Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hi Marc-Robin, On Tue, Mar 14, 2023 at 02:48:56PM +0100, Marc-Robin Wendt wrote: > Hello Salvatore, > > sorry for the delay. So lets hope, this will solve the problem for all > the affected people. > > On Fri, 2023-03-10 at 17:11 +0100, Salvatore Bonaccorso wrote: > > Hello Marc-Robin, > > > > > > > Anyhow I had some errors: > > > > > > qla2xxx :61:00.1: swiotlb buffer is full (sz: 65536 bytes), > > > total > > > 32768 (slots), used 28799 (slots) > > > > > > But I'm not sure, if this is releated and it doesn't affect > > > functions. > > > > > > Do you need further informations about the testing process? > > > > I assume this has to be tackled separately. You can confim it was not > > seen before in earlier kernels? > > > > > I checked with last working Kernel 5.10.136-1 and same errors there. So > not related to our problem. Seems like is related to LVM. > > Thanks for you efforts. Thanks for checking back to the earlier version, so at least not related to this issue and relieved the patch series resolves the problem. I'm confident will be as well the case for the other affected. I have queued them for the next Debian upload based on at least 5.10.174-1 as well. Regards, Salvatore
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hello Salvatore, sorry for the delay. So lets hope, this will solve the problem for all the affected people. On Fri, 2023-03-10 at 17:11 +0100, Salvatore Bonaccorso wrote: > Hello Marc-Robin, > > > > Anyhow I had some errors: > > > > qla2xxx :61:00.1: swiotlb buffer is full (sz: 65536 bytes), > > total > > 32768 (slots), used 28799 (slots) > > > > But I'm not sure, if this is releated and it doesn't affect > > functions. > > > > Do you need further informations about the testing process? > > I assume this has to be tackled separately. You can confim it was not > seen before in earlier kernels? > > I checked with last working Kernel 5.10.136-1 and same errors there. So not related to our problem. Seems like is related to LVM. Thanks for you efforts. Greets Marc-Robin
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hello Marc-Robin, On Thu, Mar 09, 2023 at 09:33:19PM +0100, Marc-Robin Wendt wrote: > Hello Salvatore, > > I followed instructions and it made me a linux-image-5.10.0-21-amd64- > unsigned_5.10.162-1a~test_amd64.deb, which, after installed, boots > nicely and seems to work without errors at mpt3sas anymore. > All XEN-functionality seems to work in our environment. Thanks a lot, that was helpful. Good news, Greg has now picked the series: https://lore.kernel.org/stable/zasovkmt7n9fd...@kroah.com/ > Anyhow I had some errors: > > qla2xxx :61:00.1: swiotlb buffer is full (sz: 65536 bytes), total > 32768 (slots), used 28799 (slots) > > But I'm not sure, if this is releated and it doesn't affect functions. > > Do you need further informations about the testing process? I assume this has to be tackled separately. You can confim it was not seen before in earlier kernels? Thanks, Regards, Salvatore
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hello Salvatore, I followed instructions and it made me a linux-image-5.10.0-21-amd64- unsigned_5.10.162-1a~test_amd64.deb, which, after installed, boots nicely and seems to work without errors at mpt3sas anymore. All XEN-functionality seems to work in our environment. Anyhow I had some errors: qla2xxx :61:00.1: swiotlb buffer is full (sz: 65536 bytes), total 32768 (slots), used 28799 (slots) But I'm not sure, if this is releated and it doesn't affect functions. Do you need further informations about the testing process? -- Greetings Marc-Robin On Wed, 2023-03-08 at 17:38 +0100, Salvatore Bonaccorso wrote: > Hi Marc, > > On Sat, Mar 04, 2023 at 12:09:15PM +0100, Marc-Robin Wendt wrote: > > Its actually not my game, but if it speeds things up, I will help > > testing (if anyone tells me, what to do). > > Thank you. Can you try with the following series of 4 patches. > > https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s4.2.2 > > will describe how to pass a series of patches to apply for testing. > > Thanks already in advance, > > Regards, > Salvatore
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hi Marc,
On Sat, Mar 04, 2023 at 12:09:15PM +0100, Marc-Robin Wendt wrote:
> Its actually not my game, but if it speeds things up, I will help
> testing (if anyone tells me, what to do).
Thank you. Can you try with the following series of 4 patches.
https://kernel-team.pages.debian.net/kernel-handbook/ch-common-tasks.html#s4.2.2
will describe how to pass a series of patches to apply for testing.
Thanks already in advance,
Regards,
Salvatore
>From 37904556ea6f04f48c3ca9419edf8155aa1e1899 Mon Sep 17 00:00:00 2001
From: Salvatore Bonaccorso
Date: Wed, 8 Mar 2023 17:14:56 +0100
Subject: [PATCH 1/4] Revert "scsi: mpt3sas: Fix return value check of
dma_get_required_mask()"
This reverts commit e0e0747de0ea3dd87cdbb0393311e17471a9baf1.
As noted in 1a2dcbdde82e ("scsi: mpt3sas: re-do lost mpt3sas DMA mask
fix") in mainline there was a mis-merge in commit 62e6e5940c0c ("Merge
tag 'scsi-misc' of
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi"). causing that
the fix needed to be redone later on again. To make series of patches
apply cleanly to the stable series where e0e0747de0ea ("scsi: mpt3sas:
Fix return value check of dma_get_required_mask()") was backported,
revert the afferomentioned commit.
No upstream commit exists for this commit.
Link: https://lore.kernel.org/regressions/yq1sfehmjnb@ca-mkp.ca.oracle.com/
Signed-off-by: Salvatore Bonaccorso
---
drivers/scsi/mpt3sas/mpt3sas_base.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index c1b76cda60db..18f85c963944 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -2825,7 +2825,7 @@ _base_config_dma_addressing(struct MPT3SAS_ADAPTER *ioc, struct pci_dev *pdev)
if (ioc->is_mcpu_endpoint ||
sizeof(dma_addr_t) == 4 || ioc->use_32bit_dma ||
- dma_get_required_mask(>dev) <= DMA_BIT_MASK(32))
+ dma_get_required_mask(>dev) <= 32)
ioc->dma_mask = 32;
/* Set 63 bit DMA mask for all SAS3 and SAS35 controllers */
else if (ioc->hba_mpi_version_belonged > MPI2_VERSION)
--
2.39.2
>From 8ad037f24e7a9c46105065e81503968b7bdc14b1 Mon Sep 17 00:00:00 2001
From: Sreekanth Reddy
Date: Thu, 25 Aug 2022 13:24:54 +0530
Subject: [PATCH 2/4] scsi: mpt3sas: Don't change DMA mask while reallocating
pools
commit 9df650963bf6d6c2c3fcd325d8c44ca2b99554fe upstream.
When a pool crosses the 4GB boundary region then before reallocating pools
change the coherent DMA mask to 32 bits and keep the normal DMA mask set to
63/64 bits.
Link: https://lore.kernel.org/r/20220825075457.16422-2-sreekanth.re...@broadcom.com
Signed-off-by: Sreekanth Reddy
Signed-off-by: Martin K. Petersen
Signed-off-by: Salvatore Bonaccorso
---
drivers/scsi/mpt3sas/mpt3sas_base.c | 21 ++---
1 file changed, 14 insertions(+), 7 deletions(-)
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c
index 18f85c963944..faea8001adf5 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_base.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_base.c
@@ -2822,19 +2822,26 @@ static int
_base_config_dma_addressing(struct MPT3SAS_ADAPTER *ioc, struct pci_dev *pdev)
{
struct sysinfo s;
+ u64 coherent_dma_mask, dma_mask;
- if (ioc->is_mcpu_endpoint ||
- sizeof(dma_addr_t) == 4 || ioc->use_32bit_dma ||
- dma_get_required_mask(>dev) <= 32)
+ if (ioc->is_mcpu_endpoint || sizeof(dma_addr_t) == 4 ||
+ dma_get_required_mask(>dev) <= 32) {
ioc->dma_mask = 32;
+ coherent_dma_mask = dma_mask = DMA_BIT_MASK(32);
/* Set 63 bit DMA mask for all SAS3 and SAS35 controllers */
- else if (ioc->hba_mpi_version_belonged > MPI2_VERSION)
+ } else if (ioc->hba_mpi_version_belonged > MPI2_VERSION) {
ioc->dma_mask = 63;
- else
+ coherent_dma_mask = dma_mask = DMA_BIT_MASK(63);
+ } else {
ioc->dma_mask = 64;
+ coherent_dma_mask = dma_mask = DMA_BIT_MASK(64);
+ }
+
+ if (ioc->use_32bit_dma)
+ coherent_dma_mask = DMA_BIT_MASK(32);
- if (dma_set_mask(>dev, DMA_BIT_MASK(ioc->dma_mask)) ||
- dma_set_coherent_mask(>dev, DMA_BIT_MASK(ioc->dma_mask)))
+ if (dma_set_mask(>dev, dma_mask) ||
+ dma_set_coherent_mask(>dev, coherent_dma_mask))
return -ENODEV;
if (ioc->dma_mask > 32) {
--
2.39.2
>From ecd75f575db8537651bd6ee4edaaf0f508951a8c Mon Sep 17 00:00:00 2001
From: Sreekanth Reddy
Date: Tue, 13 Sep 2022 17:35:38 +0530
Subject: [PATCH 3/4] scsi: mpt3sas: re-do lost mpt3sas DMA mask fix
commit 1a2dcbdde82e3a5f1db9b2f4c48aa1aeba534fb2 upstream.
This is a re-do of commit e0e0747de0ea ("scsi: mpt3sas: Fix return value
check of dma_get_required_mask()"), which I ended up undoing in a
mis-merge in commit 62e6e5940c0c ("Merge tag 'scsi-misc' of
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi").
