Request for reviews for 4.4-st44 4.4-cip79

Request for reviews for 4.4-st44 4.4-cip79

[Ulrich Hecht]

Hi!

Here are the manual backports for the upcoming 4.4-st44 and 4.4-cip79 releases that should be reviewed, and that can currently be found in
https://git.kernel.org/pub/scm/linux/kernel/git/cip/linux-cip.git/log/?h=linux-4.4.y-st-rc

My apologies, this is a big one. Thankfully, most backports are of the trivial sort:

** m-| 9559a3e60 f9c998 o  | x86/smp: Use dedicated cache-line for mwait_play_dead()
** m-| f175f567b f8ef12 .+ | ARM: orion5x: fix d2net gpio initialization
** m-| 4ee69c91c 6eef7a o+ | netfilter: conntrack: Avoid nf_ct_helper_hash uses after free
** m-| f66069f9d 247c3d o  | ext4: fix wrong unit use in ext4_mb_clear_bb
** m-| b3e17e6e6 a82d62 o+ | Revert "8250: add support for ASIX devices with a FIFO bug"
** m-| f35bd94b4 af73f2 o  | scsi: qla2xxx: Check valid rport returned by fc_bsg_to_rport()
** m-| 005961bd8 00eca1 o  | scsi: qla2xxx: Pointer may be dereferenced
** m-| 11b8e27ed 55c3b9 o+ | can: bcm: Fix UAF in bcm_proc_show()
** m-| 8dc52c200 8ce884 f  | posix-timers: Ensure timer ID search-loop limit is valid
** m-| 4793f9802 1aeb87 o+ | tcp: annotate data-races around tp->notsent_lowat
** m-| a276d17dc c55afc .  | usb: ohci-at91: Fix the unhandle interrupt when resume
** m-| d3009700f c2a88e o+ | drm/client: Fix memory leak in drm_client_target_cloned
** m-| f0f874147 3044b1 o+ | net/sched: cls_u32: No longer copy tcf_result on update to avoid use-after-free
** m-| 005f68d6b 949ad6 o+ | tcp_metrics: annotate data-races around tm->tcpm_stamp
** m-| 38f1f09dc 8c4d04 o+ | tcp_metrics: annotate data-races around tm->tcpm_vals[]
** m-| e28208e21 404615 o  | ext2: Drop fragment support
** m-| 368684d3b 048c79 o  | ipv6: adjust ndisc_is_useropt() to also return true for PIO
** m-| 8fad4df94 8a9896 o+ | net/packet: annotate data-races around tp->status
** m-| 4eb1fca13 d6a95d    | quota: fix warning in dqgrab()

These might need a closer look:

** mo| 70cc55f6c 1cba6c o+ | kexec: fix a memory leak in crash_shrink_memory()
** mo| 6fb4b3443 504a10 o  | gfs2: Don't deref jdesc in evict
** mo| 5df411bb8 8cda3e o+ | dmaengine: pl330: Return DMA_PAUSED when transaction is paused

There is also "x86/msr-index: remove duplicate definition of MSR_AMD64_DE_CFG", a minor fixup for a warning introduced in the previous release.

Thanks for having a look.

CU
Uli

Re: Request for reviews for 4.4-st44 4.4-cip79

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 834 bytes --]

Hi!

> Here are the manual backports for the upcoming 4.4-st44 and 4.4-cip79 releases that should be reviewed, and that can currently be found in
> https://git.kernel.org/pub/scm/linux/kernel/git/cip/linux-cip.git/log/?h=linux-4.4.y-st-rc
> 
> My apologies, this is a big one. Thankfully, most backports are of the trivial sort:

Ok, let me take a look.

> There is also "x86/msr-index: remove duplicate definition of
>MSR_AMD64_DE_CFG", a minor fixup for a warning introduced in the
>previous release.

I'd do this one the other way around -- keep the definition in the
header, as we have in mainline, and remove the one in the .c file.

Best regards,
								Pavel

-- 
DENX Software Engineering GmbH,        Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: Request for reviews for 4.4-st44 4.4-cip79

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 2845 bytes --]

Hi!

> My apologies, this is a big one. Thankfully, most backports are of the trivial sort:

