I have an HP SFF computer which has a problem with the DVD drive. Computer details: sdmesg | grep DMI: [ 0.000000] DMI: HP HP 290 G3 Small Form Factor PC/8768, BIOS F.31 03/09/2022 I am running latest version Manjaro on xfce. There is no power to the DVD unit - the drawer will not open. I use this computer for regression testing and I have a number of different operating systems on it. There is nothing wrong with the DVD drive as it works with some kernels and not others. I am running the latest BIOS. Here is a summary of what I have found. On my SDD: Windows OK Manjaro partition using kernel 5.4.201 OK Manjaro partition using kernel 5.17.9 NOT OK Manjaro partition using kernel 5.18.7.1 NOT OK Manjaro partition using kernel 5.19.1.2 NOT OK On my HDD: MX linux using kernel 5.10.70-1 OK Ubuntu using kernel 5.4.0.73 OK Old Manjaro backup partition #1 using kernel 5.15.7 OK Old Manjaro backup partition #2 using kernel 5.16.2.2 NOT OK So kernels 5.16 through 5.19 are affected. On 5.4 kernel inxi --optical gives Optical-1: /dev/sr0 vendor: hp PLDS model: DVDRW DU8AESH dev-links: cdrom Features: speed: 24 multisession: yes audio: yes dvd: yes rw: cd-r,cd-rw,dvd-r On 5.16/17/18/19 inxi --optical gives Message: No optical or floppy data found. cat /proc/scsi/scsi on 5.4 kernel Host: scsi2 Channel: 00 Id: 00 Lun: 00 Vendor: hp PLDS Model: DVDRW DU8AESH Rev: 6HSM Type: CD-ROM ANSI SCSI revision: 05 on 5.16 to 5.19 kernels, this entry does not appear. sudo desmg|grep DVD on 5.4 gives [ 1.512894] ata3.00: ATAPI: hp PLDS DVDRW DU8AESH, 6HSM, max UDMA/133 [ 1.527247] scsi 2:0:0:0: CD-ROM hp PLDS DVDRW DU8AESH 6HSM PQ: 0 ANSI: 5 Nothing on 5.18 On 5.18 kernel sudo dmesg |grep ata3 [ 0.494741] ata3: SATA max UDMA/133 abar m2048@0xb122b000 port 0xb122b200 irq 124 [ 0.805172] ata3: SATA link down (SStatus 0 SControl 300) on 5.4 kernel sudo dmesg |grep ata3 [ 1.158317] ata3: SATA max UDMA/133 abar m2048@0xb122b000 port 0xb122b200 irq 134 [ 1.472137] ata3: SATA link up 1.5 Gbps (SStatus 113 SControl 300) [ 1.481428] ata3.00: ATAPI: hp PLDS DVDRW DU8AESH, 6HSM, max UDMA/133 [ 1.487901] ata3.00: configured for UDMA/133 I lodged a query with the Manjaro forum . see https://forum.manjaro.org/t/dvd-drive-not-recognised-by-kernels-5-16-5-17-and-5-18/116386/16 Their best guess was that the problem was that an updated ACPI table was required and that this would only be achieved by a kernel update. This is not a big issue as I can use an old kernel if I want to use the optical drive. I am only trying to ensure that this bug does not show up in the next LTS. I am a user and do not have the capability to compile kernels.
Could you please attach the dmesg output of a 5.15 kernel and a later one (ideally 5.16) to this bug?
Created attachment 301665 [details] 5.15dmesgDVDOK Hi there, Thanks for looking at my problem. 2 dmesg outputs attached from kernels 5.15 and 5.16 with file names denoting which is which. Just to confirm, 5.15 DVD is OK. 5.16, no power on DVD. Best wishes, Bruce Healey On 25/08/2022 12:03, bugzilla-daemon@kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=216360 > > The Linux kernel's regression tracker (Thorsten Leemhuis) > (regressions@leemhuis.info) changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > CC| |regressions@leemhuis.info > > --- Comment #1 from The Linux kernel's regression tracker (Thorsten Leemhuis) > (regressions@leemhuis.info) --- > Could you please attach the dmesg output of a 5.15 kernel and a later one > (ideally 5.16) to this bug? >
Created attachment 301666 [details] 5.16dmesgDVDNOPOWER
(In reply to Bruce Healey from comment #3) > Created attachment 301666 [details] > 5.16dmesgDVDNOPOWER One quick further question: from the kernel name used here it looks a lot like this kernel is patched with the preempt_RT patch-set while the other is not. Have you tried a vanilla kernel, too?