The original commit message was
scsi: mpt3sas: Fix return value check of dma_get_required_mask()
Fix the incorrect return value check of dma_get_required_mask(). Due to
this incorrect check, the driver
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Its actually not my game, but if it speeds things up, I will help testing (if anyone tells me, what to do). On Sat, 2023-03-04 at 10:11 +0100, Salvatore Bonaccorso wrote: > Hi Marc-Robin, > > On Sat, Mar 04, 2023 at 09:36:35AM +0100, Marc-Robin Wendt wrote: > > Ok folks. That's it now. I have a server farm with Debian+Xen and > > I'm > > stuck in a dead end, because this Bug gets not fixed. I delayed > > kernel > > updates since 5 month now, but can't delay anymore, because of CVE- > > 2022-42328, CVE-2022-42329 and CVE-2022-3643. > > > > So I have to change to Redhat now? Means my servers will be lost > > for > > Debian for a pretty while. It's a pity, because like this, Debian > > Project goes down the drain. Maybe the maintainers should think > > about > > priorities. Bugs like this usually don't affect end consumer > > devices, > > but the big server farm provider, who btw. donate a pretty much of > > ressources to the Debian Project. > > People involved appear to prefer to rant on this bug instead of > providing help and taking note of what was mentioned in > https://bugs.debian.org/1022126#117 and earlier replies and the fact > that the discussion was brought up to upstream by the Debian > maintainers an happens on the upstream regression list: > > https://lore.kernel.org/regressions/754b030c-ba14-167c-e2d0-2f4f5bf55...@leemhuis.info/ > > I take your mail such that you are willing to test accordingly > patches > which would be followed by the above thread? Including adding > Tested-by on needed-to-backport patches? > > Regards, > Salvatore >
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hi Marc-Robin, On Sat, Mar 04, 2023 at 09:36:35AM +0100, Marc-Robin Wendt wrote: > Ok folks. That's it now. I have a server farm with Debian+Xen and I'm > stuck in a dead end, because this Bug gets not fixed. I delayed kernel > updates since 5 month now, but can't delay anymore, because of CVE- > 2022-42328, CVE-2022-42329 and CVE-2022-3643. > > So I have to change to Redhat now? Means my servers will be lost for > Debian for a pretty while. It's a pity, because like this, Debian > Project goes down the drain. Maybe the maintainers should think about > priorities. Bugs like this usually don't affect end consumer devices, > but the big server farm provider, who btw. donate a pretty much of > ressources to the Debian Project. People involved appear to prefer to rant on this bug instead of providing help and taking note of what was mentioned in https://bugs.debian.org/1022126#117 and earlier replies and the fact that the discussion was brought up to upstream by the Debian maintainers an happens on the upstream regression list: https://lore.kernel.org/regressions/754b030c-ba14-167c-e2d0-2f4f5bf55...@leemhuis.info/ I take your mail such that you are willing to test accordingly patches which would be followed by the above thread? Including adding Tested-by on needed-to-backport patches? Regards, Salvatore
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Ok folks. That's it now. I have a server farm with Debian+Xen and I'm stuck in a dead end, because this Bug gets not fixed. I delayed kernel updates since 5 month now, but can't delay anymore, because of CVE- 2022-42328, CVE-2022-42329 and CVE-2022-3643. So I have to change to Redhat now? Means my servers will be lost for Debian for a pretty while. It's a pity, because like this, Debian Project goes down the drain. Maybe the maintainers should think about priorities. Bugs like this usually don't affect end consumer devices, but the big server farm provider, who btw. donate a pretty much of ressources to the Debian Project. Greetings Marc-Robin Wendt On Mon, 2023-02-20 at 10:30 +0100, vmxevils...@gmail.com wrote: > Hello dear mantainers, > > I have stopped sending kernel packages as per your request (you > defined it spamming). > For who might be interested I am making the 6.2.0 kernel amd64 > packages myself (the 6.2.0 kernel announcement is not yet > on https://tracker.debian.org/pkg/linux). > Since this version fixes bug 1022126 (and, I am sure others), > Since all the CVE announcements on the site I read are about past > kernels.. > If you would like I can send the packages to debian mentor or > wherever is most convenient for you, for testing and review. > Sorry for my past mistakes (still have to fully comprehend where was > I failing, since I was following the manuals, including the source > etc, I'd like you to be more clear for what concerns where just > following your procedures was failing). > > > Thanks > Renato Gallo
Bug#1022126: New mainline kernel is out fixes for example 1022126 bug
Hello dear mantainers, I have stopped sending kernel packages as per your request (you defined it spamming). For who might be interested I am making the 6.2.0 kernel amd64 packages myself (the 6.2.0 kernel announcement is not yet on https://tracker.debian.org/pkg/linux). Since this version fixes bug 1022126 (and, I am sure others), Since all the CVE announcements on the site I read are about past kernels.. If you would like I can send the packages to debian mentor or wherever is most convenient for you, for testing and review. Sorry for my past mistakes (still have to fully comprehend where was I failing, since I was following the manuals, including the source etc, I'd like you to be more clear for what concerns where just following your procedures was failing). Thanks Renato Gallo
