| Message ID | 20220124164847.54002-1-zhou1615@umn.edu (mailing list archive) |
|---|---|
| State | Awaiting Upstream |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | brcmfmac: Fix a wild pointer dereference bug in brcmf_chip_recognition() | expand |
| Context | Check | Description |
|---|---|---|
| netdev/tree_selection | success | Not a local patch |
On Tue, Jan 25, 2022 at 12:48:45AM +0800, Zhou Qingyang wrote: > In brcmf_chip_recognition(), the return value of brcmf_chip_add_core() > is assigned to core and is passed to brcmf_chip_sb_corerev(). In > brcmf_chip_sb_corerev(), there exists dereference of core without check. > the return value of brcmf_chip_add_core() could be ERR_PTR on failure of > allocation, which could lead to a NULL pointer dereference bug. > > Fix this bug by adding IS_ERR check for every variable core. > > This bug was found by a static analyzer. > > Builds with 'make allyesconfig' show no new warnings, > and our static analyzer no longer warns about this code > > Fixes: cb7cf7be9eba ("brcmfmac: make chip related functions host interface independent") > Signed-off-by: Zhou Qingyang <zhou1615@umn.edu> > --- > The analysis employs differential checking to identify inconsistent > security operations (e.g., checks or kfrees) between two code paths > and confirms that the inconsistent operations are not recovered in the > current function or the callers, so they constitute bugs. > > Note that, as a bug found by static analysis, it can be a false > positive or hard to trigger. Multiple researchers have cross-reviewed > the bug. As stated before, umn.edu is still not allowed to contribute to the Linux kernel. Please work with your administration to resolve this issue.
Greg KH <greg@kroah.com> writes: > On Tue, Jan 25, 2022 at 12:48:45AM +0800, Zhou Qingyang wrote: >> In brcmf_chip_recognition(), the return value of brcmf_chip_add_core() >> is assigned to core and is passed to brcmf_chip_sb_corerev(). In >> brcmf_chip_sb_corerev(), there exists dereference of core without check. >> the return value of brcmf_chip_add_core() could be ERR_PTR on failure of >> allocation, which could lead to a NULL pointer dereference bug. >> >> Fix this bug by adding IS_ERR check for every variable core. >> >> This bug was found by a static analyzer. >> >> Builds with 'make allyesconfig' show no new warnings, >> and our static analyzer no longer warns about this code >> >> Fixes: cb7cf7be9eba ("brcmfmac: make chip related functions host >> interface independent") >> Signed-off-by: Zhou Qingyang <zhou1615@umn.edu> >> --- >> The analysis employs differential checking to identify inconsistent >> security operations (e.g., checks or kfrees) between two code paths >> and confirms that the inconsistent operations are not recovered in the >> current function or the callers, so they constitute bugs. >> >> Note that, as a bug found by static analysis, it can be a false >> positive or hard to trigger. Multiple researchers have cross-reviewed >> the bug. > > As stated before, umn.edu is still not allowed to contribute to the > Linux kernel. Please work with your administration to resolve this > issue. Thanks Greg, I didn't notice that this is from umn.edu. After seeing what kind of "research" umn.edu does I will not even look at umn.edu patches, they all will be automatically rejected without comments.