a-| 9559a3e60 f9c998 o  | x86/smp: Use dedicated cache-line for mwait_play_dead()
a -| f175f567b f8ef12 .+ | ARM: orion5x: fix d2net gpio initialization
!!incomplete,  _helper_init() needs same? -| 4ee69c91c 6eef7a o+ | netfilter: conntrack: Avoid nf_ct_helper_hash uses after free
a -| f66069f9d 247c3d o  | ext4: fix wrong unit use in ext4_mb_clear_bb
a -| b3e17e6e6 a82d62 o+ | Revert "8250: add support for ASIX devices with a FIFO bug"
would not bother -- -| f35bd94b4 af73f2 o  | scsi: qla2xxx: Check valid rport returned by fc_bsg_to_rport()
would not bother -- -| 005961bd8 00eca1 o  | scsi: qla2xxx: Pointer may be dereferenced
a -| 11b8e27ed 55c3b9 o+ | can: bcm: Fix UAF in bcm_proc_show()
a -| 8dc52c200 8ce884 f  | posix-timers: Ensure timer ID search-loop limit is valid
a -| 4793f9802 1aeb87 o+ | tcp: annotate data-races around tp->notsent_lowat
!!WRONG > ** m-| a276d17dc c55afc .  | usb: ohci-at91: Fix the unhandle interrupt when resume
a -| d3009700f c2a88e o+ | drm/client: Fix memory leak in drm_client_target_cloned
a -| f0f874147 3044b1 o+ | net/sched: cls_u32: No longer copy tcf_result on update to avoid use-after-free
a -| 005f68d6b 949ad6 o+ | tcp_metrics: annotate data-races around tm->tcpm_stamp
a -| 38f1f09dc 8c4d04 o+ | tcp_metrics: annotate data-races around tm->tcpm_vals[]
I'd not bother backporting this -- -| e28208e21 404615 o  | ext2: Drop fragment support
a -| 368684d3b 048c79 o  | ipv6: adjust ndisc_is_useropt() to also return true for PIO
a -| 8fad4df94 8a9896 o+ | net/packet: annotate data-races around tp->status
a -| 4eb1fca13 d6a95d    | quota: fix warning in dqgrab()

Most look okay to me. I'd not put too much work with READ_ONCE
annotations or qla2xxx warning tweaks, they don't really fix any real
bugs. "ext2: Drop fragment support" probably does not hurt, but surely
does not fix anything. I guess we can keep it when -stable took it,
but...

"usb: ohci-at91: Fix the unhandle interrupt when resume" is wrong, it
should do ohci_resume() once, not twice.

"netfilter: conntrack: Avoid nf_ct_helper_hash uses after free" is at
least incomplete -- nf_conntrack_helper_init() does not set the hash
to NULL in the error path. (Mainline does not have that bug, and code
change a lot there).

> These might need a closer look:

a | 70cc55f6c 1cba6c o+ | kexec: fix a memory leak in crash_shrink_memory()
a | 6fb4b3443 504a10 o  | gfs2: Don't deref jdesc in evict
a | 5df411bb8 8cda3e o+ | dmaengine: pl330: Return DMA_PAUSED when transaction is paused

These look okay to me.

Best regards,
								Pavel
-- 
DENX Software Engineering GmbH,        Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]

Re: [cip-dev] Request for reviews for 4.4-st44 4.4-cip79

[Ulrich Hecht]

Thanks for the reviews!

> On 09/04/2023 1:43 PM CEST Pavel Machek <pavel@denx.de> wrote:
> Most look okay to me. I'd not put too much work with READ_ONCE
> annotations or qla2xxx warning tweaks, they don't really fix any real
> bugs. "ext2: Drop fragment support" probably does not hurt, but surely
> does not fix anything. I guess we can keep it when -stable took it,
> but...

I'm inclined to include such things if they are trivial to adjust because I think it will save time fixing up future patches.

> "usb: ohci-at91: Fix the unhandle interrupt when resume" is wrong, it
> should do ohci_resume() once, not twice.

Correct, I'll fix that up.

> "netfilter: conntrack: Avoid nf_ct_helper_hash uses after free" is at
> least incomplete -- nf_conntrack_helper_init() does not set the hash
> to NULL in the error path. (Mainline does not have that bug, and code
> change a lot there).

Seems like that applies to all backports between 4.14 and 5.10, at least. I'll leave that out for now and ask stable about it.

CU
Uli

Re: [cip-dev] Request for reviews for 4.4-st44 4.4-cip79

[Pavel Machek]

[-- Attachment #1: Type: text/plain, Size: 1283 bytes --]

Hi!

> Thanks for the reviews!
> 
> > On 09/04/2023 1:43 PM CEST Pavel Machek <pavel@denx.de> wrote:
> > Most look okay to me. I'd not put too much work with READ_ONCE
> > annotations or qla2xxx warning tweaks, they don't really fix any real
> > bugs. "ext2: Drop fragment support" probably does not hurt, but surely
> > does not fix anything. I guess we can keep it when -stable took it,
> > but...
> 
> I'm inclined to include such things if they are trivial to adjust because I think it will save time fixing up future patches.
> 
> > "usb: ohci-at91: Fix the unhandle interrupt when resume" is wrong, it
> > should do ohci_resume() once, not twice.
> 
> Correct, I'll fix that up.
> 
> > "netfilter: conntrack: Avoid nf_ct_helper_hash uses after free" is at
> > least incomplete -- nf_conntrack_helper_init() does not set the hash
> > to NULL in the error path. (Mainline does not have that bug, and code
> > change a lot there).
> 
> Seems like that applies to all backports between 4.14 and 5.10, at least. I'll leave that out for now and ask stable about it.

Thank you!

Best regards,
								Pavel
-- 
DENX Software Engineering GmbH,        Managing Director: Erika Unter
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 195 bytes --]