, for the sake of others if not myself, it would be good to have more documentation of how to set up with ndis, and what to do when the Windows drivers don't include .inf and .sys files. Regarding two Unknown USB device messages, I believe 0x9271 is on-motherboard quasi-USB wireless adapter, and 0x3000 might be on-motherboard Bluetooth (ubt). Tom ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> If you haven't ever successfully run ndis before on this hardware, then > testing > the patch is quite useless :( Since we don't have a working reference point. > Sorry for waisting your time. > Totus tuus, Glebius. I remember trying ndis from a USB-stick installation of FreeBSD 9.2-STABLE, but it failed. I could try again, since I already have the setup, with set hint.rsu.0.disabled="1" from loader prompt. Still a small chance, but I have the setup. Maybe there is a needed firmware file in non-obvious place. Then should I svn revert the src tree and update normally, without patches? When possible, it would be good to have more documentation of how to set up with ndis, and what to do when the Windows drivers don't include .inf and .sys files. Tom ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
Thomas, On Thu, Aug 13, 2015 at 08:21:44AM +, Thomas Mueller wrote: T> > T> Thinking I might have used the wrong syntax with ifconfig, I tried again with rsu0 instead of ndis0 or other failed attempts, and was successful connecting to Internet. T> T> > AFAIU, if you got rsu(4) driver in your kernel it will attach to the hardware, T> > and the ndis(4) driver won't be able to. T> T> > Have you tried to used ndis(4) before on an unpatched vanilla FreeBSD? T> T> > Totus tuus, Glebius. T> T> I have tried to use ndis before, but never successfully. I used custom kernel, but src tree was not patched. T> T> Do I have to build a separate kernel without rsu to test ndis, or is disabling at loader good enough? T> T> set hint.rsu.0.disabled="1" T> T> Then would the device name be ndis0, or rtl8192su0, or without the 0, or something else? T> T> Would I have to run ndisgen again? Or did the presence of rsu in kernel have no effect on building the module? T> T> I might also want to test on FreeBSD-current i386. T> T> Which MS-Windows version should I prefer between 2000, XP, Vista and 7? My fallible inclination is toward XP. If you haven't ever successfully run ndis before on this hardware, then testing the patch is quite useless :( Since we don't have a working reference point. Sorry for waisting your time. -- Totus tuus, Glebius. ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> T> Thinking I might have used the wrong syntax with ifconfig, I tried again > with rsu0 instead of ndis0 or other failed attempts, and was successful > connecting to Internet. > AFAIU, if you got rsu(4) driver in your kernel it will attach to the hardware, > and the ndis(4) driver won't be able to. > Have you tried to used ndis(4) before on an unpatched vanilla FreeBSD? > Totus tuus, Glebius. I have tried to use ndis before, but never successfully. I used custom kernel, but src tree was not patched. Do I have to build a separate kernel without rsu to test ndis, or is disabling at loader good enough? set hint.rsu.0.disabled="1" Then would the device name be ndis0, or rtl8192su0, or without the 0, or something else? Would I have to run ndisgen again? Or did the presence of rsu in kernel have no effect on building the module? I might also want to test on FreeBSD-current i386. Which MS-Windows version should I prefer between 2000, XP, Vista and 7? My fallible inclination is toward XP. Tom ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
Thomas, On Wed, Aug 12, 2015 at 08:16:12PM -0700, Thomas Mueller wrote: T> GENERIB differs from GENERIC by eliminating some old devices like SCSI cards and ISA EThernet cards, but I add some wi-fi drivers like urtw, urtwn and rsu. T> T> Using the main custom kernel, I was able to build an ndis module from the WinX64 drivers net8192su.inf, rtl8192su.sys and net8192su.cat, but when I tried to ifconfig, the device ndis, ndis0, rtl8192su, rtl8192su0, were not recognized. T> T> Being in FreeBSD-current amd64, I used 64-bit WinXP (WinX64); there were also drivers for Win 32-bit, for 2000, XP, Vista and 7, all offered in both 32-bit and 64-bit, except Win 2000, 32-bit only. T> T> Hiro USB wireless adapter is H50191, their website is www.hiroinc.com . T> T> I wonder if I could have missed a firmware file but couldn't find it in the directory structure (maybe net8192su.cat ?). T> T> Thinking I might have used the wrong syntax with ifconfig, I tried again with rsu0 instead of ndis0 or other failed attempts, and was successful connecting to Internet. AFAIU, if you got rsu(4) driver in your kernel it will attach to the hardware, and the ndis(4) driver won't be able to. Have you tried to used ndis(4) before on an unpatched vanilla FreeBSD? -- Totus tuus, Glebius. ___ freebsd-current@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
UPDATE: I was successful with buildworld and kernel with the patch, both the custom kernel with ndis and the GENERIB kernel without ndis. GENERIB differs from GENERIC by eliminating some old devices like SCSI cards and ISA EThernet cards, but I add some wi-fi drivers like urtw, urtwn and rsu. Using the main custom kernel, I was able to build an ndis module from the WinX64 drivers net8192su.inf, rtl8192su.sys and net8192su.cat, but when I tried to ifconfig, the device ndis, ndis0, rtl8192su, rtl8192su0, were not recognized. Being in FreeBSD-current amd64, I used 64-bit WinXP (WinX64); there were also drivers for Win 32-bit, for 2000, XP, Vista and 7, all offered in both 32-bit and 64-bit, except Win 2000, 32-bit only. Hiro USB wireless adapter is H50191, their website is www.hiroinc.com . I wonder if I could have missed a firmware file but couldn't find it in the directory structure (maybe net8192su.cat ?). Thinking I might have used the wrong syntax with ifconfig, I tried again with rsu0 instead of ndis0 or other failed attempts, and was successful connecting to Internet. Where to go from here? I don't want to svn up the src tree until I receive further instructions, should I revert or just update on top of what's already there? But with the FreeBSD-native rsu, I did update ports and doc trees using svn. Files I got with ndisgen, and the original driver files, are, copying-pasting from vt console. If double spaced, that's because of a vt bug: root@amelia2:/RTL8188_8191_8192_SU_WindowsDriver_1086.48.0809.2011.F0049_12.P040 6_UI_1.00.0187.L/88_91_92_SU_Driver/WinX64 # ls -l total 2696 -rw-r--r-- 1 root wheel1 Aug 13 01:54 bus_if.h -rw-r--r-- 1 root wheel1 Aug 13 01:54 device_if.h -rwxr-xr-x 1 root wheel13762 Aug 19 2011 net8192su.cat -rwxr-xr-x 1 root wheel15622 Aug 13 01:50 net8192su.cat.ko -rwxr-xr-x 1 root wheel29731 Aug 13 01:48 net8192su.inf -rwxr-xr-x 1 root wheel 776808 Aug 11 2011 rtl8192su.sys -rw-r--r-- 1 root wheel 1818552 Aug 13 01:54 rtl8192su_sys.ko Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
On Wed, Aug 12, 2015 at 4:15 AM, Thomas Mueller wrote: > > I don't think anybody makes 128 MB USB sticks any more, smallest in > current production may be 4 GB or 8 GB. For the price, no use buying > anything under 8 GB or perhaps 16 GB. > The purpose of this image is just for helping people to test easily theses changes without to have to applying patch and rebuilding/reinstalling kernel. Then I've reduced at maximum the image size for faster image installation (it takes lot's of less time to dd a 128MB file than a 8GB file). And you can install this image on a bigger disk without any problem. Regards, Olivier ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> I've build a small nanobsd image (for i386 arch) that include this patch: > http://dev.bsdrp.net/FreeBSD/GLEBWIFI-0.2-full-i386-vga.img.xz > You can install it on a small (128MB USB flash) for testing it without > re-installing your system. > Installation instruction for copying this image into a USB flash disk > (/dev/da1 in this example): > fetch http://dev.bsdrp.net/FreeBSD/GLEBWIFI-0.2-full-i386-vga.img.xz > -o - | xzcat | dd of=/dev/da1 bs=128k > Then boot on this USB flash disk. > Regards, > Olivier Now there's another update from Gleb Smirnoff to try, so I think I'll try that, start "make buildworld" at bedtime. I ran svn revert -R /BETA1/usr/src and intend to update and patch within a few hours. I don't think anybody makes 128 MB USB sticks any more, smallest in current production may be 4 GB or 8 GB. For the price, no use buying anything under 8 GB or perhaps 16 GB. I have two old IDE hard drives, 341 MB and 1271 MB, that I access with Sabrent USB 2.0 enclosure. This enclosure also has eSATA interface that works with SATA but not IDE hard drives. 1271 MB drive worked when I last tried some months ago, but 341 MB drive just gives gibberish. I don't know whether the 55 sectors per track is incompatible with USB adapter, or if the drive has gone bad. Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
On Tue, Aug 11, 2015 at 09:20:08PM +, Thomas Mueller wrote: T> /usr/src/sys/modules/iwm/../../dev/iwm/if_iwm.c:3843:10: error: use of undeclared identifier 'ifp'; did you mean 'if'? T> KASSERT(ifp->if_drv_flags & IFF_DRV_RUNNING, ("not running")); T> ^~~ T> if Here is updated patch. -- Totus tuus, Glebius. net80211.diff.bz2 Description: Binary data ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
I did the update a little before your last message. svn revert by itself doesn't work, I get an error message, then run svn help revert and that gets me on the right track. I did the update and patch from NetBSD, which I believe works the same as it would from FreeBSD, subsequently booted into FreeBSD and cd /usr/src make buildworld worked again with no error, but make kernel failed. Error occurs somewhere in iwi or iwm, and that prevents the kernel from building, so I don't get to test ndis, but it may point to what needs fixing. Would it be safe to comment out iwi and other Intel wireless nics in kernel config and then try again to make buildkernel without rerunning buildworld? Or should I wait for next update? But you probably want to know where the snags are even if they don't relate to my hardware. End of the buildkernel.log: ===> iwi (all) cc -O2 -pipe -fno-strict-aliasing -Werror -D_KERNEL -DKLD_MODULE -nostdinc -DHAVE_KERNEL_OPTION_HEADERS -include /usr/obj/usr/src/sys/SANDY11NC/opt_global.h -I. -I/usr/src/sys -fno-common -g -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -I/usr/obj/usr/src/sys/SANDY11NC -mcmodel=kernel -mno-red-zone -mno-mmx -mno-sse -msoft-float -fno-asynchronous-unwind-tables -ffreestanding -fwrapv -fstack-protector -gdwarf-2 -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -Wundef -Wno-pointer-sign -D__printf__=__freebsd_kprintf__ -Wmissing-include-dirs -fdiagnostics-show-option -Wno-unknown-pragmas -Wno-error-tautological-compare -Wno-error-empty-body -Wno-error-parentheses-equality -Wno-error-unused-function -Wno-error-pointer-sign -mno-aes -mno-avx -std=iso9899:1999 -c /usr/src/sys/modules/iwi/../../dev/iwi/if_iwi.c -o if_iwi.o ctfconvert -L VERSION -g if_iwi.o ld -d -warn-common -r -d -o if_iwi.ko.debug if_iwi.o ctfmerge -L VERSION -g -o if_iwi.ko.debug if_iwi.o :> export_syms awk -f /usr/src/sys/conf/kmod_syms.awk if_iwi.ko.debug export_syms | xargs -J% objcopy % if_iwi.ko.debug objcopy --only-keep-debug if_iwi.ko.debug if_iwi.ko.symbols objcopy --strip-debug --add-gnu-debuglink=if_iwi.ko.symbols if_iwi.ko.debug if_iwi.ko ===> iwifw (all) ===> iwifw/iwi_bss (all) uudecode -p /usr/src/sys/modules/iwifw/iwi_bss/../../../contrib/dev/iwi/ipw2200-bss.fw.uu > iwi_bss iwi_bss iwi_bss cc -O2 -pipe -fno-strict-aliasing -Werror -D_KERNEL -DKLD_MODULE -nostdinc -DHAVE_KERNEL_OPTION_HEADERS -include /usr/obj/usr/src/sys/SANDY11NC/opt_global.h -I. -I/usr/src/sys -fno-common -g -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -I/usr/obj/usr/src/sys/SANDY11NC -mcmodel=kernel -mno-red-zone -mno-mmx -mno-sse -msoft-float -fno-asynchronous-unwind-tables -ffreestanding -fwrapv -fstack-protector -gdwarf-2 -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -Wundef -Wno-pointer-sign -D__printf__=__freebsd_kprintf__ -Wmissing-include-dirs -fdiagnostics-show-option -Wno-unknown-pragmas -Wno-error-tautological-compare -Wno-error-empty-body -Wno-error-parentheses-equality -Wno-error-unused-function -Wno-error-pointer-sign -mno-aes -mno-avx -std=iso9899:1999 -c iwi_bss.c -o iwi_bss.o ctfconvert -L VERSION -g iwi_bss.o ld -d -warn-common -r -d -o iwi_bss.ko.debug iwi_bss.fwo iwi_bss.o ctfmerge -L VERSION -g -o iwi_bss.ko.debug iwi_bss.fwo iwi_bss.o :> export_syms awk -f /usr/src/sys/conf/kmod_syms.awk iwi_bss.ko.debug export_syms | xargs -J% objcopy % iwi_bss.ko.debug objcopy --only-keep-debug iwi_bss.ko.debug iwi_bss.ko.symbols objcopy --strip-debug --add-gnu-debuglink=iwi_bss.ko.symbols iwi_bss.ko.debug iwi_bss.ko ===> iwifw/iwi_ibss (all) uudecode -p /usr/src/sys/modules/iwifw/iwi_ibss/../../../contrib/dev/iwi/ipw2200-ibss.fw.uu > iwi_ibss iwi_ibss iwi_ibss cc -O2 -pipe -fno-strict-aliasing -Werror -D_KERNEL -DKLD_MODULE -nostdinc -DHAVE_KERNEL_OPTION_HEADERS -include /usr/obj/usr/src/sys/SANDY11NC/opt_global.h -I. -I/usr/src/sys -fno-common -g -fno-omit-frame-pointer -mno-omit-leaf-frame-pointer -I/usr/obj/usr/src/sys/SANDY11NC -mcmodel=kernel -mno-red-zone -mno-mmx -mno-sse -msoft-float -fno-asynchronous-unwind-tables -ffreestanding -fwrapv -fstack-protector -gdwarf-2 -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual -Wundef -Wno-pointer-sign -D__printf__=__freebsd_kprintf__ -Wmissing-include-dirs -fdiagnostics-show-option -Wno-unknown-pragmas -Wno-error-tautological-compare -Wno-error-empty-body -Wno-error-parentheses-equality -Wno-error-unused-function -Wno-error-pointer-sign -mno-aes -mno-avx -std=iso9899:1999 -c iwi_ibss.c -o iwi_ibss.o ctfconvert -L VERSION -g iwi_ibss.o ld -d -warn-common -r -d -o iwi_ibss.ko.debug iwi_ibss.fwo iwi_ibss.o ctfmerge -L VERSION -g -o iwi_ibss.ko.debug iwi_ibss.fwo iwi_ibss.o :> export_syms awk -f /usr/src/sys/conf/kmod_syms.awk iwi_ibss.ko
Re: CFT: ndis(4) testers needed!