On Fri, Jan 28, 2022 at 12:31:44PM +0200, Kalle Valo wrote: > Greg KH <greg@kroah.com> writes: > > > On Tue, Jan 25, 2022 at 12:48:45AM +0800, Zhou Qingyang wrote: > >> In brcmf_chip_recognition(), the return value of brcmf_chip_add_core() > >> is assigned to core and is passed to brcmf_chip_sb_corerev(). In > >> brcmf_chip_sb_corerev(), there exists dereference of core without check. > >> the return value of brcmf_chip_add_core() could be ERR_PTR on failure of > >> allocation, which could lead to a NULL pointer dereference bug. > >> > >> Fix this bug by adding IS_ERR check for every variable core. > >> > >> This bug was found by a static analyzer. > >> > >> Builds with 'make allyesconfig' show no new warnings, > >> and our static analyzer no longer warns about this code > >> > >> Fixes: cb7cf7be9eba ("brcmfmac: make chip related functions host > >> interface independent") > >> Signed-off-by: Zhou Qingyang <zhou1615@umn.edu> > >> --- > >> The analysis employs differential checking to identify inconsistent > >> security operations (e.g., checks or kfrees) between two code paths > >> and confirms that the inconsistent operations are not recovered in the > >> current function or the callers, so they constitute bugs. > >> > >> Note that, as a bug found by static analysis, it can be a false > >> positive or hard to trigger. Multiple researchers have cross-reviewed > >> the bug. > > > > As stated before, umn.edu is still not allowed to contribute to the > > Linux kernel. Please work with your administration to resolve this > > issue. > > Thanks Greg, I didn't notice that this is from umn.edu. After seeing > what kind of "research" umn.edu does I will not even look at umn.edu > patches, they all will be automatically rejected without comments. Thank you. We could just block their emails from the mailing lists, but then that would not let us see when they send a patch and cc: the relevant maintainers, so we have to live with this way for now. I'll be contacting the umn.edu administration again and ask them what went wrong here. greg k-h
diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/chip.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/chip.c index 1ee49f9e325d..4d91cb107cd7 100644 --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/chip.c +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/chip.c @@ -986,18 +986,28 @@ static int brcmf_chip_recognition(struct brcmf_chip_priv *ci) core = brcmf_chip_add_core(ci, BCMA_CORE_CHIPCOMMON, SI_ENUM_BASE_DEFAULT, 0); + if (IS_ERR(core)) + return PTR_ERR(core); brcmf_chip_sb_corerev(ci, core); core = brcmf_chip_add_core(ci, BCMA_CORE_SDIO_DEV, BCM4329_CORE_BUS_BASE, 0); + if (IS_ERR(core)) + return PTR_ERR(core); brcmf_chip_sb_corerev(ci, core); core = brcmf_chip_add_core(ci, BCMA_CORE_INTERNAL_MEM, BCM4329_CORE_SOCRAM_BASE, 0); + if (IS_ERR(core)) + return PTR_ERR(core); brcmf_chip_sb_corerev(ci, core); core = brcmf_chip_add_core(ci, BCMA_CORE_ARM_CM3, BCM4329_CORE_ARM_BASE, 0); + if (IS_ERR(core)) + return PTR_ERR(core); brcmf_chip_sb_corerev(ci, core); core = brcmf_chip_add_core(ci, BCMA_CORE_80211, 0x18001000, 0); + if (IS_ERR(core)) + return PTR_ERR(core); brcmf_chip_sb_corerev(ci, core); } else if (socitype == SOCI_AI) { ci->iscoreup = brcmf_chip_ai_iscoreup;
In brcmf_chip_recognition(), the return value of brcmf_chip_add_core() is assigned to core and is passed to brcmf_chip_sb_corerev(). In brcmf_chip_sb_corerev(), there exists dereference of core without check. the return value of brcmf_chip_add_core() could be ERR_PTR on failure of allocation, which could lead to a NULL pointer dereference bug. Fix this bug by adding IS_ERR check for every variable core. This bug was found by a static analyzer. Builds with 'make allyesconfig' show no new warnings, and our static analyzer no longer warns about this code Fixes: cb7cf7be9eba ("brcmfmac: make chip related functions host interface independent") Signed-off-by: Zhou Qingyang <zhou1615@umn.edu> --- The analysis employs differential checking to identify inconsistent security operations (e.g., checks or kfrees) between two code paths and confirms that the inconsistent operations are not recovered in the current function or the callers, so they constitute bugs. Note that, as a bug found by static analysis, it can be a false positive or hard to trigger. Multiple researchers have cross-reviewed the bug. .../net/wireless/broadcom/brcm80211/brcmfmac/chip.c | 10 ++++++++++ 1 file changed, 10 insertions(+)