Donald Lane was convicted of being a felon in possession of a firearm in violation of 18 U.S.C. § 922(g)(1). He appeals both his conviction and his sentence. We affirm.
I. Background
Donald Lane knew that as a convicted felon and a parolee he was not allowed to possess firearms. But Lane enjoyed hunting and wanted to accompany his father and his girlfriend, Diane Stumph, on a small game hunt for rabbits. On two occasions, Lane discussed his wishes with his parole officer and eventually received permission to observe a hunt provided that he avoid handling any firearms.
On March 18, 2000, Lane and Stumph went to the Twin Oaks Tavern. While there, Lane overheard Leroy Bowen discussing his plans to sell a right-handed .22 caliber 10 shot Ruger. Lane expressed interest in purchasing it, so Bowen retrieved the gun from his nearby home. Back at the bar, Lane removed the gun from its box, held and inspected it, negotiated a price, and bought it. Lane says that he bought the gun for Stumph to use while hunting. He corroborates this by explaining that the right-handed gun was useless to Lane, a left-handed man. Stumph never. handled or examined the gun herself, but did give Lane money to buy it.
The gun ended up in Stumph’s car, although the parties disagree as to how it got there and the jury made no specific factual finding resolving the difference in the stories. The government presented Bowen’s testimony that Lane returned the gun to its box, picked it up and carried it outside himself. Lane insists that another man at the tavern, Russ Swonger, asked Lane if he could look at the gun. According to Lane, Swonger picked up the box and both men walked out of the tavern. Outside, Swonger examined the gun, returned it to the box, and without being asked, placed the box in Stumph’s car. Swonger corroborated Lane’s recollection. Lane points out that with a general verdict, we cannot be sure whether the jury concluded that he carried the gun to the car, or whether it based the felon in possession conviction exclusively on Lane’s examination of the gun.
*717Roughly one month after Lane purchased the gun, probation and police officers arrested him for violating parole. Lane allowed the officers to search the home he shared with Stumph and officers found the .22 Ruger in the garage. Lane was charged with being a felon in possession of a firearm in violation of 18 U.S.C. § 922(g)(1) and was convicted by a jury. Lane progressed to the sentencing stage of his trial. Unfortunately for Lane, the felon in possession conviction was not his first brush with the law. Lane had two prior state felony convictions, one in January of 1996 and the other in February of 1996. Realizing that his prior felonies would lengthen his sentence, Lane petitioned the court to disregard his January 1996 conviction, but the district judge denied his motion. These prior crimes boosted Lane’s basic offense level from 20 to 24, and his criminal history from a Category III to a Category VI which exposed him to a sentence between 77 and 96 months rather than a range of 41 to 51 months.
II. Discussion
A. Possession Standard
Lane first contends that the district court misunderstood the legal standard of possession and as a result, wrongly instructed the jury and prevented Lane from presenting a viable trial theory. To prove that Lane was a felon in possession of a firearm, the government was required to show that Lane (1) had been convicted of a crime punishable by a prison term exceeding one year and (2) knowingly possessed a firearm (3) that traveled in or affected interstate commerce. See 18 U.S.C. §§ 922(g)(1), 924(a)(2); United States v. Phillips, 239 F.3d 829, 847 (7th Cir.2001). We review Lane’s legal challenge de novo. See United States v. Stott, 245 F.3d 890, 904 (7th Cir.2001). At trial, the government agreed to focus only on the incidents at the Twin Oaks Tavern and not the gun’s presence in Lane’s garage. We likewise limit our inquiry.
At trial, the judge held that momentarily handling a gun satisfied the legal definition of “possession” as a matter of law. While possession can be actual or constructive, see United States v. Kitchen, 57 F.3d 516, 520 (7th Cir.1995), the government has pressed the actual possession theory. Actual possession occurs when a defendant “knowingly has direct physical control over a thing at a given time.” United States v. Walls, 225 F.3d 858, 864 (7th Cir.2000) (citing Kitchen, 57 F.3d at 520). Lane contends that although it is possible for physical contact to constitute possession, it does not do so as a matter of law. According to Lane, he was entitled to argue that the gun was not under his control when he held it because he was merely inspecting a gun owned by someone else. Lane relies heavily on Kitchen, 57 F.3d at 518-23, a drug case, to bolster his reasoning and to provide an example of a situation when momentarily holding contraband did not prove control or possession.
In Kitchen, the defendant appealed his conviction for possession of cocaine with intent to distribute. See 57 F.3d at 518-19. Kitchen’s conviction stemmed from a would-be drug transaction, when Kitchen, a drug buyer, met undercover federal agents to purchase cocaine from them. See id. at 519. The agents showed Kitchen the cocaine, which Kitchen picked up and inspected for two or three seconds. See id. However, before Kitchen could complete the transaction, the federal agents arrested him. See id. Kitchen contested his conviction arguing that he never possessed the drugs despite briefly handling and inspecting them. We reversed the conviction, reasoning that *718Kitchen’s momentary handling of the cocaine did not constitute possession in the context of the 21 U.S.C. § 841(a)(1) charge because Kitchen never exhibited assent to the drug transaction and therefore never demonstrated that he had the authority to exercise control over the cocaine. See id. at 523. We explained that “we require ... some factor indicating that [the defendant] had the authority or the ability to exercise control over the contraband.” Id. at 523.
In Kitchen we also made clear that “[w]e do not attempt to use the present case to formulate a rule workable for all circumstances.” Id. We find Lane’s case to be distinguishable. There is a marked difference between the steps necessary to exercise control over drugs and those necessary to control a gun. To deal with the “growing menace of drug abuse in the United States,” H.R.Rep. No. 91-1444, at 1 (1970), 1970 U.S.C.C.A.N. 4566, Congress made it illegal to possess drugs with the intent to distribute them. To obtain control over drugs in this context, a defendant needs more than just mere physical contact; he must have the perceived right among the criminals with whom he is interacting to deal, use, transport, or otherwise control what happens to the drugs. See Kitchen, 57 F.3d at 524. Such control is not a foregone conclusion when a defendant’s sole physical contact with drugs is momentary inspection of drugs he does not own or over which he did not have recognized authority. There is a meaningful distinction between physical contact and the ability or authority to control the drugs, so we require proof of a factor beyond mere physical contact to show that the defendant exerted authority or the ability to physically control the drugs.
In contrast to drugs, it is much easier to obtain control and therefore possession of a gun in the context of 18 U.S.C. § 922(g)(1). Congress originally passed this law as the Federal Firearms Act of 1938 “to ‘prevent the crook and gangster, racketeer and fugitive from justice from being able to purchase or in any way come in contact with firearms of any kind.’ ” Barrett v. United States, 423 U.S. 212, 220, 96 S.Ct. 498, 46 L.Ed.2d 450 (1975) (quoting S.Rep. No. 1189, at 33 (1937)). Throughout this Act’s subsequent history, this purpose has remained constant. See S.Rep. No. 90-1501, at 22 (1968); H.R.Rep. No. 99-495, at 1-3. From the purpose of the Act, it is reasonable to infer that Congress intended to prohibit felons from exercising any physical control over a gun.
Physical control over a gun is remarkably easy to effect. Once the gun is in the defendant’s hands he need only pull the trigger, an act which can be completed in a split second and which is controlled and influenced by nothing more than the defendant’s whim. Lane protests that the circumstances surrounding his inspection of the gun show that he did not possess it. He points out that he was merely inspecting the gun and that when he held the gun it belonged to Bowen. But none of these-circumstances bear on Lane’s ability to shoot the gun. Felons handling guns, unlike defendants who have touched drugs, do not need recognition of their authority or any extra element to obtain the ability to shoot the gun. Lane had just as much control over the gun when he inspected it while it belonged to Bowen as he would have if he as the gun’s owner took aim at a rabbit. Because a defendant can shoot a gun so quickly and easily once he holds it in his hands, we conclude that evidence showing that a felon held a gun is by itself a “factor indicating that the defendant had the ability to exercise direct control over the [firearm].” The distinction between holding a gun and obtaining control over a gun as required to prove possession is academic. We do not address whether *719touching a gun as opposed to holding a gun mandates the same result.
Lane asserts that in United States v. Wilson, 922 F.2d 1336, 1338-39 (7th Cir.1991) we implicitly decided that momentarily holding a gun does not constitute possession as a matter of law. In dicta, we offered as an example of “innocent contact” between a felon and a gun the scenario of a felon momentarily handling a gun while taking it away from children who were playing with it. We do not decide whether this example is correct because it may implicate a defense not relevant to Lane’s case. But as a general proposition of law, this statement was dicta and put forth without any reasoning to justify it and we decline to proceed down this path.