On Mon, Aug 10, 2015 at 11:27 PM, Thomas Mueller wrote: > > I guess I need to do, from src tree base, > > svn revert > svn up . > patch < /path/to/net80211.diff > > Does that look correct? > > I've build a small nanobsd image (for i386 arch) that include this patch: http://dev.bsdrp.net/FreeBSD/GLEBWIFI-0.2-full-i386-vga.img.xz You can install it on a small (128MB USB flash) for testing it without re-installing your system. Installation instruction for copying this image into a USB flash disk (/dev/da1 in this example): fetch http://dev.bsdrp.net/FreeBSD/GLEBWIFI-0.2-full-i386-vga.img.xz -o - | xzcat | dd of=/dev/da1 bs=128k Then boot on this USB flash disk. Regards, Olivier ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
Thomas, On Mon, Aug 10, 2015 at 09:27:44PM +, Thomas Mueller wrote: T> You sent a patch, net80211.diff.bz2 . T> T> Before I extract and apply the patch, how does that differ from D2655.diff ? Is it an updated version, does it include ndis or only FreeBSD-native drivers? It is updated version. It includes updates for all drivers, including ndis. T> Do you want me to test Hiro H50191 with native rsu, or ndis, or both separately? Only ndis please. T> I guess I need to do, from src tree base, T> T> svn revert T> svn up . T> patch < /path/to/net80211.diff T> T> Does that look correct? Yes, this is correct. -- Totus tuus, Glebius. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
You sent a patch, net80211.diff.bz2 . Before I extract and apply the patch, how does that differ from D2655.diff ? Is it an updated version, does it include ndis or only FreeBSD-native drivers? Do you want me to test Hiro H50191 with native rsu, or ndis, or both separately? I guess I need to do, from src tree base, svn revert svn up . patch < /path/to/net80211.diff Does that look correct? Packages/ports have fallen behind, due to problems with rsu driver and not being able to use Ethernet in FreeBSD (now OK), still on png-1.5.18; xorg-server is still on 1.12.4_7,1 . An attempt to upgrade using portmaster in accordance with UPDATING 20141225 instructions hit a snag, could possibly have rendered xorg-server nonrunnable. Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
Thomas, On Mon, Aug 10, 2015 at 01:29:49PM +, Thomas Mueller wrote: T> I didn't mean to send the same old message a second time (sorry); I was at a difficult-to-read interface, this is what I meant to send: T> T> UPDATE, typing from FreeBSD vt and X crashing system: T> T> I ran svn up on the source tree, applied the patch, make buildworld succeeded, but building the kernels all failed, with and without ndis, so I dare not make installworld, which would put userland out of sync with kernels and likely give me nonworking system. T> T> Now, I try to startx as nonroot, then there is no more signal to monitor, but keyboard is recognized, at least CapsLock and ScrollLock light and unlight the LEDs. Returning Ctrl-Alt-F1 blindly and typing "shutdown -r now" does nothing, I must press Reset button. T> T> Now I want to undo the patch, unless there is something better you can suggest, like maybe a fix? T> T> I want to get a working system again. T> T> svn revert --depth=infinity ? Yes, 'svn revert --depth=infinity .' in the root of sources will undo any local changes. T> I don't really know what happened that X won't start any more, it's on a Western Digital Green Drive bought at the end of May 2013, maybe it's nearing end of life? T> T> Or maybe something with the buildworld and buildkernel had an adverse effect on the file system? buildworld and buildkernel can't affect your installation. -- Totus tuus, Glebius. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> I ended with doing the aforementioned 2 changes as one. > The diff is living here: > https://reviews.freebsd.org/D2655 > It spans the net80211 layer and converts (almost) all 802.11 drivers > to new KPI. Now it also converts if_ndis, and this is the change I > am asking you to test. > Unlike all other drivers, converting if_ndis wasn't trivial. There are > likely bugs introduced, so it may happen that we will need several > rounds of testing. > Totus tuus, Glebius. I have Hiro H50191 USB WLAN adapter, including .inf and .sys files. I also have, on MSI MPOWER motherboard, Atheros AR9271 quasi-USB WLAN, and Realtek 8111E Ethernet, with MS-Windows drivers but no .INF and .SYS files on DVD that came with the motherboard. Also, after applying the patch to the source tree, what if anything do I have to do to not be messed up for subsequent svn updates to source tree? I didn't mean to send the same old message a second time (sorry); I was at a difficult-to-read interface, this is what I meant to send: UPDATE, typing from FreeBSD vt and X crashing system: I ran svn up on the source tree, applied the patch, make buildworld succeeded, but building the kernels all failed, with and without ndis, so I dare not make installworld, which would put userland out of sync with kernels and likely give me nonworking system. Now, I try to startx as nonroot, then there is no more signal to monitor, but keyboard is recognized, at least CapsLock and ScrollLock light and unlight the LEDs. Returning Ctrl-Alt-F1 blindly and typing "shutdown -r now" does nothing, I must press Reset button. Now I want to undo the patch, unless there is something better you can suggest, like maybe a fix? I want to get a working system again. svn revert --depth=infinity ? I don't really know what happened that X won't start any more, it's on a Western Digital Green Drive bought at the end of May 2013, maybe it's nearing end of life? Or maybe something with the buildworld and buildkernel had an adverse effect on the file system? I think I'm missing something in the References headers but was doing the best I can without graphic interface. Text DRMKMS screen I get when booting NetBSD-current looks the same as what I get in FreeBSD-current after returning from X, and much better-looking than the text vt screen I get upon booting FreeBSD before starting X. Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
Thomas, On Mon, Aug 10, 2015 at 08:46:06AM +, Thomas Mueller wrote: T> What do I need to download? There is a lot of stuff on the page you cite, and I see a link to the right, Download Raw Diff. Is that what I'm supposed to download? Here is patch attached to this email. T> What do you want me to test it on? T> T> I have Hiro H50191 USB WLAN adapter, including .inf and .sys files. If this device is now working with FreeBSD 11, then it is the device to test. So please just test on Hiro and that would be enough. T> I also have, on MSI MPOWER motherboard, Atheros AR9271 quasi-USB WLAN, and Realtek 8111E Ethernet, with MS-Windows drivers but no .INF and .SYS files on DVD that came with the motherboard. T> T> Is there a way to convert MS-Windows drivers that come with no .inf and .sys files to ndis-compatible format? Or just test on the Hiro H50191? Testing on Hiro would be enough. T> Also, after applying the patch to the source tree, what if anything do I have to do to not be messed up for subsequent svn updates to source tree? You can always revert all changes to your tree with 'svn revert'. -- Totus tuus, Glebius. net80211.diff.bz2 Description: Binary data ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> I ended with doing the aforementioned 2 changes as one. > The diff is living here: > https://reviews.freebsd.org/D2655 > It spans the net80211 layer and converts (almost) all 802.11 drivers > to new KPI. Now it also converts if_ndis, and this is the change I > am asking you to test. > Unlike all other drivers, converting if_ndis wasn't trivial. There are > likely bugs introduced, so it may happen that we will need several > rounds of testing. > Totus tuus, Glebius. What do I need to download? There is a lot of stuff on the page you cite, and I see a link to the right, Download Raw Diff. Is that what I'm supposed to download? What do you want me to test it on? I have Hiro H50191 USB WLAN adapter, including .inf and .sys files. I also have, on MSI MPOWER motherboard, Atheros AR9271 quasi-USB WLAN, and Realtek 8111E Ethernet, with MS-Windows drivers but no .INF and .SYS files on DVD that came with the motherboard. Is there a way to convert MS-Windows drivers that come with no .inf and .sys files to ndis-compatible format? Or just test on the Hiro H50191? Also, after applying the patch to the source tree, what if anything do I have to do to not be messed up for subsequent svn updates to source tree? Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
