Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
On Wed, Oct 09, 2019 at 09:03:25AM -0400, Dave Anderson wrote: > > > - Original Message - > > On Tue, Oct 08, 2019 at 04:55:48PM -0400, Dave Anderson wrote: > > > > > > > - Original Message - > > > > > Hi Dave, > > > > > > > > > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > > > > > > > > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. > > > > >Does anyone work on fixing it on kernel side? > > > > >(just adding one or two lines though) > > > > > > > > > > 2. With a tweak above, I still fail to run crash with vmcore > > > > >seeing the following errors; > > > > >(I didn't dig into details) > > > > > > > > > >crash-arm64: invalid kernel virtual address: 800010c689a8 > > > > > type: > > > > >"kernel_config_data" > > > > >WARNING: cannot read kernel_config_data > > > > >crash-arm64: invalid kernel virtual address: 80001177a160 > > > > > type: > > > > >"possible" > > > > >WARNING: cannot read cpu_possible_map > > > > >crash-arm64: invalid kernel virtual address: 80001177a140 > > > > > type: > > > > >"present" > > > > >WARNING: cannot read cpu_present_map > > > > >crash-arm64: invalid kernel virtual address: 80001177a120 > > > > > type: > > > > >"online" > > > > >WARNING: cannot read cpu_online_map > > > > >crash-arm64: invalid kernel virtual address: 80001177a188 > > > > > type: > > > > >"active" > > > > >WARNING: cannot read cpu_active_map > > > > >crash-arm64: invalid kernel virtual address: 80001196afe0 > > > > > type: > > > > >"shadow_timekeeper xtime_sec" > > > > >crash-arm64: invalid kernel virtual address: 800011782b3c > > > > > type: > > > > >"init_uts_ns" > > > > >WARNING: invalid linux_banner pointer: 800010c500e8 > > > > >crash: vmlinux.1003 and vmcore.1003 do not match! > > > > > > > > > > Have you noticed this issue? > > > > > > > > I have been waiting for our internal 5.4-rc1 based ARK kernel to be > > > > built, > > > > and it looks like it should be available soon. But I did not expect > > > > crash > > > > to work on arm64 in 5.4. > > > > > > > > I put in a bunch of 52-bit support into the crash utility last spring, > > > > but > > > > it was based upon an early proposed patch-set from Steve Capper that > > > > Chris > > > > von Recklinghausen backported to our 4.18.0-80.el8 kernel at the time. > > > > I believe Chris may have added the VA_BITS_ACTUAL vmcoreinfo entry (?), > > > > and I'm presuming there will be other VM-related things that are > > > > different from that original kernel patch. > > > > > > > > Thanks, > > > > Dave > > > > > > I still haven't got a 5.4-rc1 kernel to test with, but Chris Von > > > Recklinghausen > > > has backported the latest upstream 52-bit patch-set to a 4.18-based > > > kernel. > > > However, it did require one additional patch needed for crash to run live > > > on /proc/kcore, which he posted on the arm list today: > > > > > > [PATCH] arm64: remove arm64 definitions of kc_vaddr_to_offset and > > > kc_offset_to_vaddr. > > > > > > http://lists.infradead.org/pipermail/linux-arm-kernel/2019-October/685109.html > > > > > > With his patch-set (including the above), crash-7.2.7 comes up unmodified > > > on > > > a live system -- albeit one which downgrades from 52 bits to 48 bits. > > > Chris is > > > going to test crash running live on a 52 bit simulator tomorrow. > > > > I may misunderstand something, but even after applying this patch, > > nothing has changed, at least, against /proc/vmcore. (I see the same > > errors.) > > The patch has nothing to do with /proc/vmcore -- did you mean /proc/kcore? That's why I said that I might misunderstand something. I have concerns on /proc/vmcore as I'm now pushing my kexec_file_load- based kdump patch. Thanks, -Takahiro Akashi > Dave > > > > > > Thanks, > > -Takahiro Akashi > > > > > > > Dave > > > > > > -- > > > Crash-utility mailing list > > > Crash-utility@redhat.com > > > https://www.redhat.com/mailman/listinfo/crash-utility > > > > -- > > Crash-utility mailing list > > Crash-utility@redhat.com > > https://www.redhat.com/mailman/listinfo/crash-utility > > > > -- > Crash-utility mailing list > Crash-utility@redhat.com > https://www.redhat.com/mailman/listinfo/crash-utility -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
- Original Message - > On Tue, Oct 08, 2019 at 04:55:48PM -0400, Dave Anderson wrote: > > > > > - Original Message - > > > > Hi Dave, > > > > > > > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > > > > > > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. > > > >Does anyone work on fixing it on kernel side? > > > >(just adding one or two lines though) > > > > > > > > 2. With a tweak above, I still fail to run crash with vmcore > > > >seeing the following errors; > > > >(I didn't dig into details) > > > > > > > >crash-arm64: invalid kernel virtual address: 800010c689a8 type: > > > >"kernel_config_data" > > > >WARNING: cannot read kernel_config_data > > > >crash-arm64: invalid kernel virtual address: 80001177a160 type: > > > >"possible" > > > >WARNING: cannot read cpu_possible_map > > > >crash-arm64: invalid kernel virtual address: 80001177a140 type: > > > >"present" > > > >WARNING: cannot read cpu_present_map > > > >crash-arm64: invalid kernel virtual address: 80001177a120 type: > > > >"online" > > > >WARNING: cannot read cpu_online_map > > > >crash-arm64: invalid kernel virtual address: 80001177a188 type: > > > >"active" > > > >WARNING: cannot read cpu_active_map > > > >crash-arm64: invalid kernel virtual address: 80001196afe0 type: > > > >"shadow_timekeeper xtime_sec" > > > >crash-arm64: invalid kernel virtual address: 800011782b3c type: > > > >"init_uts_ns" > > > >WARNING: invalid linux_banner pointer: 800010c500e8 > > > >crash: vmlinux.1003 and vmcore.1003 do not match! > > > > > > > > Have you noticed this issue? > > > > > > I have been waiting for our internal 5.4-rc1 based ARK kernel to be > > > built, > > > and it looks like it should be available soon. But I did not expect > > > crash > > > to work on arm64 in 5.4. > > > > > > I put in a bunch of 52-bit support into the crash utility last spring, > > > but > > > it was based upon an early proposed patch-set from Steve Capper that > > > Chris > > > von Recklinghausen backported to our 4.18.0-80.el8 kernel at the time. > > > I believe Chris may have added the VA_BITS_ACTUAL vmcoreinfo entry (?), > > > and I'm presuming there will be other VM-related things that are > > > different from that original kernel patch. > > > > > > Thanks, > > > Dave > > > > I still haven't got a 5.4-rc1 kernel to test with, but Chris Von > > Recklinghausen > > has backported the latest upstream 52-bit patch-set to a 4.18-based kernel. > > However, it did require one additional patch needed for crash to run live > > on /proc/kcore, which he posted on the arm list today: > > > > [PATCH] arm64: remove arm64 definitions of kc_vaddr_to_offset and > > kc_offset_to_vaddr. > > > > http://lists.infradead.org/pipermail/linux-arm-kernel/2019-October/685109.html > > > > With his patch-set (including the above), crash-7.2.7 comes up unmodified > > on > > a live system -- albeit one which downgrades from 52 bits to 48 bits. Chris > > is > > going to test crash running live on a 52 bit simulator tomorrow. > > I may misunderstand something, but even after applying this patch, > nothing has changed, at least, against /proc/vmcore. (I see the same errors.) The patch has nothing to do with /proc/vmcore -- did you mean /proc/kcore? Dave > > Thanks, > -Takahiro Akashi > > > > Dave > > > > -- > > Crash-utility mailing list > > Crash-utility@redhat.com > > https://www.redhat.com/mailman/listinfo/crash-utility > > -- > Crash-utility mailing list > Crash-utility@redhat.com > https://www.redhat.com/mailman/listinfo/crash-utility > -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
On Tue, Oct 08, 2019 at 04:55:48PM -0400, Dave Anderson wrote: > > > - Original Message - > > > Hi Dave, > > > > > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > > > > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. > > >Does anyone work on fixing it on kernel side? > > >(just adding one or two lines though) > > > > > > 2. With a tweak above, I still fail to run crash with vmcore > > >seeing the following errors; > > >(I didn't dig into details) > > > > > >crash-arm64: invalid kernel virtual address: 800010c689a8 type: > > > "kernel_config_data" > > >WARNING: cannot read kernel_config_data > > >crash-arm64: invalid kernel virtual address: 80001177a160 type: > > > "possible" > > >WARNING: cannot read cpu_possible_map > > >crash-arm64: invalid kernel virtual address: 80001177a140 type: > > > "present" > > >WARNING: cannot read cpu_present_map > > >crash-arm64: invalid kernel virtual address: 80001177a120 type: > > > "online" > > >WARNING: cannot read cpu_online_map > > >crash-arm64: invalid kernel virtual address: 80001177a188 type: > > > "active" > > >WARNING: cannot read cpu_active_map > > >crash-arm64: invalid kernel virtual address: 80001196afe0 type: > > > "shadow_timekeeper xtime_sec" > > >crash-arm64: invalid kernel virtual address: 800011782b3c type: > > > "init_uts_ns" > > >WARNING: invalid linux_banner pointer: 800010c500e8 > > >crash: vmlinux.1003 and vmcore.1003 do not match! > > > > > > Have you noticed this issue? > > > > I have been waiting for our internal 5.4-rc1 based ARK kernel to be built, > > and it looks like it should be available soon. But I did not expect crash > > to work on arm64 in 5.4. > > > > I put in a bunch of 52-bit support into the crash utility last spring, but > > it was based upon an early proposed patch-set from Steve Capper that Chris > > von Recklinghausen backported to our 4.18.0-80.el8 kernel at the time. > > I believe Chris may have added the VA_BITS_ACTUAL vmcoreinfo entry (?), > > and I'm presuming there will be other VM-related things that are > > different from that original kernel patch. > > > > Thanks, > > Dave > > I still haven't got a 5.4-rc1 kernel to test with, but Chris Von > Recklinghausen > has backported the latest upstream 52-bit patch-set to a 4.18-based kernel. > However, it did require one additional patch needed for crash to run live > on /proc/kcore, which he posted on the arm list today: > > [PATCH] arm64: remove arm64 definitions of kc_vaddr_to_offset and > kc_offset_to_vaddr. > > http://lists.infradead.org/pipermail/linux-arm-kernel/2019-October/685109.html > > With his patch-set (including the above), crash-7.2.7 comes up unmodified on > a live system -- albeit one which downgrades from 52 bits to 48 bits. Chris > is > going to test crash running live on a 52 bit simulator tomorrow. I may misunderstand something, but even after applying this patch, nothing has changed, at least, against /proc/vmcore. (I see the same errors.) Thanks, -Takahiro Akashi > Dave > > -- > Crash-utility mailing list > Crash-utility@redhat.com > https://www.redhat.com/mailman/listinfo/crash-utility -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
> - Original Message - > > Hi Dave, > > > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. > >Does anyone work on fixing it on kernel side? > >(just adding one or two lines though) > > > > 2. With a tweak above, I still fail to run crash with vmcore > >seeing the following errors; > >(I didn't dig into details) > > > >crash-arm64: invalid kernel virtual address: 800010c689a8 type: > > "kernel_config_data" > >WARNING: cannot read kernel_config_data > >crash-arm64: invalid kernel virtual address: 80001177a160 type: > > "possible" > >WARNING: cannot read cpu_possible_map > >crash-arm64: invalid kernel virtual address: 80001177a140 type: > > "present" > >WARNING: cannot read cpu_present_map > >crash-arm64: invalid kernel virtual address: 80001177a120 type: > > "online" > >WARNING: cannot read cpu_online_map > >crash-arm64: invalid kernel virtual address: 80001177a188 type: > > "active" > >WARNING: cannot read cpu_active_map > >crash-arm64: invalid kernel virtual address: 80001196afe0 type: > > "shadow_timekeeper xtime_sec" > >crash-arm64: invalid kernel virtual address: 800011782b3c type: > > "init_uts_ns" > >WARNING: invalid linux_banner pointer: 800010c500e8 > >crash: vmlinux.1003 and vmcore.1003 do not match! > > > > Have you noticed this issue? > > I have been waiting for our internal 5.4-rc1 based ARK kernel to be built, > and it looks like it should be available soon. But I did not expect crash > to work on arm64 in 5.4. > > I put in a bunch of 52-bit support into the crash utility last spring, but > it was based upon an early proposed patch-set from Steve Capper that Chris > von Recklinghausen backported to our 4.18.0-80.el8 kernel at the time. > I believe Chris may have added the VA_BITS_ACTUAL vmcoreinfo entry (?), > and I'm presuming there will be other VM-related things that are > different from that original kernel patch. > > Thanks, > Dave I still haven't got a 5.4-rc1 kernel to test with, but Chris Von Recklinghausen has backported the latest upstream 52-bit patch-set to a 4.18-based kernel. However, it did require one additional patch needed for crash to run live on /proc/kcore, which he posted on the arm list today: [PATCH] arm64: remove arm64 definitions of kc_vaddr_to_offset and kc_offset_to_vaddr. http://lists.infradead.org/pipermail/linux-arm-kernel/2019-October/685109.html With his patch-set (including the above), crash-7.2.7 comes up unmodified on a live system -- albeit one which downgrades from 52 bits to 48 bits. Chris is going to test crash running live on a 52 bit simulator tomorrow. Dave -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
- Original Message - > Hi Dave, > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. >Does anyone work on fixing it on kernel side? >(just adding one or two lines though) > > 2. With a tweak above, I still fail to run crash with vmcore >seeing the following errors; >(I didn't dig into details) > >crash-arm64: invalid kernel virtual address: 800010c689a8 type: > "kernel_config_data" >WARNING: cannot read kernel_config_data >crash-arm64: invalid kernel virtual address: 80001177a160 type: > "possible" >WARNING: cannot read cpu_possible_map >crash-arm64: invalid kernel virtual address: 80001177a140 type: > "present" >WARNING: cannot read cpu_present_map >crash-arm64: invalid kernel virtual address: 80001177a120 type: > "online" >WARNING: cannot read cpu_online_map >crash-arm64: invalid kernel virtual address: 80001177a188 type: > "active" >WARNING: cannot read cpu_active_map >crash-arm64: invalid kernel virtual address: 80001196afe0 type: > "shadow_timekeeper xtime_sec" >crash-arm64: invalid kernel virtual address: 800011782b3c type: > "init_uts_ns" >WARNING: invalid linux_banner pointer: 800010c500e8 >crash: vmlinux.1003 and vmcore.1003 do not match! > > Have you noticed this issue? I have been waiting for our internal 5.4-rc1 based ARK kernel to be built, and it looks like it should be available soon. But I did not expect crash to work on arm64 in 5.4. I put in a bunch of 52-bit support into the crash utility last spring, but it was based upon an early proposed patch-set from Steve Capper that Chris von Recklinghausen backported to our 4.18.0-80.el8 kernel at the time. I believe Chris may have added the VA_BITS_ACTUAL vmcoreinfo entry (?), and I'm presuming there will be other VM-related things that are different from that original kernel patch. Thanks, Dave > > Thanks, > -Takahiro Akashi > > -- > Crash-utility mailing list > Crash-utility@redhat.com > https://www.redhat.com/mailman/listinfo/crash-utility > -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
Hi Bhupesh, On Thu, Oct 03, 2019 at 11:25:25AM +0530, Bhupesh Sharma wrote: > Hi Akashi, > > On Thu, Oct 3, 2019 at 10:49 AM AKASHI Takahiro > wrote: > > > > Hi Dave, > > > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. > >Does anyone work on fixing it on kernel side? > >(just adding one or two lines though) > > I am working on the changes in the vmcoreinfo after the 52-bit VA > changes from Steve were accepted in Linux 5.4-rc1 ("Support for 52-bit > virtual addressing in kernel space"). Okay. I was trying to make sure that my kexec_file_load-based kdump patch would work on v5.4 before reposting it to kernel ML. Thanks, -Takahiro Akashi > The suggestions during the last set of reviews (my v3 patchset) was to > introduce arch specific details like TCR_EL1.TxSZ in the vmcoreinfo > and also to wait for the 52-bit changes to make way upstream (as I > raised a couple of issues/kernel panic issues) earlier with the 52-bit > VA changes from Steve. > > I plan to send out the v4 patchset later this week (and will Cc you). > > Thanks, > Bhupesh > > > > 2. With a tweak above, I still fail to run crash with vmcore > >seeing the following errors; > >(I didn't dig into details) > > > >crash-arm64: invalid kernel virtual address: 800010c689a8 type: > > "kernel_config_data" > >WARNING: cannot read kernel_config_data > >crash-arm64: invalid kernel virtual address: 80001177a160 type: > > "possible" > >WARNING: cannot read cpu_possible_map > >crash-arm64: invalid kernel virtual address: 80001177a140 type: > > "present" > >WARNING: cannot read cpu_present_map > >crash-arm64: invalid kernel virtual address: 80001177a120 type: > > "online" > >WARNING: cannot read cpu_online_map > >crash-arm64: invalid kernel virtual address: 80001177a188 type: > > "active" > >WARNING: cannot read cpu_active_map > >crash-arm64: invalid kernel virtual address: 80001196afe0 type: > > "shadow_timekeeper xtime_sec" > >crash-arm64: invalid kernel virtual address: 800011782b3c type: > > "init_uts_ns" > >WARNING: invalid linux_banner pointer: 800010c500e8 > >crash: vmlinux.1003 and vmcore.1003 do not match! > > > > Have you noticed this issue? > > > > Thanks, > > -Takahiro Akashi > > > > -- > > Crash-utility mailing list > > Crash-utility@redhat.com > > https://www.redhat.com/mailman/listinfo/crash-utility > > -- > Crash-utility mailing list > Crash-utility@redhat.com > https://www.redhat.com/mailman/listinfo/crash-utility -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
Re: [Crash-utility] arm64: doesn't work for kernel v5.4?
Hi Akashi, On Thu, Oct 3, 2019 at 10:49 AM AKASHI Takahiro wrote: > > Hi Dave, > > I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. > > 1. VA_BITS_ACTUAL is missing in vmcoreinfo. >Does anyone work on fixing it on kernel side? >(just adding one or two lines though) I am working on the changes in the vmcoreinfo after the 52-bit VA changes from Steve were accepted in Linux 5.4-rc1 ("Support for 52-bit virtual addressing in kernel space"). The suggestions during the last set of reviews (my v3 patchset) was to introduce arch specific details like TCR_EL1.TxSZ in the vmcoreinfo and also to wait for the 52-bit changes to make way upstream (as I raised a couple of issues/kernel panic issues) earlier with the 52-bit VA changes from Steve. I plan to send out the v4 patchset later this week (and will Cc you). Thanks, Bhupesh > 2. With a tweak above, I still fail to run crash with vmcore >seeing the following errors; >(I didn't dig into details) > >crash-arm64: invalid kernel virtual address: 800010c689a8 type: > "kernel_config_data" >WARNING: cannot read kernel_config_data >crash-arm64: invalid kernel virtual address: 80001177a160 type: > "possible" >WARNING: cannot read cpu_possible_map >crash-arm64: invalid kernel virtual address: 80001177a140 type: > "present" >WARNING: cannot read cpu_present_map >crash-arm64: invalid kernel virtual address: 80001177a120 type: > "online" >WARNING: cannot read cpu_online_map >crash-arm64: invalid kernel virtual address: 80001177a188 type: > "active" >WARNING: cannot read cpu_active_map >crash-arm64: invalid kernel virtual address: 80001196afe0 type: > "shadow_timekeeper xtime_sec" >crash-arm64: invalid kernel virtual address: 800011782b3c type: > "init_uts_ns" >WARNING: invalid linux_banner pointer: 800010c500e8 >crash: vmlinux.1003 and vmcore.1003 do not match! > > Have you noticed this issue? > > Thanks, > -Takahiro Akashi > > -- > Crash-utility mailing list > Crash-utility@redhat.com > https://www.redhat.com/mailman/listinfo/crash-utility -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility
[Crash-utility] arm64: doesn't work for kernel v5.4?
Hi Dave, I tried to use crash v7.2.7 with kernel v5.4-rc1 on arm64. 1. VA_BITS_ACTUAL is missing in vmcoreinfo. Does anyone work on fixing it on kernel side? (just adding one or two lines though) 2. With a tweak above, I still fail to run crash with vmcore seeing the following errors; (I didn't dig into details) crash-arm64: invalid kernel virtual address: 800010c689a8 type: "kernel_config_data" WARNING: cannot read kernel_config_data crash-arm64: invalid kernel virtual address: 80001177a160 type: "possible" WARNING: cannot read cpu_possible_map crash-arm64: invalid kernel virtual address: 80001177a140 type: "present" WARNING: cannot read cpu_present_map crash-arm64: invalid kernel virtual address: 80001177a120 type: "online" WARNING: cannot read cpu_online_map crash-arm64: invalid kernel virtual address: 80001177a188 type: "active" WARNING: cannot read cpu_active_map crash-arm64: invalid kernel virtual address: 80001196afe0 type: "shadow_timekeeper xtime_sec" crash-arm64: invalid kernel virtual address: 800011782b3c type: "init_uts_ns" WARNING: invalid linux_banner pointer: 800010c500e8 crash: vmlinux.1003 and vmcore.1003 do not match! Have you noticed this issue? Thanks, -Takahiro Akashi -- Crash-utility mailing list Crash-utility@redhat.com https://www.redhat.com/mailman/listinfo/crash-utility