Our conclusion is further bolstered because it is consistent with holdings of two sister circuits. In United States v. Adkins, 196 F.3d 1112, 1115, 1117-18 (10th Cir.1999), the Tenth Circuit held that a convicted felon violated 18 U.S.C. § 922(g)(1) when, in his friend’s presence, he carried the gun she had just purchased from the gun store to her car. Explaining the law, the court stated that once it was established that a felon held a gun even “for a mere second or two,” the felon was guilty of possessing a firearm unless he truly did not know that the item was a firearm. See id. In United States v. Scales, 599 F.2d 78, 80 (5th Cir.1979), the defendant was a convicted felon who purchased a gun for his wife to use to defend herself. The defendant advanced the theory that he never actually possessed the gun because he never intended to exercise control over it himself, but only to deliver it to his wife for her use. See id. He requested jury instructions stating that possession required actual control, care, and management of the gun, not passing control, fleeting and shadowy in its nature. The court held that the district court correctly refused to issue this instruction. See id. at 81. These cases are virtually indistinguishable from Lane’s.
For the reasons above, we find that holding a firearm establishes possession as a matter of law in the context of a charge under 18 U.S.C. § 922(g)(1) and that neither of the district court’s challenged rulings were erroneous. This holding depends heavily on both the purpose of the law in question and on the physical nature of guns. For these reasons, we do not intend to make a rule that is applicable to other statutes or physical situations without analysis specific to both.
B. Motive Testimony
The parties’ struggle over the definition of possession spills into Lane’s next argument, that the judge abused her discretion by striking as irrelevant testimony about Lane’s motive for purchasing the gun. We review the district court’s decision to exclude evidence, such as the motive testimony at issue here, for an abuse of discretion. See United States v. Walton, 217 F.3d 443, 449 (7th Cir.2000). Parties who challenge a district court’s evidentiary ruling face an uphill battle. See id.
Lane offered the motive evidence to show that he did not possess the gun because he.did not intend to exercise control and dominion over it. As Lane points out, we commonly admit evidence of the defendant’s motive to prove possession in actual and constructive gun possession cases. See, e.g., United States v. Joy, 192 F.3d 761, 768 (7th Cir.1999) (finding no error when a district court admitted evidence that the defendant committed a burglary immediately before his arrest, thereby providing a motive for gun possession, making his possession more likely); United States v. Butler, 71 F.3d 243, 250-51 (7th Cir.1995) (finding no error when the *720district judge admitted evidence that defendant acted as security in a gang because it provided a motive for his gun possession, making it more likely). These cases differ in one important way from Lane’s — the motive evidence was offered by the government to make more likely the conclusion that the defendant possessed a gun. Although at face value, it seems fair to allow a defendant to make use of the same type of evidence as the government, a deeper analysis taking into account the evidentiary burdens of each party lead us to the opposite conclusion.
Because the government has the burden of proving the elements of the crime beyond a reasonable doubt, the relevance of its evidence depends on the elements of the crime. The defendant bears no such burden, but if he wishes, a defendant may rebut the government’s theories or evidence. Therefore, the relevance of the defendant’s evidence necessarily depends in part on the evidence the government has advanced. In this case, the government endeavored to prove possession by bringing direct evidence that Lane held the .22 Ruger in his hands. Defense evidence relevant to possession was needed to rebut the government’s theory. Had Lane used motive testimony to argue that he never physically held the gun, the evidence would have been relevant. But Lane took a different approach. He admitted that he knowingly held the gun but wanted to use the motive testimony to show that he did not intend to exert control over the gun. But as we discussed above, to prove possession, the government need only establish that Lane knowingly held the gun. Lane’s motive evidence was not relevant to the question of whether he held the gun and was therefore appropriately refused.
Lane protests that 18 U.S.C. § 922(g)(1) is not a strict liability crime and that his motive is therefore relevant to the element of possession. We disagree. 18 U.S.C. § 924(a)(2) incorporates a general intent requirement, namely that a defendant must have known that the object he possessed was a gun. See Bryan v. United States, 524 U.S. 184, 193, 118 S.Ct. 1939, 141 L.Ed.2d 197 (1998); United States v. Dodd, 225 F.3d 340, 344 (3d Cir.2000); United States v. Deleveaux, 205 F.3d 1292, 1298 (11th Cir.2000); United States v. Miller, 105 F.3d 552, 555 (9th Cir.1997). But this is the only intent requirement imposed by the statute. See United States v. Jones, 143 F.3d 1417, 1419 (11th Cir.1998); Dodd, 225 F.3d at 347; United States v. Langley, 62 F.3d 602, 604-05 (4th Cir.1995) (en banc). Lane’s intent to purchase the gun for Stumph had no bearing on the knowledge requirement. Because Lane’s motive fails to address either the possession element or the knowledge requirement, it is irrelevant, and the district judge did not err in excluding it.