On Thu, Aug 06, 2015 at 07:49:56AM +, Thomas Mueller wrote: T> > I ended with doing the aforementioned 2 changes as one. T> T> > The diff is living here: T> T> > https://reviews.freebsd.org/D2655 T> T> > It spans the net80211 layer and converts (almost) all 802.11 drivers T> > to new KPI. Now it also converts if_ndis, and this is the change I T> > am asking you to test. T> T> > Unlike all other drivers, converting if_ndis wasn't trivial. There are T> > likely bugs introduced, so it may happen that we will need several T> > rounds of testing. T> T> What do I need to download? There is a lot of stuff on the page you cite, and I see a link to the right, Download Raw Diff. Is that what I'm supposed to download? Correct, the link "Download Raw Diff". T> What do you want me to test it on? T> T> I have Hiro H50191 USB WLAN adapter, including .inf and .sys files. T> T> I also have, on MSI MPOWER motherboard, Atheros AR9271 quasi-USB WLAN, and Realtek 8111E Ethernet, with MS-Windows drivers but no .INF and .SYS files on DVD that came with the motherboard. T> T> Is there a way to convert MS-Windows drivers that come with no .inf and .sys files to ndis-compatible format? Or just test on the Hiro H50191? Testing on Hiro would be fine. T> Also, after applying the patch to the source tree, what if anything do I have to do to not be messed up for subsequent svn updates to source tree? If you want to revert your svn tree to pristine state, you need to say in the root of the tree: svn revert --depth=infinity . -- Totus tuus, Glebius. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> I ended with doing the aforementioned 2 changes as one. > The diff is living here: > https://reviews.freebsd.org/D2655 > It spans the net80211 layer and converts (almost) all 802.11 drivers > to new KPI. Now it also converts if_ndis, and this is the change I > am asking you to test. > Unlike all other drivers, converting if_ndis wasn't trivial. There are > likely bugs introduced, so it may happen that we will need several > rounds of testing. > Totus tuus, Glebius. What do I need to download? There is a lot of stuff on the page you cite, and I see a link to the right, Download Raw Diff. Is that what I'm supposed to download? What do you want me to test it on? I have Hiro H50191 USB WLAN adapter, including .inf and .sys files. I also have, on MSI MPOWER motherboard, Atheros AR9271 quasi-USB WLAN, and Realtek 8111E Ethernet, with MS-Windows drivers but no .INF and .SYS files on DVD that came with the motherboard. Is there a way to convert MS-Windows drivers that come with no .inf and .sys files to ndis-compatible format? Or just test on the Hiro H50191? Also, after applying the patch to the source tree, what if anything do I have to do to not be messed up for subsequent svn updates to source tree? Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"
Re: CFT: ndis(4) testers needed!
> I need to make couple of non-functional but rather large changes > to the if_ndis driver and will appreciate if anyone signs up to > test my changes. Please contact me if you can provide help. > Totus tuus, Glebius. I have two possible candidates for testing: Hiro H50191 USB wi-fi adapter, but this seems to work as rsu in FreeBSD-current but not 10-STABLE. Atheros AR9271 on-motherboard quasi-USB wi-fi, sometimes works in NetBSD-current but most of the time fails to load firmware. I have .inf and .sys files for Hiro but not for Atheros AR9271 ("install" under wine and get .inf and .sys that way?). Are these suitable for testing ndis? My Realtek re(4) Ethernet now works with recent updates in current, I think also 10-STABLE but have to check that again. I only sporadically do anything with FreeBSD 10-STABLE, nothing with FreeBSD < 10.0. Tom ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to "freebsd-current-unsubscr...@freebsd.org"