Commit c732623b authored by Tim Sell's avatar Tim Sell Committed by Greg Kroah-Hartman

staging: unisys: visorbus: tolerate larger-than-expected controlvm channel

Use the dynamic size of the controlvm channel (struct channel_header.size)
instead of the statically computed sizeof(struct controlvm_channel) when
determining the valid bounds for visorchannel_read() and
visorchannel_write().

This prevents an observed problem where kdump was failing because
controlvm_channel.local_crash_msg_offset was pointing beyond the statically
computed size of the channel, even though the channel was physically large
enough.  This was causing visorchannel_read() to unecessarily fail, because
we thought we were attempting to access memory outside of the channel.
Signed-off-by: default avatarTimothy Sell <timothy.sell@unisys.com>
Signed-off-by: default avatarDavid Kershner <david.kershner@unisys.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
parent a8deaef3
......@@ -2251,7 +2251,6 @@ visorchipset_init(struct acpi_device *acpi_device)
{
int rc = 0;
u64 addr;
int tmp_sz = sizeof(struct spar_controlvm_channel_protocol);
uuid_le uuid = SPAR_CONTROLVM_CHANNEL_PROTOCOL_UUID;
addr = controlvm_get_channel_address();
......@@ -2261,8 +2260,10 @@ visorchipset_init(struct acpi_device *acpi_device)
memset(&busdev_notifiers, 0, sizeof(busdev_notifiers));
memset(&controlvm_payload_info, 0, sizeof(controlvm_payload_info));
controlvm_channel = visorchannel_create_with_lock(addr, tmp_sz,
controlvm_channel = visorchannel_create_with_lock(addr, 0,
GFP_KERNEL, uuid);
if (!controlvm_channel)
return -ENODEV;
if (SPAR_CONTROLVM_CHANNEL_OK_CLIENT(
visorchannel_get_header(controlvm_channel))) {
initialize_controlvm_payload();
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment