On my Toshiba Portege M400, "echo crt_out:1 > /proc/acpi/toshiba/video" does not enable external video output. Tested on gentoo-sources-2.6.32-r7. This might be due to the issue described here: http://memebeam.org/toys/ToshibaAcpiVideo (I have disassembled my ACPI DSDT and confirmed that the \_SB_.VALX.DSSX method is not defined). It used to be possible to enable output to an external monitor using the toshset utility, but recent kernels (tested on gentoo-sources-2.6.32-r7) break compatibility with toshset. Despite the note about the patch being included in the mainline kernel, the patch at http://memebeam.org/free-software/toshiba_acpi/ was nonetheless still necessary to get toshset to work on older kernels, e.g. gentoo-sources-2.6.29-r6. However, it no longer applies cleanly to more recent kernels (tested on gentoo-sources-2.6.32-r7), and though it can be made to apply cleanly again with a minor edit, this led to a kernel oops when inserting the toshiba_acpi module. Reproducible: Always Steps to Reproduce: 1. Acquire a Toshiba Portege M400 laptop (there are reports of this also affecting other Toshiba laptops, but I don't have access to any other models) 2. Install gentoo-sources-2.6.32-r7 3. Configure kernel with CONFIG_TOSHIBA=m and CONFIG_ACPI_TOSHIBA=m 4. Compile and install kernel 5. Boot into new kernel 6a. echo lcd_out:1;crt_out:1 > /proc/acpi/toshiba/video 6b. toshset -video on Actual Results: External video output remains disabled (step 6a). # echo lcd_out:1;crt_out:1 > /proc/acpi/toshiba/video # cat /proc/acpi/toshiba/video lcd_out: 1 crt_out: 0 tv_out: 0 Toshset utility reports that computer is not supported (step 6b): # toshset -video on toshset: this computer is not supported Expected Results: External video output should be enabled. # uname -a Linux c3po 2.6.32-gentoo-r7 #1 SMP Thu May 6 12:30:29 BST 2010 i686 Intel(R) Core(TM)2 CPU T7200 @ 2.00GHz GenuineIntel GNU/Linux # paludis -q gentoo-sources * sys-kernel/gentoo-sources gentoo: 2.6.16-r13 {:2.6.16-r13} 2.6.25-r9 {:2.6.25-r9} 2.6.26-r4 {:2.6.26-r4} 2.6.27-r8 {:2.6.27-r8} 2.6.27-r10 {:2.6.27-r10} 2.6.28-r5 {:2.6.28-r5} 2.6.28-r6 {:2.6.28-r6} 2.6.29-r5 {:2.6.29-r5} 2.6.29-r6 {:2.6.29-r6} 2.6.30-r8 {:2.6.30-r8} (2.6.30-r9)K {:2.6.30-r9} 2.6.31-r10 {:2.6.31-r10} (2.6.32-r6)K {:2.6.32-r6} 2.6.32-r7 {:2.6.32-r7} (2.6.32-r8)K {:2.6.32-r8} (2.6.33)K {:2.6.33} (2.6.33-r1)K {:2.6.33-r1} (2.6.33-r2)K {:2.6.33-r2} installed: 2.6.29-r6* {:2.6.29-r6} Description: Full sources including the Gentoo patchset for the 2.6 kernel tree Homepage: http://dev.gentoo.org/~mpagano/genpatches License: GPL-2 !deblob? ( freedist ) Installed time: Thu May 06 11:28:37 BST 2010 Use flags: (-build) (-deblob) (-symlink) Build Options: -trace From repositories: gentoo Installed using: paludis-0.46.0
Created attachment 230593 [details] Kernel config
Thanks for the report, i'm forwarding this to the kernel team.
It sounds like toshset needs an update
This is not a gentoo kernel issue. I suggest taking this to the toshet upstream developers. http://www.schwieters.org/toshset/
I reported this upstream to the toshset dev. It seems an experimental version of the toshiba_acpi driver is required for toshset to work, but the latest available version of that patch doesn't apply cleanly to recent kernels. However, the changes required to get the patch to apply cleanly to 2.6.34 were trivial. After modifying the patch, applying it, and recompiling the kernel, I now have toshset working under tuxonice-sources-2.6.34. (I've also tested that the patch at least applies cleanly to vanilla-sources-2.6.34, though I haven't tried compiling or running that one.) I'm posting the updated patch here for reference. It should also be available soon from the toshset website linked above.
Created attachment 237071 [details, diff] Experimental toshiba_acpi patch required by toshset.
BTW, the toshiba_acpi patch required by toshset was apparently rejected by the kernel devs over a year ago, on the basis that they would support Toshiba hardware directly in toshiba_acpi without needing toshset at all. Some things *are* now supported natively (e.g. bluetooth...more or less), but other important hardware features such as external video output are *still* broken in the latest kernels, and can only be made to work with toshset. Despite the patch allowing toshset to work again, this issue ought nonetheless to be reported upstream to the kernel devs. It only affects certain Toshiba laptops, so they may not even be aware that there is still work to do on toshiba_acpi before it covers all Toshiba hardware and toshset can be mothballed. Would it be better for the Gentoo devs to upstream this bug report to the kernel devs, or shall I do it myself?
(In reply to comment #7) > BTW, the toshiba_acpi patch required by toshset was apparently rejected by the > kernel devs over a year ago, on the basis that they would support Toshiba > hardware directly in toshiba_acpi without needing toshset at all. Some things > *are* now supported natively (e.g. bluetooth...more or less), but other > important hardware features such as external video output are *still* broken in > the latest kernels, and can only be made to work with toshset. > > Despite the patch allowing toshset to work again, this issue ought nonetheless > to be reported upstream to the kernel devs. It only affects certain Toshiba > laptops, so they may not even be aware that there is still work to do on > toshiba_acpi before it covers all Toshiba hardware and toshset can be > mothballed. > > Would it be better for the Gentoo devs to upstream this bug report to the > kernel devs, or shall I do it myself? > The policy for gentoo-sources is to reject patches that have been rejected from upstream developers. You need to open a bug upstream ( or poke the toshset developer to do it ) and paste the link here so we can track the progress Thank you
(In reply to comment #8) > (In reply to comment #7) > > Despite the patch allowing toshset to work again, this issue ought nonetheless > > to be reported upstream to the kernel devs. It only affects certain Toshiba > > laptops, so they may not even be aware that there is still work to do on > > toshiba_acpi before it covers all Toshiba hardware and toshset can be > > mothballed. > > > > Would it be better for the Gentoo devs to upstream this bug report to the > > kernel devs, or shall I do it myself? > > > > The policy for gentoo-sources is to reject patches that have been rejected from > upstream developers. You need to open a bug upstream ( or poke the toshset > developer to do it ) and paste the link here so we can track the progress You've misunderstood what I wrote. I'm *not* suggesting the patch should be pushed upstream. It's already been rejected by the kernel devs. They don't want to support toshset. End of story. I'm saying that *bug report* about broken external video output should be reported upstream, so that the kernel devs are at least aware of the fact that there's still work to be done before Toshiba laptops are fully supported. My understanding was that bugs in software installed via Gentoo should be reported first to Gentoo, so that Gentoo devs can upstream the bug report if appropriate. The lack of external video output support on certain Toshiba laptop models is an issue that can only be addressed by fixing the kernel. The kernel devs have already attempted to get this working via /proc/acpi/video, but this remains broken on some laptop models. They should at least be made aware of this, otherwise how will they know to fix it? It therefore seems appropriate to report it upstream. However, I'm happy to open a kernel bug myself, if I've misunderstood how the process is supposed to work.