and are the 5.16+ kernels all build from a similar configuration as the 5.15 kernel? `make oldconfig` is your friend here. For details see https://docs.kernel.org/admin-guide/reporting-regressions.html
All kernels I have used are loaded from the Manjaro site loaded using Pamac. I do not perform any sort of patching or kernel modification. All kernels are vanilla. I am a kernel user and in no way a kernel developed. All my work is at the application level. Any variations in kernel are from Manjaro and not me. On 26/08/2022 06:51, bugzilla-daemon@kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=216360 > > --- Comment #4 from The Linux kernel's regression tracker (Thorsten Leemhuis) > (regressions@leemhuis.info) --- > (In reply to Bruce Healey from comment #3) >> Created attachment 301666 [details] >> 5.16dmesgDVDNOPOWER > One quick further question: from the kernel name used here it looks a lot > like > this kernel is patched with the preempt_RT patch-set while the other is not. > Have you tried a vanilla kernel, too? >
As with the response to question #4 all kernels are vanilla Manjaro On 26/08/2022 06:53, bugzilla-daemon@kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=216360 > > --- Comment #5 from The Linux kernel's regression tracker (Thorsten Leemhuis) > (regressions@leemhuis.info) --- > and are the 5.16+ kernels all build from a similar configuration as the 5.15 > kernel? `make oldconfig` is your friend here. For details see > https://docs.kernel.org/admin-guide/reporting-regressions.html >
(In reply to Bruce Healey from comment #7) > As with the response to question #4 all kernels are vanilla Manjaro > > On 26/08/2022 06:53, bugzilla-daemon@kernel.org wrote: > > https://bugzilla.kernel.org/show_bug.cgi?id=216360 > > > > --- Comment #5 from The Linux kernel's regression tracker (Thorsten > Leemhuis) > > (regressions@leemhuis.info) --- > > and are the 5.16+ kernels all build from a similar configuration as the > 5.15 > > kernel? `make oldconfig` is your friend here. For details see > > https://docs.kernel.org/admin-guide/reporting-regressions.html > > This problem is persisting with kernel 6.0.0.1. Is there any resolution to this? I would like to point out that my kernel 5.16 was a 'real time kernel' vanilla from Manjaro. All later kernels 5.17, 1.18, 5.19 and 6.0 are vanilla unmodified Manjaro kernels. Manjaro do not have seem to permit bugs to be logged against the kernel so I have nowhere else to go on this. The last advice was to 'make oldconfig' - how does this apply? I have not built any kernels myself ever. I am a kernel user not a developer. My DVD unit used to work straight away with an out of the box kernel and now it doesn't although it still works with older kernels and Windows. I don't understand whether this thread has been dropped or I am supposed to do something.
(In reply to Bruce Healey from comment #8) > This problem is persisting with kernel 6.0.0.1. Do you mean 6.0.1? If not: There is no 6.0.0.1 at kernel.org, might thus be a distro-kernel. > Is there any resolution to this? Pretty unlikely to happen, unless someone performs a git bisect or finds the root cause by different means (but that is even harder). Sorry, that's just how it is. Regressions like this shouldn't happen, but humans make errors when working on code. :-( This seem to have happen here. And this error apparently only affects your setup (maybe it's the kernel, maybe its config or something), otherwise somebody else likely would have found and report the problem. > I have not built any kernels myself ever. It's not that hard, sadly we don't have docs for this (yet -- it's pretty high on my todo list). But there are tons of howtos on the net; some even mention how to perform a bisection. See for example: https://wiki.debian.org/DebianKernel/GitBisect https://wiki.gentoo.org/wiki/Kernel_git-bisect
(In reply to The Linux kernel's regression tracker (Thorsten Leemhuis) from comment #9) > (In reply to Bruce Healey from comment #8) > > > This problem is persisting with kernel 6.0.0.1. > > Do you mean 6.0.1? If not: There is no 6.0.0.1 at kernel.org, might thus be > a distro-kernel. > > > Is there any resolution to this? > > Pretty unlikely to happen, unless someone performs a git bisect or finds the > root cause by different means (but that is even harder). > > Sorry, that's just how it is. Regressions like this shouldn't happen, but > humans make errors when working on code. :-( This seem to have happen here. > And this error apparently only affects your setup (maybe it's the kernel, > maybe its config or something), otherwise somebody else likely would have > found and report the problem. > > > I have not built any kernels myself ever. > > It's not that hard, sadly we don't have docs for this (yet -- it's pretty > high on my todo list). But there are tons of howtos on the net; some even > mention how to perform a bisection. See for example: > > https://wiki.debian.org/DebianKernel/GitBisect > https://wiki.gentoo.org/wiki/Kernel_git-bisect Is there anything you would like me do to try and resolve this problem? Please make any instructions clear enough for someone who has never worked on a kernel before.
I have additional information about this problem. I still have no power to my DVD drive with Manjaro kernels 6.0, 6.1, 6.2, 6.3 and 6.4. I do have two other distros loaded which I have recently updated. DEBIAN 12 kernel 5.10 - DVD drive powered kernel 6.1 - No power to DVD drive MX 21 kernel 5.10 - DVD drive powered kernel 6.0 - No power to DVD drive kernel 6.1 - No power to DVD drive So Debian and MX identically follow the observations I have seen on Manjaro. Please note all kernels are downloaded unmodified using pamac or synaptic according to the distro - I have not done anything to any of them.
(In reply to Bruce Healey from comment #10) > (In reply to The Linux kernel's regression tracker (Thorsten Leemhuis) from > comment #9) > > (In reply to Bruce Healey from comment #8) > > > > > This problem is persisting with kernel 6.0.0.1. > > > > Do you mean 6.0.1? If not: There is no 6.0.0.1 at kernel.org, might thus be > > a distro-kernel. > > > > > Is there any resolution to this? > > > > Pretty unlikely to happen, unless someone performs a git bisect or finds > the > > root cause by different means (but that is even harder). > > > > Sorry, that's just how it is. Regressions like this shouldn't happen, but > > humans make errors when working on code. :-( This seem to have happen here. > > And this error apparently only affects your setup (maybe it's the kernel, > > maybe its config or something), otherwise somebody else likely would have > > found and report the problem. > > > > > I have not built any kernels myself ever. > > > > It's not that hard, sadly we don't have docs for this (yet -- it's pretty > > high on my todo list). But there are tons of howtos on the net; some even > > mention how to perform a bisection. See for example: > > > > https://wiki.debian.org/DebianKernel/GitBisect > > https://wiki.gentoo.org/wiki/Kernel_git-bisect > > Is there anything you would like me do to try and resolve this problem? > Please make any instructions clear enough for someone who has never worked > on a kernel before. The prerequisite for bisection is you need to successfully installed vanilla kernel. To build it, see instructions on ArchWiki (since Manjaro is derived from Arch) [1]. Then follow one of two bisection howtos above (although these are distro-specific; you can omit any distro-specific steps there such as installing build dependencies). [1]: https://wiki.archlinux.org/title/Kernel/Traditional_compilation
Apologies for the inordinate delay in getting back to you on this one. The DVD drawer is powered for all releases of kernel 5.15 including 5.15.138. The first kernel to not power the DVD drawer is 5.16.1 Please let me know if there is additional information I can supply.
(In reply to Bruce Healey from comment #13) > Please let me know if there is additional information I can supply. As mentioned in earlier comments already: please perform a bisection using "git bisect" between 5.15 and 5.16 (with a config derived from a working 5.15 kernel). Without one it's unlikely that anyone will look into this.
Excuse me??? I thought the information you wanted was for me to find which release of kernel the bug first appeared. AS STATED ABOVE the last kernel to work properly was 5.15.138. As this was the very last kernel in the 5.15 series the very next kernel was 5.16.1 which now has the problem. I was not aware that you wanted a finer level of detail than that.
