| Message ID | 20241010144703.113728-1-zhengzengkai@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ACPI: GTDT: Tighten the check for the first platform timer entry | expand |
On Thu, 10 Oct 2024 15:47:03 +0100, Zheng Zengkai <zhengzengkai@huawei.com> wrote: > > As suggested by Marc and Lorenzo, first we need to check whether > the platform_timer pointer is within gtdt bounds (< gtdt_end) before > de-referencing what it points at to detect the (first) platform > timer entry length and check that next platform_timer pointer is > within gtdt_end too. Now we do that only in next_platform_timer() > for subsequent platform timers. > > So add check against table length (gtdt_end) for the first platform > timer entry. > > Suggested-by: Marc Zyngier <maz@kernel.org> > Suggested-by: Lorenzo Pieralisi <lpieralisi@kernel.org> > Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com> > --- > drivers/acpi/arm64/gtdt.c | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/drivers/acpi/arm64/gtdt.c b/drivers/acpi/arm64/gtdt.c > index c0e77c1c8e09..f249af1ed1cd 100644 > --- a/drivers/acpi/arm64/gtdt.c > +++ b/drivers/acpi/arm64/gtdt.c > @@ -177,7 +177,8 @@ int __init acpi_gtdt_init(struct acpi_table_header *table, > } > > platform_timer = (void *)gtdt + gtdt->platform_timer_offset; > - if (platform_timer < (void *)table + sizeof(struct acpi_table_gtdt)) { > + if (platform_timer < (void *)table + sizeof(struct acpi_table_gtdt) || > + platform_timer >= acpi_gtdt_desc.gtdt_end) { > pr_err(FW_BUG "invalid timer data.\n"); > return -EINVAL; > } You are only checking the base pointer for the platform_timer array. This doesn't say anything about the *size* of that array (or at least its first element), and whether that actually fits in the table. M.
在 2024/10/10 23:52, Marc Zyngier 写道: > On Thu, 10 Oct 2024 15:47:03 +0100, > Zheng Zengkai <zhengzengkai@huawei.com> wrote: >> As suggested by Marc and Lorenzo, first we need to check whether >> the platform_timer pointer is within gtdt bounds (< gtdt_end) before >> de-referencing what it points at to detect the (first) platform >> timer entry length and check that next platform_timer pointer is >> within gtdt_end too. Now we do that only in next_platform_timer() >> for subsequent platform timers. >> >> So add check against table length (gtdt_end) for the first platform >> timer entry. >> >> Suggested-by: Marc Zyngier <maz@kernel.org> >> Suggested-by: Lorenzo Pieralisi <lpieralisi@kernel.org> >> Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com> >> --- >> drivers/acpi/arm64/gtdt.c | 3 ++- >> 1 file changed, 2 insertions(+), 1 deletion(-) >> >> diff --git a/drivers/acpi/arm64/gtdt.c b/drivers/acpi/arm64/gtdt.c >> index c0e77c1c8e09..f249af1ed1cd 100644 >> --- a/drivers/acpi/arm64/gtdt.c >> +++ b/drivers/acpi/arm64/gtdt.c >> @@ -177,7 +177,8 @@ int __init acpi_gtdt_init(struct acpi_table_header *table, >> } >> >> platform_timer = (void *)gtdt + gtdt->platform_timer_offset; >> - if (platform_timer < (void *)table + sizeof(struct acpi_table_gtdt)) { >> + if (platform_timer < (void *)table + sizeof(struct acpi_table_gtdt) || >> + platform_timer >= acpi_gtdt_desc.gtdt_end) { >> pr_err(FW_BUG "invalid timer data.\n"); >> return -EINVAL; >> } > You are only checking the base pointer for the platform_timer > array. This doesn't say anything about the *size* of that array (or at > least its first element), and whether that actually fits in the table. > > M. Yes, I will send v2 to check against gtdt_end for both entry and length of each array element. Thanks!
diff --git a/drivers/acpi/arm64/gtdt.c b/drivers/acpi/arm64/gtdt.c index c0e77c1c8e09..f249af1ed1cd 100644 --- a/drivers/acpi/arm64/gtdt.c +++ b/drivers/acpi/arm64/gtdt.c @@ -177,7 +177,8 @@ int __init acpi_gtdt_init(struct acpi_table_header *table, } platform_timer = (void *)gtdt + gtdt->platform_timer_offset; - if (platform_timer < (void *)table + sizeof(struct acpi_table_gtdt)) { + if (platform_timer < (void *)table + sizeof(struct acpi_table_gtdt) || + platform_timer >= acpi_gtdt_desc.gtdt_end) { pr_err(FW_BUG "invalid timer data.\n"); return -EINVAL; }
As suggested by Marc and Lorenzo, first we need to check whether the platform_timer pointer is within gtdt bounds (< gtdt_end) before de-referencing what it points at to detect the (first) platform timer entry length and check that next platform_timer pointer is within gtdt_end too. Now we do that only in next_platform_timer() for subsequent platform timers. So add check against table length (gtdt_end) for the first platform timer entry. Suggested-by: Marc Zyngier <maz@kernel.org> Suggested-by: Lorenzo Pieralisi <lpieralisi@kernel.org> Signed-off-by: Zheng Zengkai <zhengzengkai@huawei.com> --- drivers/acpi/arm64/gtdt.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)