C. Uncounseled Conviction
Last, Lane argues that the court erred when it took his uncounseled prior felony conviction into account when determining his sentence. Lane petitioned the district court to exclude from its consideration for sentencing purposes his January 1996 conviction. Lane’s January 1996 trial occurred after the case had been pending for two years. On the morning of the trial, Lane asked the state district court for permission to dismiss his attorneys. Lane’s counsel warned him against the dismissal and the prosecutor made clear that he wanted the trial to proceed that day even if Lane had to represent himself pro se because the witnesses were present. Lane persisted in his wish to dismiss his counsel and the judge allowed him to do so. Although Lane persistently and clearly requested a delay in the trial so he could *721hire new counsel, the state judge forced Lane to proceed pro se. In February of 1996, Lane was again in state court as a defendant in another felony case. Represented by counsel, Lane pled guilty to this offense. As part of his plea agreement, Lane agreed to withdraw any appeals regarding the January 1996 conviction and to let the conviction stand.
Before the district court, Lane argued that the January 1996 conviction was obtained in violation of his Sixth Amendment right to counsel and therefore should not be considered part of his criminal history for sentencing purposes. The judge denied Lane’s petition for three reasons: (1) Lane was not denied his Sixth Amendment right to counsel at his 1996 trial; (2) a defendant cannot collaterally attack a prior state conviction during sentencing; and (3) Lane waived his right to challenge his 1996 conviction in a subsequent plea agreement. Lane claims that all three reasons are erroneous.
We can dispose of this issue by addressing only the waiver argument. Because the judge essentially determined that Lane procedurally defaulted on his claim by agreeing to waive it in the February 1996 plea agreement, we review the issue de novo. See Braun v. Powell, 227 F.3d 908, 911 (7th Cir.2000). In the February 1996 plea agreement, Lane agreed to withdraw “any appeal that has been heretofore filed on the convictions that were obtained following his jury trial in January 25, 1996 ... [and] to let those convictions stand as part of his plea agreement.” The government correctly contends that by the plain language of the February 1996 plea agreement, Lane waived his right to appeal directly or collaterally attack his January 1996 conviction.
Lane protests that while' the plea agreement waived his right to direct appeal, he retained the ability to collaterally attack the conviction. He bolsters his argument by pointing to several cases where plea agreements explicitly waived both the right of future direct and collateral appeal. But such explicit language is not required to waive the right to collaterally attack a conviction. We settled long ago that waiving or foregoing a direct appeal bars collateral attack on the basis of most issues, even many constitutional issues, that could have been raised on direct appeal. See Johnson v. United States, 838 F.2d 201, 202 (7th Cir.1988) (holding that a doctor who was convicted of prescribing amphetamines and barbiturates without a prescription, and who had filed an appeal and withdrew it could not later mount a collateral attack in the form of a writ of corum nobis based on issues that were available for direct appeal); United States v. Behrman, 235 F.3d 1049, 1051-51 (7th Cir.2000) (holding that a defendant who signed a plea agreement waiving his right to appeal also gave up his right to collaterally attack his sentence on constitutional grounds not implicating the validity of the plea agreement); Daniels v. United States, 54 F.3d 290, 292-93 (7th Cir.1995) (holding that a defendant who attempted to mount a collateral attack on his guilty plea because the judge purportedly failed to conduct an adequate Rule 11(d) colloquy waived his right to do so by failing to pursue the issue on direct appeal). Lane was aware of his Sixth Amendment issue when he signed the plea agreement; thus he had an opportunity to appeal and freely waived it.
We make an exception to the waiver rule for parties who can show that they had cause for failing to appeal directly and that they suffered prejudice from this omission. See Wainwright v. Sykes, 433 U.S. 72, 84-85, 97 S.Ct. 2497, 53 L.Ed.2d 594 (1977); Johnson, 838 F.2d at *722202-05. Lane does not meet this standard in part because he cannot show cause. Lane knew about the purported Sixth Amendment violation when he signed the February 1996 plea agreement and freely chose, on the advice of his lawyer, to waive his right to appeal. We routinely allow defendants to waive constitutional rights in plea agreements, see Behrman, 235 F.3d at 1052, so the constitutional nature of Lane’s issue gives us no particular pause.
Affirmed.