(In reply to Bruce Healey from comment #15) > Excuse me??? > I thought the information you wanted was for me to find which release of > kernel the bug first appeared. AS STATED ABOVE the last kernel to work > properly was 5.15.138. As this was the very last kernel in the 5.15 series > the very next kernel was 5.16.1 which now has the problem. > I was not aware that you wanted a finer level of detail than that. I repeat what I and Thorsten already said: please *DO* bisection. Then, check if latest mainline (currently v6.6) still exhibits your regression.
I have had some time to get back to this and have got stuck. 5.15.147 is the final release where the DVD drawer operates. 5.16.1 is the first where it does not. You have been very insistent that I should perform a GIT BISECT for a reason I don't understand. However, I have attempted to do this with little success. Here are the commands I entered. $ git clone https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ Cloning into 'linux'... remote: Enumerating objects: 12158717, done. remote: Counting objects: 100% (11350/11350), done. remote: Compressing objects: 100% (4825/4825), done. remote: Total 12158717 (delta 9001), reused 7787 (delta 6509), pack-reused 12147367 Receiving objects: 100% (12158717/12158717), 4.83 GiB | 3.63 MiB/s, done. Resolving deltas: 100% (9706993/9706993), done. Updating files: 100% (83463/83463), done. $ git bisect start $ git bisect good v5.15.147 $ git bisect bad v5.16.1 Bisecting: a merge base must be tested [8bb7eca972ad531c9b149c0a51ab43a417385813] Linux 5.15 I have no idea what this is telling me and I cannot find any idiot level documentation online to get me any further. I have no idea how to create a merge base, no idea what the hex number is and no idea why GIT BISECT is pointing to 5.15 when clearly the problem is in 5.16. Can you point me at any documentation that would help me get any further?
On 1/27/24 22:14, bugzilla-daemon@kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=216360 > > --- Comment #17 from Bruce Healey (bhealey@live.co.uk) --- > I have had some time to get back to this and have got stuck. 5.15.147 is the > final release where the DVD drawer operates. 5.16.1 is the first where it > does > not. You have been very insistent that I should perform a GIT BISECT for a > reason I don't understand. However, I have attempted to do this with little > success. > Here are the commands I entered. > $ git clone https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ > Cloning into 'linux'... > remote: Enumerating objects: 12158717, done. > remote: Counting objects: 100% (11350/11350), done. > remote: Compressing objects: 100% (4825/4825), done. > remote: Total 12158717 (delta 9001), reused 7787 (delta 6509), pack-reused > 12147367 > Receiving objects: 100% (12158717/12158717), 4.83 GiB | 3.63 MiB/s, done. > Resolving deltas: 100% (9706993/9706993), done. > Updating files: 100% (83463/83463), done. > $ git bisect start > $ git bisect good v5.15.147 > $ git bisect bad v5.16.1 > Bisecting: a merge base must be tested > [8bb7eca972ad531c9b149c0a51ab43a417385813] Linux 5.15 > > I have no idea what this is telling me and I cannot find any idiot level > documentation online to get me any further. I have no idea how to create a > merge base, no idea what the hex number is and no idea why GIT BISECT is > pointing to 5.15 when clearly the problem is in 5.16. Can you point me at any > documentation that would help me get any further? > tl;dr: Thorsten has a bisection guide [1]. Please have a look and follow the instructions. Bye! [1]: https://lore.kernel.org/linux-doc/c763e15e-e82e-49f8-a540-d211d18768a3@leemhuis.info/
Please mark this bug as CLOSED: UNRESOLVED. I have been unable to find a procedure using GIT Bisect which tells me how to compile a kernel to a particular COMMIT which is at a level I can use. The bug is in kernel 5.16.0 and appears in every kernel since - I have tried every major release. I will use windows or kernel 5.15 if I need to use my DVD drive.
On 5/28/24 21:38, bugzilla-daemon@kernel.org wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=216360 > > --- Comment #19 from Bruce Healey (bhealey@live.co.uk) --- > Please mark this bug as CLOSED: UNRESOLVED. > I have been unable to find a procedure using GIT Bisect which tells me how to > compile a kernel to a particular COMMIT which is at a level I can use. The > bug > is in kernel 5.16.0 and appears in every kernel since - I have tried every > major release. I will use windows or kernel 5.15 if I need to use my DVD > drive. > tl;dr: read and follow **carefully** [1]. [1]: https://docs.kernel.org/admin-guide/verify-bugs-and-bisect-regressions.html