| Message ID | 20210112132449.22243-1-tiwai@suse.de |
|---|---|
| Headers | show |
| Series | iwlwifi: Fix a crash at loading | expand |
On Tue, 12 Jan 2021 16:50:54 +0100, Kalle Valo wrote: > > Takashi Iwai <tiwai@suse.de> writes: > > > The ucode TLV data may be read-only and should be treated as const > > pointers, but currently a few code forcibly cast to the writable > > pointer unnecessarily. This gave developers a wrong impression as if > > it can be modified, resulting in crashing regressions already a couple > > of times. > > > > This patch adds the const prefix to those cast pointers, so that such > > attempt can be caught more easily in future. > > > > Signed-off-by: Takashi Iwai <tiwai@suse.de> > > So this need to go to -next, right? Yes, this isn't urgently needed for 5.11. > Does this depend on patch 1 or can > this be applied independently? It depends on the first patch, otherwise you'll get the warning in the code changing the const data (it must warn -- that's the purpose of this change :) So, if applying to a separate branch is difficult, applying together for 5.11 would be an option. thanks, Takashi
On Tue, 2021-01-12 at 17:05 +0100, Takashi Iwai wrote: > On Tue, 12 Jan 2021 16:50:54 +0100, > Kalle Valo wrote: > > > > Takashi Iwai <tiwai@suse.de> writes: > > > > > The ucode TLV data may be read-only and should be treated as const > > > pointers, but currently a few code forcibly cast to the writable > > > pointer unnecessarily. This gave developers a wrong impression as if > > > it can be modified, resulting in crashing regressions already a couple > > > of times. > > > > > > This patch adds the const prefix to those cast pointers, so that such > > > attempt can be caught more easily in future. > > > > > > Signed-off-by: Takashi Iwai <tiwai@suse.de> > > > > So this need to go to -next, right? > > Yes, this isn't urgently needed for 5.11. Acked-by: Luca Coelho <luciano.coelho@intel.com> > > Does this depend on patch 1 or can > > this be applied independently? > > It depends on the first patch, otherwise you'll get the warning in the > code changing the const data (it must warn -- that's the purpose of > this change :) > > So, if applying to a separate branch is difficult, applying together > for 5.11 would be an option. It doesn't matter to me how you apply it. Applying together is obviously going to be easier, but applying separately wouldn't be that hard either. You'd just have to track when 1/2 went into net-next before applying this one. Kalle's call. -- Cheers, Luca.
"Coelho, Luciano" <luciano.coelho@intel.com> writes: > On Tue, 2021-01-12 at 17:05 +0100, Takashi Iwai wrote: >> On Tue, 12 Jan 2021 16:50:54 +0100, >> Kalle Valo wrote: >> > >> > Takashi Iwai <tiwai@suse.de> writes: >> > >> > > The ucode TLV data may be read-only and should be treated as const >> > > pointers, but currently a few code forcibly cast to the writable >> > > pointer unnecessarily. This gave developers a wrong impression as if >> > > it can be modified, resulting in crashing regressions already a couple >> > > of times. >> > > >> > > This patch adds the const prefix to those cast pointers, so that such >> > > attempt can be caught more easily in future. >> > > >> > > Signed-off-by: Takashi Iwai <tiwai@suse.de> >> > >> > So this need to go to -next, right? >> >> Yes, this isn't urgently needed for 5.11. > > Acked-by: Luca Coelho <luciano.coelho@intel.com> > > >> > Does this depend on patch 1 or can >> > this be applied independently? >> >> It depends on the first patch, otherwise you'll get the warning in the >> code changing the const data (it must warn -- that's the purpose of >> this change :) >> >> So, if applying to a separate branch is difficult, applying together >> for 5.11 would be an option. > > It doesn't matter to me how you apply it. Applying together is > obviously going to be easier, but applying separately wouldn't be that > hard either. You'd just have to track when 1/2 went into net-next > before applying this one. Kalle's call. Ok, I'll apply this to wireless-drivers-next after wireless-drivers is merged to -next. It might take a while. -- https://patchwork.kernel.org/project/linux-wireless/list/ https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches